Talk:Roman numerals/Archive 8

Characters not showing

 * Firefox, Internet Explorer and Chrome do not show the characters for 50 and 500 at the end of the sentence "The use of separate symbols for 5, 50, 500 etc. is a feature of the old Greek numbers system..." in the section "Origin of the system". George Rodney Maruri Game (talk) 07:48, 5 May 2019 (UTC)
 * Browsers depend on fonts installed on your machine to render character codes. Not all fonts have all Unicode characters. Those characters that you mention are in the Ancient_Greek_Numbers_(Unicode_block).  Their high code values (starting at 10140 hexadecimal) imply that they were added somewhat recently. Unfortunatey we don't have any easy way to display those characters in the text without resorting to Unicode.  You will have to find and download fonts that cover those codes. All the best, --Jorge Stolfi (talk) 11:32, 5 May 2019 (UTC)
 * This seems like an obstinate refusal to provide an alternate, more compatible way of displaying the characters - for example, a simple graphical table or suchlike. The fact that, after trying both Chrome under Windows, and the default Webkit browser on my phone, I have not seen anything rendered for the "missing" / high codenumber characters than a blank box (or a box with a cross in it) suggests that these characters simply will not show up for the greater majority of WP readers, as these are by far the world's most popular OSes and browsers... hardly anything unusual or outdated. If it was a common thing for up-to-date browsers to come with a font showing these characters, I'd be seeing them right now. Thus, the page has been designed for use with exceptional browsers (or by exceptional readers), which is surely a mistake.
 * Relying on the majority of everyday users to go to the time and trouble of seeking out and installing custom fonts - which, ignoring the fact that it's a technical tasks that will be beyond a lot of them (and possibly completely blocked for others, especially in schools), is quite a heavy download and additional strain on system resources (in order to work properly, it would have to replace the browser's default font, and be used for all pages), and one I doubt I could make work on my rather memory and CPU bound handset - just so a couple of paragraphs on a single wiki page will render correctly, and so the authors of the page don't have to put themselves to the trouble of creating and inserting a graphic, is to my eye somewhere between madness and just plain rude.
 * The characters show up just fine - as readymade graphics - on UnicodeTable.com when I select them and do a websearch for each, so it wouldn't be particularly burdensome to whizz up a table of them. Might even do it myself if I have time later on, but I don't have an account so I'm not entirely sure how I'd then insert it. Alternatively it could probably be built from just arranging the UT files in a grid, or maybe even inlining them into the text (with a considerable downsize in their image tag), or even just providing a suitable link to each character's page on UT behind the relevant errorbox. 146.199.60.87 (talk) 17:22, 22 August 2019 (UTC)


 * Edit: Article is semi-protected, so I can't commit any of those suggested fixes regardless. Going to have to rely on the sense and diligence of registered editors in this case.
 * If it's any use, here's links for the first two blocks and ; and direct to the images... oh... huh. The page must have a special font embedded into it (which is actually a thing these days; in this case, "u10000.woff", which presumably only encodes a certain subset of the larger table, and so provides the specific capability quite efficiently), because the large versions that display properly are actually characters not images (whilst the small ones, which aren't HTML coded to use the special font, are still rendered only as empty boxes). Can we not attempt that on WP somehow? Surely there must be some provision for that kind of capability in the wiki software... and if not, it's something that it's well worth petitioning for the addition of? 146.199.60.87 (talk) 17:35, 22 August 2019 (UTC)
 * The Number Forms article is already using small images for these, you can copy them from there.Spitzak (talk) 01:03, 23 August 2019 (UTC)

General "Variant forms" heading
This was lost in a flurry of edits - I have put it back (without altering any of the other changes, useful or inconsequential, made at the same time). The main value of this can be seen in the contents list - it is now much more clearly evident, especially to one of the 4,000 or so people who actually consult it on a daily basis - where the part of the article dealing with "non-standard" RNs actually starts. --Soundofmusicals (talk) 09:22, 23 August 2019 (UTC)
 * Thanks for all your work looking after this article. I have been watching since 2010 but haven't done much apart from create rn. However, I support your edits such as the one mentioned. Johnuniq (talk) 10:29, 23 August 2019 (UTC)
 * Can't say how much I appreciate ANY support for the otherwise more or less thankless headbanging I have gone through over this one - but especially is it welcome from an editor so eminent (any apparent crawling and sarcasm not intended - well not very much anyway). Best wishes. --Soundofmusicals (talk) 12:20, 23 August 2019 (UTC)

Etruscan numerals
There is a strong case, IMHO, for including a brief mention of the Etruscans in this context - but the information about other numeral systems - less directly related to the topic of this article as they are - was not only stubbornly uncited, but also not really necessary. If anyone really misses anything I have excised, and feels contrained to restore any of it - please cite everything to a reliable source, and ensure that it has a direct connection to the subject of the article (which was, last time I looked, ROMAN numerals). --Soundofmusicals (talk) 07:03, 28 August 2019 (UTC)

"IV vs IIII" revisited (copied from my talk page)
I take your point that we should not "ape" references and rather use our own words, however nor should we alter or inflate the meaning of them (seeWP:STICKTOSOURCE). "Most" has a very different meaning to "many" as I'm sure you will agree. As such, I'll come an alternative way of expressing more accurately the meaning of the reference. Cheers Jschnur (talk) 02:55, 12 August 2019 (UTC)


 * By all means (I take YOUR point too, of course) - or, and this may be an even better idea, find an alternative reference that is closer to our text! In the aftermath of making my latest edit I had fun looking at some of these (just try Googling "Roman numeral clock IIII" ) - interestingly, MANY (if not MOST) of these use a phrase that includes MOST rather than MANY.
 * But let's get this into proportion - the best and most important reason for having citations at all is to increase the accuracy of this encyclopedia by cutting down on "original research" (in the sense of editors' pet notions). But the "notion" that not just MOST, but practically ALL Roman numeral clocks use "IIII" rather than "IV" is hardly one crying out for specific verification - "public" exceptions like Big Ben are actually very rare indeed - most exceptions are in the "private" or "deliberately non-traditional" category! --Soundofmusicals (talk) 03:26, 12 August 2019 (UTC)


 * I took your suggestion of Googling "Roman numeral clock IIII" and sure enough, I found myself falling into a rabbit hole full of twists and turns involving at least four different theories (more like speculations really) on why most clocks seem to use IIII. I now think you were right to revert my edit, I just wish we had a citation that used that word rather than many (and that my kitchen clock didn't use a "IV"). Jschnur (talk) 07:16, 12 August 2019 (UTC)


 * I think the current citation is the best of a bad lot. Remember this is the "lede" - we treat the question more comprehensively in the body of the article, where we can better sort out what "citations" to refer to. A citation that is at best a blog deserves less reverence than (say) a prestigious print encyclopedia. [edit: in fact I've substituted a "reputable newspaper one]
 * Another point is that the difference between "many" and "most" (in this context, at least) is as points on a continuum running between "few", "some", "many", "most" and "practically all". None of these are "different" in a precise (or, in this case, measurable) sense.
 * Before the current trend surfaced for wrist watches, kitchen clocks and such (what I referred to as "private" timepieces in my last post) to use Roman Numerals, virtually all RN clock faces appeared on public clocks and used the traditional form (each numeral oriented radially rather than vertically, as well as using IIII instead of IV). To be fair, most still do. --Soundofmusicals (talk) 13:39, 12 August 2019 (UTC)
 * We now cut the Gordian knot by not using either word. --Soundofmusicals (talk) 07:11, 28 August 2019 (UTC)
 * Much better :-) Jschnur (talk) 21:22, 28 August 2019 (UTC)

Subtractive notation reason
Apparently nobody is allowed to state the blatantly obvious reason for subtractive notation (that it is shorter) because that is "speculation". However for some reason the ridiculous idea that it is based on the Latin or Etruscan (!) pronounciation of numbers is allowed??? That explanation is stupid because subtractive notation is used for plenty of numbers (like 4 and 9) that are not pronounced that way. If "shorter" is disallowed "speculation" than certainly this pronounciation explanation is as well.Spitzak (talk) 15:58, 23 August 2019 (UTC)
 * The actual edit you made in the text was "well picked" (hence my thanks). The "shorter" bit is SO bloody obvious (excuse Australian) that it doesn't seem to NEED a "reliable source" type citation. This kind of thing (which of course abounds thoughout Wikipedia) used to drive me crazy, too, so I deeply empathise. On the other hand the Etruscan bit IS (apparently) cited, and therefore not to be removed except in extremis - at first blush Wiki rules on this kind of thing may seem pedantic and illogical but try to imagine the alternative, with completely open slather for everyone's pet notions! In some ways each article having a single editor, as in a conventional encyclopedia, would definitely be advantageous - but for all its frustrations the way things are oranised here DOES have advantages too. In any case I long ago gave up "campaigning". --Soundofmusicals (talk) 21:55, 24 August 2019 (UTC)
 * Although we did manage to keep the "shorter" (and even the "more distinctive") explanations in a footnote. Also see the next topic - which is about the Etruscans in another context. --Soundofmusicals (talk) 07:08, 28 August 2019 (UTC)

Copyright law
This came up while I was looking for a reference that it is actually possible to simply "misspell" a Roman numeral! This (IMHO) DOES actually add something useful to the article, which is more than can be said of most recent nitpicky edits!!! -Soundofmusicals (talk) 04:25, 6 September 2019 (UTC)

"Three score" and "four score". (and "Nineteen")
"Score" for twenty - and "three score" for sixty, "four score" for eighty etc. are all still current (if rather literary/archaic) English. "Three-twenty" is not English at all. In fact probable confusion with "three and twenty" is another reason to avoid word-coinage in this context. Similarly - we just don't say "ten-nine" in English - the word is "nineteen". Nothing "colloquial" about this - there isn't a colloquial English equivalent for "nineteen", or a more formal equivalent either for that matter. "Dix-neuf" on the other hand is the precise equivalent (i.e. the most "literal" translation possible!) of this in French. Might one (very humbly) suggest that whatever the function of an on-line encyclopedia might be, coining new words is not one of them? --Soundofmusicals (talk) 01:45, 6 September 2019 (UTC)


 * "Score" is indeed standard English and is an ideal translation as in "quatre-vingt" (four score). Furthermore, it illustrates that the pattern was/is common in English.  The real issue is with the translations from French where the French is being used to show a pattern.  If you were saying that 19 is dix-neuf in French and nineteen in English then that would also be uncontentious.  However the two phrases in question are explicitly using the French pattern of language to illustrate a point, then providing an English translation of the French, not the English version of the number.  " for 99, reflecting the French reading of that number as quatre-vingt-dix-neuf (four-score-ten-nine)" takes the number, provides an unambiguous modern version (99) and then compares the French form "quatre-vingt-dix-neuf".  The parenthesised "(four-score-ten-nine)" is clearly an interpretation of the French for those who may not be familiar with it.  To labour the point a little; you could recast it as " for 99, reflecting the French reading of that number as quatre-vingt-dix-neuf or in English four score and nineteen" but why?  One can assume that the reader has that already from "99".  Finally as regards "coining new words", the last time "nine" and "ten" were neologisms Alfred was still burning his cakes.  Let's simply go back to the format at the start of this month, the status quo ante.  Martin of Sheffield (talk) 08:03, 6 September 2019 (UTC)
 * This remains the English version of Wikipedia, and we cannot assume that "quatre-vingt-dix-neuf" will be anything but gibberish to readers with no French. On the other hand readers who CAN at least count in French will be very well aware that "dix-neuf" is French for nineteen, from "dix" (ten) and neuf (nine). Translating "dix-neuf" as "ten-nine" imparts no new information to a French speaker (or at least a "French counter") - in fact it may be mildly confusing. To someone with as much French as I have Tibetan, on the other hand, it offers nothing but total confusion. "Ten-nine" has the added drawback, as I tried to point out in my last post, of not being English. Nineteen is not just the BEST English language equivalent of "dix-neuf" (not to mention the best English name for the number represented by the numeral "") it is the ONLY equivalent, unless we are determined to coin a new word ("ten-nine") for the number between eighteen and twenty. The improvement represented by the current text over the status quo ante may not be of major importance, but it is definitely worth hanging onto - don't you think? --Soundofmusicals (talk) 00:39, 7 September 2019 (UTC)
 * Having some fun with the idea of forcing numerals (Roman or Arabic) into a shape determined by linguistic rather than mathematical patterns! In a language I happen to be fluent in due to a period of foreign residence in my (very distant) youth "ninety nine" comes out as "taurahani-ta ahui taurahani-ta" or "twice-four and one tens twice-four and one". Just imagine THAT lot in Medieval Roman numerals - perhaps ""? --Soundofmusicals (talk) 01:21, 7 September 2019 (UTC)
 * I understood the "four-score-ten-nine" or "four-twenty-ten-nine" not as a translation of the whole, but rather showing what each part of the French term meant individually. It shows that the French term is built up from smaller numbers. Perhaps something more like "the French word for 99 (quatre-vingt-dix-neuf) is constructed from the French words for 4, 20, 10, and 9" would better convey the meaning. --Khajidha (talk) 23:56, 19 September 2019 (UTC)
 * "Showing what each part of the French term meant individually" is not really relevant to an English language encyclopedia article on Roman numerals. A person with at least enough French to count to twenty will probably get the point without us labouring it - but a person with no French will probably not get the point at all. Much better to stick to English - which we can assume is reasonably familiar to all our readers, than insert something specifically for the French speakers, which many users will find confusing. Sorry to be a bit repetitive, but ... --Soundofmusicals (talk) 09:33, 21 September 2019 (UTC)

"Original research"
A very innocent little paragraph (which shall remain nameless to protect the guilty) is STILL being battered with requests for citation. Even if it should be decided (by whom?) that a statement that boils down to a gentle reminder that ancient (medieval, renaissance etc. etc.) scribes and stone masons were every bit as human (i.e. prone to error) as Wikipedia editors, and may even have made mistakes at times! is so very contentious and unlikely that it could come under the heading of "research" (original or otherwise) then why not simply find a [naughty Australianism suppressed to protect the innocent] reference themselves? In fact this bit is so VERY "the sky is blue" obvious that one wonders if the kindest thing is simply to delete it. Personally, I (my "first-person" self, to ditch for the moment all pretence of academic objectivity) think that this would be rather a shame. A hypothetical user, faced with the problem of what a particular "Roman numeral" actually means - may very well actually need this final fall-back - "hey - it may just be a mistake! - people have always made them!". A constructive attempt to word this better, or even supply a sensible citation, or, indeed, anything that actually improves the usefulness of this rather frequently consulted article - would certainly attract no argument from me (myself and I). --Soundofmusicals (talk) 01:45, 6 September 2019 (UTC)


 * @Soumyabrata. Since you STILL don't think it's worth your bother adding edit summaries I really thought you'd fiddled things much more than you had, so my latest "re-edit" gets things back to pretty much how you left them - apart from the "cn" template. I had already started a new discussion on this one, so you really were out of order restoring this - at least without adding to the existing thread with an account of WHY this looks like OR to you. I'd appreciate your reading my comments and pointing out why you can't accept this without a citation. Perhaps we need a few comments from other editors. In any case it's not the kind of thing you CAN cite - all we are saying, as I mentioned above all we are saying is that some alleged "Roman numerals" are simply wrong - people do make mistakes and always have - in ancient and medieval times because so few people were literate, and (especially) nowadays, when most people are not that fluent with Roman numerals anyway, and are liable get the wrong end of the stick. We need to either decide that this paragraph needs to go altogether - or leave it in without the "cn". -Soundofmusicals (talk) 19:54, 8 September 2019 (UTC)

Statue of Liberty tablet date
I think that this National Park Service brochure would be a more reliable source. https://www.nps.gov › stli › upload › STLI-Statue-Stats_Rev --Khajidha (talk) 15:48, 19 September 2019 (UTC)
 * By all means - would you like to add the reference yourself, since you found it? - shame we can't use one of those excellent graphics, but I expect they are all copyright. --Soundofmusicals (talk) 23:13, 19 September 2019 (UTC)
 * Nah, you go ahead. Citation templates give me headaches. --Khajidha (talk) 23:46, 19 September 2019 (UTC)
 * I think a clear photo of the tablet and its inscription would be helpful. Such a photo is on Commons, which comes from a Historic American Engineering Record of the statue made during its renovation in 1984–86. A link to it without adding the picture is: Inscription on tablet. — Joe Kress (talk) 00:31, 20 September 2019 (UTC)
 * US government publications are usually public domain. Does that not apply to their images? --Khajidha (talk) 00:41, 20 September 2019 (UTC)
 * You are correct. The cited image, Inscription on tablet, would not be on Commons unless it was in the public domain. It can be added to the article at will. It is a government project to photograph the details of all National Park Service structures via HABS (Historic American Buildings Survey) or HAER (Historic American Engineering Record). I have used the many corresponding photos of details of the Washington Monument in its article. They are stored by the Library of Congress on its website. Many/most have already been copied to Commons simply because they are in the public domain, even without any anticipated use in a Wikipedia article. — Joe Kress (talk) 19:26, 20 September 2019 (UTC)
 * This is NOT, I feel, a place where a citation is really required anyway. 1776 as the year of American Independence surely does NOT require verification (!?!) Trying to fit an illustration here, in the middle a list of examples, would be very fiddly indeed, the picture would have to be presented in a very small format and would lose a good deal of its clarity. All in all - if someone to whom this is really vital wants to make the changes, and can do so neatly, then go for your life. As far as I am concerned it is not worth the bother. --Soundofmusicals (talk) 10:21, 21 September 2019 (UTC)
 * I only specified a link to the photo, not the photo itself, because I thought the photo would be too obtrusive. The link can be unobtrusive in a ref. — Joe Kress (talk) 04:25, 22 September 2019 (UTC)

Pointless?
The use of "Roman numeral letters" in non-numeric senses has caused various kinds and degrees of confusion (not to mention wailing and gnashing of teeth) in the real world (for just one example see No. 29 Squadron RAF) - links to common numerals like XXX and XL, which may very well NOT be the numerals for 30 and 40 (depending on context) are at least as valuable as many other factoids in this section - since both the numerals themselves and the "non-numeric equivalents" are still in common use and not just enshrined on a single obscure inscription! Suggest we keep this, anyway. --Soundofmusicals (talk) 23:07, 19 September 2019 (UTC)


 * Not only keep it, in fact, but retain its context! This is not really about the way Roman numerals are still used - but what strange and unfamiliar "Roman numerals" a user may encounter. Well worth pointing out that "apparent" combinations of "Roman numeral letters" may bear other, non-numeric interpretations. Having said that, the new text under this heading may be a little clearer. --Soundofmusicals (talk) 09:39, 21 September 2019 (UTC)

A "base 10" system
The numerals are not really a base ten system, they are more correctly described as bi-quinary system. I'm keeping half an eye open for a citable reference and if I find one I'll update the page. Martin of Sheffield (talk) 15:06, 18 June 2019 (UTC)
 * In a sense all decimal counting can be called bi-quinary as the very idea of counting in tens is based on the fingers of two human hands of course - but the use of "V", "L" and "D" (sometimes described as "quinary numerals") as non-repeating tally marks clearly marks them as an intervening abbreviation device between the primary symbols "I", "X", "C" (and "M", although M is a special case, as it lacks a quinary symbol of its own). It would be helpful if you ran anything counter to this past the other editors of the page first, at least, rather than rushing in where angels fear to tread. --Soundofmusicals (talk) 09:43, 19 June 2019 (UTC)
 * Decimal counting with our familiar digits is in no sense bi-quinary. There are ten digits (including zero) and the first change to the pattern occurs at ten.  The hands issue is a sideline; children and the innumerate will often use all ten fingers and thumbs as a simple 1-10 tally with no splitting for left and right.  Stephenson's work here is highly relevant where he shows the use of spaces on the abacus, plus of course the hand abacus (illustrated) is clearly bi-quinary.  Indeed if (and I have neither citation nor proof at the moment) the Romans used the bi-quinary finger counting method then you can tally up to 100 - a fair sized flock of sheep. Martin of Sheffield (talk) 10:03, 19 June 2019 (UTC)
 * Of course "Arabic" numerals are not in any sense bi-quinary, especially in the sense of being base 5 rather than base 10. The point is that neither are Roman ones! The "quinary Roman numerals" are all non-repeating ("VV", "LL" and "DD" do not ocurr) - plus the repeated pattern which can be seen in the numerals for the powers of ten, and the way each power is written separately clearly show this. --Soundofmusicals (talk) 10:21, 19 June 2019 (UTC)
 * The numerals are not the quinary part, they are the bi part, as in binary.  And just as in binary the digits are absent (binary 0) or present (binary 1).  The quinary part is .  Compare a true quinary system: 0, 1, 2, 3, 4, 10, 11, 12, 13, 14, 20 ... 44, 100. Martin of Sheffield (talk) 10:50, 19 June 2019 (UTC)


 * Regardless, it cannot be described as decimal (where's the decimal point?) or n-ary for any n other than 1, as it is not a positional/place-value number system, it's a sign-value notation with symbols for some groupings of multiples of 2 and 5. Not sure if a simple terminology or taxonomy exists to describe it, but it is simply inaccurate to refer to it by notions of base/radix. I'm removing "decimal", as it's inaccurate and also inconsistent with the article on numeral systems — Preceding unsigned comment added by 74.96.154.122 (talk) 19:23, 26 September 2019 (UTC)


 * A decimal point is not needed for a system to be regarded as decimal. According to decimal separator positional decimal fractions appear in the 10th century, would you dismiss Hindu-Arabic numbers prior to that as non-decimal?  Roman numbers are not a positional system, but mechanism is distinct from basis, see numeral system.  The mechanism is non-positional but the basis is decimal.  Describing the system as a unary numeral system is clearly wrong.  The first three digits are derived from tally marks, but then so are 1, 2 and 3 (see §4.2 of Hindu–Arabic numeral system and associated illustration).  Numbers from IV upwards do not fit the unary pattern.  Numeral system refers you to list of numeral systems where the table in §1 gives the base of Roman numerals as 10.  Bi-qinary is the term I've heard used, but lacking a citable authority I'm reverting to decimal as the best description, accurate and consistent. Martin of Sheffield (talk) 20:29, 26 September 2019 (UTC)

The passage in question was meant to be a descriptive (rather than an analytical or a prescriptive) summary of how RNs actually work - I hope all this theorising won't render it incomprehensible. Or is it the Decimal article - which seems pretty incompatible with this - that needs rewriting? It would be nice to get this right. --Soundofmusicals (talk) 05:16, 27 September 2019 (UTC)
 * I think the basic problem is the confusion which has arisen, generally not just here, between a decimal system (base 10) and the Decimal (a specific development of the Hindu-Arabic numbers). Unfortunately Decimal rather falls into this trap, as did  above.  The passage in the main article is meant to be fairly stable, which is why this talk page is the appropriate venu for discussion. Martin of Sheffield (talk) 07:23, 27 September 2019 (UTC)
 * The part of the system (of RNs) we DO still use is concerned entirely with integers. "Roman fractions" are indeed duodecimal, and for that reason quite unusable in the year MMXIX! (except, perhaps, for =½). This has nothing to do with the "base 10" nature (of RNs). So yes, I think we are on pretty much the same wavelength at this point. What I fear is a descent into tiresome, irrelevant and confusing technicality. This is essentially a "practical" article - "higher" mathematics are incongruous in a discussion of the numeric system of an ancient people as uninclined to such things as the Romans! --Soundofmusicals (talk) 19:15, 27 September 2019 (UTC)
 * Notwithstanding the above - I have added a judicious "in that" to hint at the respects in which RNs are "base 10" and "decimal". Hope this adds to clarity. --Soundofmusicals (talk) 22:00, 15 November 2019 (UTC)

Big Numbers
If this has been discussed before, you have my apologies. I’ve heard about multiple bars over letters like a billion is two bars over M. I didn’t read the article to well. Maybe it’s been covered. Redding7 (talk) 21:19, 10 February 2020 (UTC)


 * See next thread (although this arose in a different context)
 * There is no need for a Roman numeral larger than 3,999 (MMMCMXCIX).
 * There is no standard way of writing one anyway (although there are any number of ways, historical or hypothetical, of doing so if you really wanted to. --Soundofmusicals (talk) 23:23, 26 February 2020 (UTC)

Vinculum
On the whole I approve of the recent deletion of hypothetical "this might be what it would look like if we used it now" matter under this heading. This (on reflection) always belonged on the talk page rather than in the article itself. The trouble is that quite a lot of people (if you have a look through the rubbish a google search will bring up you'll see what I mean) assume vinculum is part of a "current standard". It actually IS the neatest way of writing really large numbers in "Roman numerals" but then who really wants to anyway? So long as we can write the current year, which we will be able to do for another MCMLXXIX (or 1,979) years, the ability to write a number larger than MMMCMXCIX (3,999) is redundant, one might have thought. --Soundofmusicals (talk) 23:14, 26 February 2020 (UTC)

"Largest number"?
I don't understand why 3,999 is supposed to be the "largest numeral". As the article says, the original way to write 4 was IIII, until they abbreviated it to IV. So why could not one write MMMMCM for 4,900? Indeed, I see no reason a person couldn't write MMMMMMMCM for 7,900. There is no rule forbidding writing numbers higher than 5 without using the special symbol for that number. One can write XIIIII for 15, the V is just for convenience. I would also like a bit of clarification; I gather from reading this that there is no 'right' way to use Roman numerals (not sure if that's the intent, but it's how it reads), but I saw the date MCMILXX for 1969 in a book. Somehow this seems wrong to me. ILXX apparently is supposed to mean '1 before 70', but shouldn't it read MCMLXIX, for 1000 + 900 + 60 + 9? That's the way I always thought it worked, but I see no confirmation either way in this article...other than to tell me that it apparently depends on who is writing it. That seems like it creates a lot of potential for confusion.

64.223.93.202 (talk) 20:19, 6 May 2020 (UTC)


 * because you're only allowed to sequentially iterate up to 3 times, any exception to that is considered variant/unorthodox. Yes, 1969 should be MCMLXIX; ILXX is not a legal combination (the subtraction is smaller than one tenth, and is not applied to the rightmost numeral). If you want to go high (4k and up) you'll need to apply vinculums. Check the technical ruleset I put up, that should answer any questions. Xcalibur (talk) 07:09, 14 May 2020 (UTC)


 * This was perfectly well covered already in the text of the article, which actually included the passage (we put it into a note to make it "flow" better - perhaps this was a mistake - assuming notes don't get read! (Even by people trying to answer questions?)
 * Since the largest Roman numerals commonly used today are year numbers up to thepresent, the need for larger Roman numerals seldom arises. For larger numbers more "M"s, as required, could be added without ambiguity, although this would quickly prove cumbersome. During the centuries that Roman numerals remained the standard way of writing numbers throughout Europe, there were various extensions to the system designed to indicate larger numbers, none of which were ever standardised. In any case we have restored the note to the main text. --Soundofmusicals (talk) 12:13, 14 May 2020 (UTC)


 * Incidentally, the answer to to OP's other question - no, there IS no fixed set of "rules" - you read that bit quite correctly (although we might have added that one usually tries to stick by the standard, none the less). Your "corrected" version for 1969 of "MCMLXIX" is at least much more usual! Under "description" there is in fact already a very full description of this kind of construction. No way could (or should) we describe every "mistake" it might be possible to make. --Soundofmusicals (talk) 14:24, 14 May 2020 (UTC)

Ruleset for Roman Numerals, revisited
I've been away for some time. I didn't intend to take a wikibreak, but it happens. Hopefully I can ease back into this, and in light of that, I'd like to make another attempt at adding formal rules to this article. My ruleset has been renovated and fine-tuned, much improved; perhaps relevant parties will take a different view this time. This article would benefit from having two different approaches side-by-side: a layman's description, and a more thorough treatment. I'm offering a more advanced description of the system which paraphrases RS. Given that there's plenty of interest in modern rules, and these are strongly supported and comprehensively described, I think this is a reasonable addition. Let me know if there are any issues, if there are, we can always move the relevant content here and discuss it. Xcalibur (talk) 06:57, 14 May 2020 (UTC)
 * Text like "formal, detailed approach" indicates a WP:NOTHOWTO problem. Is there anything missing in the article that the proposed text fixes? Johnuniq (talk) 07:30, 14 May 2020 (UTC)
 * You bring up valid concerns. first, this doesn't violate WP:NOTHOWTO because it's offering a detailed description of the system, not a step-by-step guide for converting a value to its Roman Numeral equivalent (which would indeed trip over that restriction). The article lays out basic ground rules, which is quite effective as an introduction for a layman. However, it doesn't have the comprehensiveness or precision of my content, which is useful for answering the many questions that come up on this topic (eg why can't I write RN this way, which string of RN is correct, etc. plenty of which you'll see on this talk page). In addition, my content paraphrases what the RS have to say on the topic, which I cited in the section lede. For these reasons, I believe my contribution is worthwhile. Xcalibur (talk) 07:45, 14 May 2020 (UTC)


 * I wish we had had a chance to debate this properly BEFORE the article was changed unilaterally! Anyway, let's just look at a few problems. Some if not all of these were covered in our original debate.
 * There are innumerable different (and often contradictory) "sets of rules" out there (many of which can be "cited"). The problem is that the whole notion of "rules" sits very ill as a "companion" piece to the very simple description of the current text. The descriptive (describing what actually happens) can never sit well with the proscriptive (a list of rules and instructions of what (not) to do). This is of course classic WP:NOTHOWTO - although I hesitate as a rule to argue from any Wiki guideline, especially in a dogmatic way, in a case like this. Point is, anyway, that if the description has described properly, then you don't need a "how to". It has to be one or the other (all else confusion!). In the previous debate I even suggested (or I should have) that a deficiency in the description might be better added there than in an alternative take on the whole thing.
 * There IS no universally recognised set of rules anyway - and never has been. The current (more or less) standard "orthography" (not sure this is actually the right word anyway, but let's not quibble) did not "arise" in recent times but is very close indeed to the usual form the numerals had already taken by the early Empire period, (say the century BCE 50 - CE 150). The (sometimes quite startling) innovations of the medieval and early modern periods have been, in fact, more or completely discarded and forgotten. In any case, we describe them specifically later in the article. Usage sine the age of enlightenment, when there was a minor revival of interest in roman numerals has conformed to a rather dogmatic idea of revived 'classical' usage. In any case "variant" forms (what you and I might class as straight "mistakes" remain common.
 * Any set of rules actually NEEDS jargon (in the sense of words used in a specialised and more or less technical way) if it to be reasonably concise and at the same time intelligible. But we cannot presume a general reader will be able to make sense of our jargon if the explanations are unclear or contradictory (or even just not there!). What is our reader here supposed to make of terms like "decimal/quinary numerals" or "additive/subtractive notation", for example - especially when these terms are used in ways that clash with the meanings they are given elsewhere in the article?
 * Roman fractions and the use of the Vinculum or barring (only one way the Romans wrote large number anyway!) have no place in any set of rules of current usage - simply because we don't use them nowadays at all - no we don't (really)! The cryptic reference to the them here is in any case badly out out of place in that both are explained properly elsewhere in sections of their own.


 * Sorry if this is seen by anyone as less than duly patient, or taking a sledgehammer to a not, but... --Soundofmusicals (talk) 13:22, 14 May 2020 (UTC)


 * Trying to be as fair as I can - have a look at the current "Standard forms" section, which I have now tweaked to cover the specific question that raised the purported need for prescriptive rules. The real reason this time rather than "because you're not allowed". --Soundofmusicals (talk) 15:32, 14 May 2020 (UTC)


 * First of all, I was only adding content, not editing or deleting, which is why I acted boldly; otherwise I would've asked on the talk page first. I disagree with many of your points, which I shall now address:
 * I see no problems with combining descriptive and proscriptive, as long as each is in its own section. They are two different approaches to describing the same system. The contrast should provide clarity, not confusion, as long as the two approaches are both separate and consistent with one another, which they are. I don't see this as a WP:NOTHOWTO issue, since it's a detailed description taken from RS, rather than a step-by-step guide to constructing a numeral. it's left up to the reader to apply the rules correctly.
 * But there is a universal set of rules, which are described quite consistently by multiple RS, including a scholarly journal. I believe I have described these rules correctly and thoroughly, so that only one permutation is legal for each value. If you can find any exceptions to this, by all means let me know. And on the contrary, my sources indicate that there was quite a bit of variant usage in classical antiquity, while the standard form only took shape in the past few centuries.
 * Yes, there needs to be jargon, which is why the technical approach needs a basic overview to complement it. I did what I could to make these terms accessible by putting the relevant numerals in parentheses, so that the first mention of 'decimal numerals' is followed by (I, X, C, M) and that of 'quinary numerals' is followed by (V, L, D), in order to clear up confusion. I see no problems with additive/subtractive, the relation of these words to addition/subtraction should be apparent. I also don't see how they clash with the rest of the article, if you can cite an example, by all means do so.
 * The vinculum and especially fractions rules are not required for my ruleset to work, they're merely extensions to the system (going higher and lower). It is possible to combine numerals with fractions, as I've seen in my sources, and the RS frequently refer to vinculums. Not only that, but I've seen a fair amount of use of iterated vinculums, including here on WP. Since the general limit is three sequential repetitions, it seemed logical to apply this to vinculums as well. With three vinculums, this could bring Roman Numerals up to the low trillions (one less than 4 trillion, to be exact) which should be sufficient.
 * I'm willing to negotiate my 'extended rules' on vinculums and fractions, if that's the issue. but the core rules are fine. They're backed by RS, there's no gaps in their logic, they're consistent with the universal convention, they illuminate the topic further, and are reasonably intelligible, especially when used in complementary fashion. They are also within the scope of an encyclopedia, and do not explicitly violate WP:NOTHOWTO. For these reasons, I believe my content is worth adding. Xcalibur (talk) 01:36, 15 May 2020 (UTC)
 * It occurred to me that this would be easier to follow if I linked the relevant content and sources. Thus:
 * Current orthography the proposed section (current link further down)
 * supporting RS (there are other consistent sources out there, such as Lee K. Seitz but I didn't want to overload on citations).
 * Xcalibur (talk) 03:23, 15 May 2020 (UTC)
 * I just made a breakthrough and significantly improved my ruleset! While the old one was logically consistent, it was also somewhat inefficient. Now it's much more efficient, thanks to a rule I borrowed from a source I had glossed over. I may post it on this talk page later. Xcalibur (talk) 04:41, 15 May 2020 (UTC)
 * This has been argued a dozen times before. IMHO any "rules" longer than the basic description is wrong. I think the citations for "rules" might be useful, and your introductory sentence is much better than the one in the current article.
 * Here is an example of simple rules, with the many redundancies in your set removed:
 * 1. Symbols are written in order from highest to lowest and their sum is the value of the number
 * 2. I, X, and C can be written at most once each directly before a symbol with 10x or 5x value (thus out of order), and is subtracted from the sum.
 * 3. The shortest pattern that sums to the correct value is used.
 * Many of the "variants" can be described as relaxing rule 2 toward "any symbol written before a larger one is subtracted", and perhaps not using the shortest pattern.

Spitzak (talk) 19:54, 15 May 2020 (UTC)


 * Your approach is interesting and quite logical. However, explicit rules are supported in Reliable Sources, and I think they're better at answering specific questions (eg why is 99 XCIX, and not IC?). Moreover, while I thought I had fine-tuned it, my recent alteration has been a leap forward. In the earlier discussion, you rightly pointed out that describing a decimal system with more than 10 rules is inefficient. this is because there was a weakness in my ruleset, which I had patched over with a few cumbersome rules (any amount, not redundant, rightmost). In reviewing a source I had overlooked, I found the rule I was missing: addition must be less than subtraction for a given numeral. Thus, I replaced 3 rules with this one, which slid perfectly into place. Now I've got 13 rules, 3 of which are about fractions and vinculums. Ignoring the extended rules, I can now describe the basic system with 10 rules, which is efficient. With this latest advance, there's even more reason to add this, alongside clarity and RS. Xcalibur (talk) 01:44, 17 May 2020 (UTC)


 * Read my last addition (deleted by Spitzak as redundant, with which I actually concur) - it DOES cover the only thing that could be described as "missing" from current text. "Following RS" is all very well - but as I pointed out there are any number of different on-line approaches to RMs - many of which could well qualify as "RS". Forgetting the forums, and those sources that actually make incompatible statements, these different approaches are mostly very much WP:NOTHOWTO - and very properly indeed, since they ARE basically "how to" manuals rather than encyclopedic descriptions. The "trouble" with your rules isn't so much that they are written in technical language that won't be understood by the very people who come to an article like this for information (although that happens to be true). It's NOT that they lay stress things like "Roman fractions" and vinculum which are not part of normal modern use (although both these things have their own sections elsewhere, which, just quietly, you might like to (re-)read). It's NOT that in making the rules more "efficient" (=shorter) they have been robbed of whatever comprehensibility they might have once possessed. It's not even that they interrupt the flow of the article - rather the real problem is that this interruption is unnecessary, unless there is something missing from the description. In which case a tweak or two of the description, which is something you have never suggested, through all your myriad posts to this and previous threads on this topic, would be better than a new section that basically starts again from the beginning and tells the same story in a different way. Even if it were a clearer and/or more comprehensive way, which is patently just not so (and which, to be fair, you have never claimed). --Soundofmusicals (talk) 04:13, 17 May 2020 (UTC)
 * @Spitzak (talk) I've added another tweak to the article itself - expanding the introductory note to the "Standard forms" section with reference to the introduction to Xcalibur's new "rules". Is this helpful, do you feel? Far from essential, in my point of view, if you feel it is "redundant verbiage". --Soundofmusicals (talk) 05:30, 17 May 2020 (UTC)


 * (pardon me for fixing your margins. I think talk page discussions are easier to follow if you only change the margin for each new author). The problem, Soundofmusicals, is that almost none of that is correct. There is, in fact, a single modern convention for Roman Numerals, and the various RS are entirely consistent on this matter. Thus, an accurate description of this convention which paraphrases the RS is a perfectly reasonable addition. it's providing relevant info on the topic, which dispels the "real problem" of it being unnecessary. The rest doesn't really ring true either: it fits right into the article and is fairly short, it's a bit technical but not excessively so, and I don't think there's any serious conflict with WP:NOTHOWTO, at least, no more than there is for the current description. it is in fact an encyclopedic description of the modern universal standard, as given by the RS, especially now that I've filled in the missing rule and taken away the scaffolding, so that it's even more cohesive and efficient. Yes, vinculums/fractions are both extensions of the basic system, but they too are mentioned in the RS -- one of them mentions that fractions should only be added, and most of them mention vinculums. The rules are comprehensive, and they're not meant to be a replacement. Rather, "telling the same story in a different way" is exactly what would benefit the article; the basic intro and the more advanced description complement one another. To reiterate, the ruleset I've offered is quite correct (and more efficient than before), it accurately paraphrases RS, and helps explain the consistent modern standard. And yes, there is a consistent standard, and all the sources agree on that, without any incompatibilities that I've seen, yet for some reason you don't seem to believe this.
 * In light of my significant change, it's pertinent that I link the ruleset as it currently stands: Current orthography (I added then reverted so that I could link it here, I also struck through the earlier link to keep all interested parties on the same page). Xcalibur (talk) 09:25, 17 May 2020 (UTC)
 * It seems you've been helpful after all. One of your major criticisms is that the style is too technical for easy reading; while I downplayed this before, I now concede this is true. I focused too much on formal accuracy, and not enough on ease and clarity. Thus, I've re-written the ruleset in plainer, more direct English. I owe a small debt to both you and Spitzak for offering critiques which spurred me into making significant improvements.
 * Current orthography Much improved, I'm sure you'll agree (I'm using mainspace because my sandbox is a mess, and to show how it fits into the article as a whole). Xcalibur (talk) 00:51, 20 May 2020 (UTC)
 * Your own references somehow manage to restrict themselves to 3 rules, how come you cannot? Almost all the rules are covered by "use the shortest allowed pattern". In any case there should NOT be two descriptions of how to write numbers, if your "rules" are better then you should replace the other text, having two texts just leads readers to try to figure out what the differences between them are. And stop duplicating the intro information about this being modern usage. I do think your intro is much better though.Spitzak (talk) 15:48, 22 May 2020 (UTC)
 * Your own references somehow manage to restrict themselves to 3 rules, how come you cannot? Almost all the rules are covered by "use the shortest allowed pattern". In any case there should NOT be two descriptions of how to write numbers, if your "rules" are better then you should replace the other text, having two texts just leads readers to try to figure out what the differences between them are. And stop duplicating the intro information about this being modern usage. I do think your intro is much better though.Spitzak (talk) 15:48, 22 May 2020 (UTC)


 * That's because my refs tend to compound and summarize rules, while I spell them out explicitly. You'll find that the coverage is not that different, especially the last link (which I should probably turn into a ref). As I said, your approach is logical and effective, but it's not as detailed or as well-supported by RS. I'm not duplicating any text, at least not intentionally. Most importantly, I fundamentally disagree that there has to be only one description -- two different approaches to the same system sheds more light on the subject. I doubt it would cause problems for readers, especially since they're entirely consistent with each other -- one is outlining the general pattern, the other is more in-depth. This seems to be a philosophical disagreement more than anything. Xcalibur (talk) 16:56, 22 May 2020 (UTC)


 * You still need a consensus it you want to make this major change - neither Spitzak nor I (not to mention Johnuniq) agree with you so you are a minority of one. The fact that the sources don't find it necessary to "codify" to this level of detail - rather cuts the ground from under your plea that you are "following sources" better than the rest of us.
 * ANYWAY I am trying hard to use all this to improves the article - and am in the process of using my sandbox to (try to) craft a better article from this mess. (see next thread).


 * Granted, I do need consensus. I only restored the content because I made significant stylistic and structural changes, which resolved a couple major criticisms. I thought this might be enough, so I re-started WP:BRD. I certainly won't attempt to override others, which goes against policy.
 * It really doesn't undercut me. Read closely, and you'll see that my sources cover similar rules-based ground, just with less detail in some cases. My ruleset corresponds very well with two of my best sources, namely Allen Shaw - Note on Roman Numerals & Lee K. Seitz which have a similar degree of detail. So yes, I'm following RS, by weaving together the various sources into a single work. Also relevant is that those two linked sources take the 'dual approach' I recommended, with both rules-based analysis and a simple description of the decimal pattern. I see no reason why this article can't follow the same concept. Xcalibur (talk) 07:00, 23 May 2020 (UTC)

Sound of Musicals's sandbox!
1. Description (introduction).

I have basically reduced the sources to three. All are R.S. - and all refer to the points made by the text to which they are appended. "Sources" I have cut are redundant, or contain major factual errors (one "clearly" states that Roman numerals have no "fractional" forms at all - which is contradicted by a whole section we devote to them! They (the sources) are certainly not more "professional" or "comprehensive" than others we already use elsewhere in the article to make specific points (just have a look at them!)
 * There really ISN'T a "universial standard" - (the point of the Adams source) In fact Adams is not really a totally "reliable" source - he seems to favour the use any form you like! But we now hint that this is probably a minority view ("some writers" standing for Adams and his ilk).


 * The "Hayes' source actually refers directly to the U.S. copyright example (just one reason why using standard "orthography" (pace Adams) makes such good sense!


 * I have kept the best (at least the best looking) of the new references!

2. Heading "Individual decimal places".


 * I actually prefer this for a heading! But it does need a separate note about the "base 10" nature of RNs to make sense of what follows. As I said in a separate edit summary it is possible the paragraph concerned could be better worded.


 * I finally yield to the "1 to 9" count rather than "1 to 10" - although this rankles. Strict "base 10" counting goes from "0 to 9", of course, but RNs don't have a "zero" so "1 to 10" was actually better in this context.

Hopefully Vinculum explanation will suffice - I think some people have actually suggested its use as "completing RNs for modern use - a totally unnecessary exercise!! In any case no need to complicate things here.

PLEASE forgive mad inconsequential thrashing produced by silly sandbox thrashing!

Soundofmusicals (talk)


 * Pardon me for interjecting, but the whole point of the sandbox is to experiment. By all means, act as you'd like within its bounds. Xcalibur (talk) 09:10, 24 May 2020 (UTC)


 * What I appologised for is what I actually did - which was to pile up 20 virtually identical versions of the article in mainspace - an administrator or someone might like to wipe them, I don't know how. No bearing whatever on our current disagreement - just a tired old man making a dog's dinner of the very sound technique of making multiple edits in a sandbox copy of an article and then copying the whole (edited) sandbox back into the article. I've used said technique for years without the least problem. Somehow let myself get rattled - but never mind, the end result is what I intended (I think)!--Soundofmusicals (talk) 14:59, 25 May 2020 (UTC)

Let's actually look at the famous "ruleset"
The following seems to be the latest form of the proposed rulesset:


 * Repeated decimal numerals (I, X, C, M) are added together.
 * Up to three (3) decimal numerals may be repeated in sequence per power (there may be up to four (4) non-sequential repetitions per power).
 * Smaller numerals placed to the right are also added.
 * Smaller decimal numerals placed to the left are subtracted.
 * Quinary numerals (V, L, D) may not be repeated.
 * Quinary numerals may not be subtracted, only added.
 * Only one (1) decimal numeral may be subtracted per power.
 * Subtraction must be by 1/5 or 1/10 (i.e. the next lower decimal numeral).
 * Addition must be less than subtraction for any given numeral.
 * Powers of ten are dealt with separately, ordered from greatest to smallest and from left to right.
 * Fractions may only be added to the right of all numerals (not subtracted).
 * Placing a bar (vinculum) over a numeral multiplies it x1000.
 * The vinculum is preferred for values of 4,000 and greater (it is not used up to 3,999); and may be iterated up to three (3) times.


 * Quite apart from the fact that they add nothing whatever to the article not already explained much better in the description (and explained much less cryptically and in a proper sequence) let's look, purely academically, at how what is wrong with each rule!


 * Repeated decimal numerals (I, X, C, M) are added together.


 * All very well, but what is a "decimal number" in this context? Listing them (as I, X, C, M) is much less helpful than you seem to think. We need, following this kind of approach, to have a "rule" distinguishing the symbols used in Roman numbers as "decimal", and "quinary" (V, L, D). And what does "added together" (or "additive") actually mean? We probably could not explain what actually happens here at all without a few examples (as in I, II, III) - certainly that is the best way of doing so - but that is getting us right back to how we do this in the current "description", isn't it?


 * Up to three (3) decimal numerals may be repeated in sequence per power (there may be up to four (4) non-sequential repetitions per power)..


 * Again, a simple example example, like say "III" is needed here - but then, isn't that what we do in the current description? In fact these first two rules seem to overlap a bit, don't they? Finally - what exactly is "four non-sequential repetitions per power"? Are you referring to the use of IIII for 4? But they ARE sequential, surely? Or perhaps you mean that fact that the RN for a particular digit can be represented by four symbols (as VIII)? But there are still only three "repetitions" here. And what does "power" mean in this context? Again, an explanation of the significance of the powers of 10 needs to come before this rule if it to make any sense at all.


 * Smaller numerals placed to the right are also added.


 * Yet again, what do you mean? Adding a "I" to the right of "II" make "III" or 3? Isn't this a bit redundant with the first two rules? Or are we talking about adding "I" to "V" to make "VI" or 6? Or "I" to "XX" to make "XXI" or 21? Sort of, although not the clearest way to put it. Are we allowed, by this rule, to add "VI" to the right of "XXI" to make "XXIVI", or 27? (the first "I" is added to "XX" and the second "I" to "V". This last is in fact quite a common type of beginner's mistake. You have mentioned "powers" (of what) but not explained that each power (of 10 - this is important) is separate, we don't add or subtract between powers (in this case units and tens).


 * Smaller decimal numerals placed to the left are subtracted.


 * Your intentions here are impeccable - this is your explanation of subtractive notation. But can we subtract "I" from "C" to make 99? (NO, because we would be mixing, or "crossing" powers). And what is wrong with "VC" for 95? You do mention "decimal numbers" and thus by implication disqualify "V", "L" and "D" - but a rule needs to be specific. In fact there are only six permissible subtractive notations in standard RNs "IV" and "IX" in the units, "XL" and "XC" in the tens, and "CD" and "CM" in the hundreds. To explain this in a single "rule" would be difficult. With all respect, you don't come close, even allowing for the other "subtractive" rules in this set. To explain it using examples, as the current article does, in the context of the common template followed by each power of 10, is not only clear, simpler, and more concise - but it heads off user questions about "subtractivity" far more effectively.


 * Quinary numerals (V, L, D) may not be repeated.


 * Logically this must be so. What can "VV" mean but 10 ("X" )- or "DD" than "M"). But much more importantly - a "legal" RN may not include more than one instance of a quinary number, repeated or not (note the real-life example of "MDCDIII" for 1903 on a respected art museum of all places)!


 * Quinary numerals may not be subtracted, only added.


 * Indeed. But surely you have already said so. what is the point of having a rule that "only" decimal numbers can be subtracted if you're going to have another that quinary ones can't? Or is there a third class of RNs that can? (or can't?)


 * Only one (1) decimal numeral may be subtracted per power.


 * This one is very closely connected with your inadequate little rule about subtractive notation. It needs either to go - or be integrated with that rule (or at the very least to follow it! And you STILL haven't explained what power means in the context of RNs (and why).


 * Subtraction must be by 1/5 or 1/10 (i.e. the next lower decimal numeral).


 * Clumsy, cryptic, ambiguous, and its meaning (assuming I have read it correctly) needs to be part of at least one other rule, so it is potentially redundant as well. We already cover this much better in the current article, using examples.


 * Powers of ten are dealt with separately, ordered from greatest to smallest and from left to right.


 * This needs to be rule no.1 - several of the other rules need to refer to it. (hence it coming at the beginning of our existing description).


 * Fractions may only be added to the right of all numerals (not subtracted).


 * This is superfluous nonsense. We don't use Roman fractions at all as part of a "standard, modern RN", so a rule about where they go in a ruleset about standard RNs is plain silly. The Romans themselves only used an inscribed "numeral" fraction to indicate the value of their small change. (see the section on fractions in the article). There is one exception to this, modern pharmacists use "S" (= 6/12 or a half) I'm told. But this is hardly "standard orthography" in anyone's book.


 * Placing a bar (vinculum) over a numeral multiplies it x1000.
 * The vinculum is preferred for values of 4,000 and greater (it is not used up to 3,999); and may be iterated up to three (3) times.


 * But vinculum is NOT part of "standard, modern orthography". Its use by the Romans is speculative - its use in medieval times is largely irrelevant (we have ditched most uniquely medieval forms) - and its use in modern times is hypothetical (we don't need it).


 * Nor, trying hard to be kind here, do we need these rules! --Soundofmusicals (talk) 08:47, 23 May 2020 (UTC)


 * First of all, the rules must be taken together. Many of the 'loopholes' you pointed out are filled by other rules. I'll admit, you did find a chink in my armor -- I didn't specify that quinary numerals may not be repeated non-sequentially, and I'll have to tinker to remove that bit of vagueness. for the record, non-sequential repetition can be seen in a numeral like 39: XXXIX, which has 4 Xs, but no more than 3 in a row. The fractions rule is supported by one of my sources, and the vinculum is widely supported by RS, so it belongs. we do in fact need vinculums if we want to extend coverage beyond 3,999 up to the low trillions. You also didn't comment on the subtraction>addition rule, which is very significant to structural integrity.
 * Aside from the quinary fix, I'd be willing to rearrange and fold rules together. I was thinking myself that perhaps the 'powers of ten' rule could be set at the beginning rather than at the end. To re-iterate, the ruleset and decimal pattern complement one another, as seen in my sources. I'll be back with a revision. Xcalibur (talk) 09:52, 23 May 2020 (UTC)


 * Powers of ten are dealt with separately, ordered from greatest to smallest and from left to right.
 * Repeated decimal numerals (I, X, C, M) are added together, with up to three (3) permitted in sequence (there may be up to four (4) non-sequential repetitions per power).
 * Quinary numerals (V, L, D) may only be added once per power (not repeated nor subtracted).
 * Smaller numerals placed to the right are added, while smaller decimal numerals placed to the left are subtracted.
 * Only one (1) decimal numeral may be subtracted per power, and this must be by 1/5 or 1/10 (i.e. the next lower decimal numeral).
 * Addition must be less than subtraction for any given numeral.
 * Fractions may only be added to the right of all numerals (not subtracted).
 * Placing a bar (vinculum) over a numeral multiplies it x1000.
 * The vinculum is used for values of 4,000 and greater (it is not used up to 3,999), and may be iterated up to three (3) times.
 * Behold, significant renovations have been made. Does this address your concerns? Xcalibur (talk) 10:37, 23 May 2020 (UTC)


 * BUT it doesn't really address any of my "concerns" at all, in fact apart from my mention of the fact that one of the rules ought to come first it doesn't seriously pretend to do anything of the kind.
 * I have repeatedly pointed out (with reasons) why ANY rule about fractions or vinculum is nonsense in this context. To repeat - neither is part of any 'modern, standardised" RN notation. If you are determined to keep the last three rules at least give some kind of reason! Even an example of a "real" modern inscription or document using either would be a step in the right direction.
 * Several of your rules are still misleading or meaningless except in combination with other "rules". If a rule has a "loophole" it needs to be "plugged" as part of the original rule.
 * The other "renovationed" rules are still (in themselves) highly cryptic, if not meaningless. I pointed out that this could be improved, if not fixed, with an example or two.
 * My real "concern" is one I expressed at the beginning, and have repeated innumerable times now - a rules set that is any use at all will actually look very like the description we already have. If the description is satisfactory then we don't need to boil our cabbages again - even if the rules really were "compatible with the description". In fact the more compatible they were, the less they would be needed. --Soundofmusicals (talk) 20:53, 23 May 2020 (UTC)


 * I was thinking of splitting the rules into two parts, 6 basic rules and then an extended section to cover fractions/vinculums. would that be acceptable? either way, I would like to keep the extra rules, given their representation in RS and elsewhere (including WP, there's a doubled vinculum in the infobox for 1,000,000,000). On the contrary, I addressed your concerns quite well, clearing up the quinary rule, compounding rules so that they're less fragmented, and re-writing for ease of understanding. Yes, they must be taken together, which is less of an issue when there's only 6 primary rules. My re-formatting has also brought the ruleset closer in line with the RS; they really should be self-explanatory in their current form. finally, two different angles is exactly what's needed to illuminate the topic, and the 'dual approach' can be seen in my best sources. Xcalibur (talk) 07:47, 24 May 2020 (UTC)

Can we please watch the terminology
Quinary means base 5 so that you count 1, 2, 3, 4, 10, 11, 12, 13, 14, 20. It does not mean 5x10^n Decimal number is also suspect, powers of 10 is better. The whole business of a ruleset is getting far too technical for ordinary readers, and are mathematicians interested? The only reason that I can see for this detailed set is if you are using as the basis for some sort of programming. Martin of Sheffield (talk) 11:29, 24 May 2020 (UTC)


 * I appreciate your feedback. Yes, I understand base numbers well, and I guess my use of 'decimal' and 'quinary' is a misnomer. Maybe I could refer to them as 'fives numerals' and 'tens numerals' instead? I'll admit, my initial approach was overly technical, in fact I originally thought of it in terms of programming logic. However, I also believe the 6 basic rules I've narrowed it down to are reasonably accessible, well-supported, and help answer the various questions that come up on this topic. Xcalibur (talk) 12:02, 24 May 2020 (UTC)

Fractions and "barring" as part of modern, current usage!!!
Have you noticed that we already have separate sections on fractions and vinculum? Read them - and consider adding anything that we don't already cover to your satisfaction there, rather than under a heading that has nothing to do with either. But please cite any changes - which doesn't mean vague references to "RS" that on examination prove to either to have nothing to do with the case, actually lie closer to the current article, or treat the subject in a puerile or erratic way that is very far from being "reliable" anyway. There is an awful lot of contradictory nonsense on the web on this subject. In the meantime why not just cut these so-called 'advanced rules" out altogether - however important they might be to the topic as a whole they have nothing to do with "current orthography" because - and I am getting a little fed up repeating this, they do NOT form a part of the Roman numerals in current use. Just saying that an unnamed "RS" mentions them is no argument - so do we, in the right place. --Soundofmusicals (talk) 09:56, 24 May 2020 (UTC)


 * Yes, but I see no reason we can't discuss fractions/vinculums in the rules section while fleshing them out further down. I'd be willing to consider paring down the rules to the basic pattern, although I'd rather have more content than less. I cited my sources above [Duplicated set of "refs" cut here] in fact, there is a universal modern standard for RNs, and this is supported by a consensus among Reliable Sources, referenced above. You'll note that they refer to vinculums and fractions, and are quite consistent with my rules, even moreso now that I've re-formatted. I suppose I could dig for even more sources, but these should be sufficient (I don't want to overcite). my proposed rule section is well-sourced, logically sound, cohesive, adds something original to the article, is not overly long, and is better than ever thanks to constructive criticism. I'm not sure why you're still in opposition. Xcalibur (talk) 10:43, 24 May 2020 (UTC)


 * Can you please refer to the text on any of those websites that supports (or in fact relates to) any specific point you have been making here. One of them actually contradicts your "fractions" argument, stating that the Romans didn't have notation for fractions at all (well, they did, but the only examples we have are on small value coins). And to repeat - aspects of Roman numerals that are no part of current usage (and neither Roman fractions nor vinculum are in current use) don't belong in a description (whether formatted as "rules" or not). Exactly what part of that don't you get? --Soundofmusicals (talk) 13:45, 24 May 2020 (UTC)


 * it's right there in plain sight. even if you don't have JSTOR access, you should be able to see the first page of the scholarly journal. None of my sources contradict fractions; I believe you're referring to the MathForum section on Calculation, specifically the first paragraph which states that the Romans didn't have standard means of writing fractions using their numerals (as we do). That's true, they used the separate system of unciae and semi, which is explained by the next two paragraphs! it seems like you skimmed over without reading it properly. as for your point, I get that, but I'm not sure if it's correct to rule out fractions/vinculums from modern use, especially since my sources discuss them, and I've seen references to iterated vinculums, and you mentioned pharmaceutical use of S. Xcalibur (talk) 14:13, 24 May 2020 (UTC)


 * Let me address the point more directly. How can you be sure that fractions/vinculums are not part of modern orthography? my sources mention both, especially vinculums. with that said, if that's the sticking point, I'd be willing to put up the basic rules without the extended, at least for the time being. Xcalibur (talk) 05:18, 25 May 2020 (UTC)


 * If they were "part of modern orthography" then we'd still be using them - which we're not (so they aren't). Otherwise, how exactly do you define "modern orthography"? If you want to concoct a "set of rules" (on any subject) it needs NOT to stray from its topic. What about a further "rule 1." stating "Roman numerals are no longer the principal way of indicating numeric values - but remain in use in some specialised contexts for integers between 1 and 4000". --Soundofmusicals (talk) 10:13, 25 May 2020 (UTC)


 * that 'rule 1' is background/context rather than a rule. in any case, if the sticking point is fractions/vinculums, I could leave those out, and just put up the basic 6 rules. Do you assent? Xcalibur (talk) 14:00, 25 May 2020 (UTC)


 * Assent to what? Assent to your apparent recognition that fractions and the use of vinculum have nothing to do with the case? Definitely. In which case by all means have a look at our current write up on these (separate and discrete) subjects (neither of which has anything to do with "current orthography") and improve them if you find them inadequate. --Soundofmusicals (talk) 14:47, 25 May 2020 (UTC)


 * Perhaps we've come to an agreement? I've restored the section with basic rules, leaving out extended rules and referring the reader to the sections on fractions/vinculums instead. is this acceptable? Xcalibur (talk) 15:07, 25 May 2020 (UTC)

"Context" for rules
I think what is actually "wrong" with those rules of is that they do not make sense at all without "context". Add enough context to make them clear and meaningful and you'd actually have something very like the current text, only ten times as long and not one half as comprehensible. But no - unless you can get "assent" from a consensus of other users that your set of rules IS an improvement, then no. That's how it works. --Soundofmusicals (talk)


 * The context is provided by the article, so that my rules complement the basic description (and again, the 'dual approach' is used by my best sources). For a moment, I thought we had an agreement, your continued objection is rather unfortunate.Xcalibur (talk) 15:35, 25 May 2020 (UTC)


 * Of course by "context" I meant specific context, such as supplied by the introductory sentences to your new section. --Soundofmusicals (talk) 17:15, 25 May 2020 (UTC)

For reference, this is my latest draft: (comments by Soundofmusicals in italics)
 * Modern Orthography
 * The modern era has seen the emergence of a standardized orthography for roman numerals, which permits only one permutation for any given value. While exceptions can be made (notably IIII instead of IV on clockfaces), the modern convention is widely recognized and adhered to, and may be described by the following ruleset:
 * This bit really IS redundant - in fact if you had noticed I actually incorporated the relevant part of it into the introductory remarks to the description, so it is now a permanent part of the article!
 * Powers of ten are dealt with separately, ordered from greatest to smallest and from left to right.
 * How about "A number containing several decimal digits is represented by the notation for each digit, from the highest to the lowest power" oops - that's what the current article already says!
 * Repeated 'tens' numerals (I, X, C, M) are added together, with up to three (3) permitted in sequence (there may be up to four (4) non-sequential repetitions per power).
 * The objection here was to the "jargon" use of "decimal". Coining a new jargon term ("tens") doesn't really work either. Better might be something like "The numerals governing each power, (I, X, C, M) may be repeated up to three times (but no more) - within the part of the numeral corresponding to a single digit". But the effect of this is made much more clearly in the description, the difference being merely in the fact that we go by line, rather than by column.
 * 'Fives' numerals (V, L, D) may only be added once per power (not repeated nor subtracted).
 * As was pointed out, "quinary" really has another meaning altogether, but coining a new jargon word yourself doesn't really work here. The fact is that no "standard" Roman numeral can logically contain more than one example of any of the "quinary" numbers. Is this rule therefore redundant? Or might we restate it as something like "None of the symbols for the numbers representing the powers of 5 (V, L, D) are ever repeated within a conventionally expressed Roman numeral, nor are they ever used subtractively." still rather silly really.
 * Smaller numerals placed to the right are added, while smaller decimal numerals placed to the left are subtracted.
 * Smaller than what, and to the right/left of what, and when? Totally and comprehensively nonsensical - at least without context or an example or two to show what you are talking about
 * Only one (1) decimal numeral may be subtracted per power, and this must be by 1/5 or 1/10 (i.e. the next lower decimal numeral).
 * "Standard" subtractive notation is limited to powers of 4 (IV, XL, CD) and 9 (IX, XC, CM) - why not just say that?
 * Addition must be less than subtraction for any given numeral.
 * What exactly does this mean? I'm serious, it's very late/early here but I honestly don't know quite what you're talking about - an example or two - clearer if slightly "windier" prose?
 * These rules describe the basic pattern from 1 - 3,999. The system can be extended by employing fractions and vinculums, which are detailed in later sections.
 * I really thought we had agreed on that point. Any attempt to append a fraction to a Roman numeral, or express a number over 3999 may be wonderful fun, but it goes into quite another realm than "modern standard orthography"!


 * This is the result of a fair amount of fine-tuning and constructive criticism, and in its current form would make a useful addition. the 'dual approach' of describing the same system two different ways is quite useful, and it specifically answers the many queries that come up on this topic.
 * Alas, no - the whole notion is basically flawed - and, even more importantly, unnecessary - although no one is saying the current text is perfect and might not be improved it does the same job, and at least for the time being much better! --Soundofmusicals (talk) 18:54, 25 May 2020 (UTC)


 * I provided examples in the other thread. I'm not sure if 'powers of 5' or 'powers of 4 or 9' is correct. jargon seems to be the best way to describe the relevant numerals; 'fives' and 'tens' are intuitive and accurate. as for the right/left rule being nonsensical, it makes perfect sense in the context of the 'powers of ten' rule. subtraction can be described either way, my way is more formal and backed by sources. as for subtract>add, is that not self-explanatory? IXX is wrong because X is greater than I, XIX is correct because nothing is added to the rightmost X; IXI is wrong because then addition and subtraction are equal; MCMD is illegal because the addition is larger. it's not flawed, it's perfectly sound, and backed by RS. I still think extended rules for vinculums/fractions are worthwhile (as they're mentioned alongside other rules in the RS), but I'm willing to compromise on that point. Xcalibur (talk) 20:19, 25 May 2020 (UTC)


 * If you're going to use jargon at all (rather than just spell out what you mean) at least use a term someone else has used - making up your own jargon is patent OR however "intuitive" and "accurate" its inventor may feel it is. And anyway you reverted to "decimal" in the following rules! Have a look again at that rule I called "nonsensical". The reason that IXX is "wrong" is because it breaks the "separate notation for each digit, and in the normal order" rule. "I" is always a "unit" value - "X" only has a unit value when preceded by a "I" - as in IX, which means 9. So IXX looks like an attempt to write 19 - (either 1 subtracted from 20, just not the way it's done, or nine and ten, which is out of order - tens always go before units. A further problem is that it is an occasional ancient/medieval variant of 21! Lets not get into that one here. XIX is the correct way to write 19 first the "tens" (X) then the units (IX). All nothing to do with the strange mess you'd made of that rule - sorry but calling a spade a spade for once! MCMD is nonsense because you already have a complete "hundreds" notation in CM (900). Adding another hundreds numeral D (500) doesn't make any sense at all - is it an error for "L"? or is it a fanciful way of writing MMCD. Again - no obvious connection with your "rule". This isn't really quite what I meant by "flawed" however. The very idea of a set of rules for Roman numerals that will have a one for one correspondence with every possible "mistake" a clumsy user might make is the "flawed" idea. On the one hand no one really designed the system - it grew like Topsy, and on the other hand as such things go it's just not that complicated anyway - get hold of one or two basic ideas and everything else follows without any real problems.  --Soundofmusicals (talk) 23:02, 25 May 2020 (UTC)


 * You're right, I didn't change every use of 'decimal', thanks for catching that. I don't see a problem with using different terms, especially when they're similar to a term the RS use, namely 'powers of ten'. in any case, I could work around that if needed. as for the rest, I'm not sure if you're following things correctly. there's nothing flawed about the concept of rules that answer every possible error, because it's quite possible in a limited system such as this, and I believe (with the help of sources) I've pulled it off. Xcalibur (talk) 12:21, 26 May 2020 (UTC)

Latest reversion
I thought we had mutual "assent" to the principle of WP:BRD. We need a change of consensus here - not a "wearing down" process. --Soundofmusicals (talk) 15:33, 25 May 2020 (UTC)
 * Of course I accept BRD. I thought you agreed that the section could go up, so long as fractions/vinculums were left out. I took you the wrong way. I'm not attempt to 'wear down' anyone; in fact, I've made numerous fixes and improvements in response to constructive criticism. it's my hope that I can achieve a consensus on this, but you seem to have a basic philosophical disagreement with my content. Xcalibur (talk) 15:38, 25 May 2020 (UTC)
 * Could one of the other parties who seem to be watching this please chime in here with a suggestion about how this dispute might be resolved - since the two of us seem to have stopped at the stage of wishing the other one would go away? --Soundofmusicals (talk) 16:31, 25 May 2020 (UTC)
 * @Xcalibur Please, once more, just read and assess the current "section we have instead of rules" - see what we say, and list your objections to it. In what way(s) is it inadequate? Don't just say "it is a different approach" again, because in fact that really isn't so. The emphasis might be different but the basic ground covered is the same. We DON'T need to do it twice, even in an encyclopedia. The current text "arose" (if you have the patience to get a really early form of this article you can check this) out of something much closer to a "set of rules" like yours. If you like, it REMAINS a set of rules - only more in the form of a narrative than a list. Narratives are very widely favoured in Wikipedia and lists often deprecated (although I don't like generalised arguments like this as a rule I think this is one case...). Nonetheless I have always been very open to the idea of replacing what we have with something better. A number of people (well two or three anyway) have pointed out things that they think are wrong with your "rules"; in spite of what you say you have made no real attempt to meet this "constructive criticism", which would have required a complete rethink rather than a little tinkering at the edges. On the other hand nobody (even you) has had any real problems with what we have, although (as always) it has been subject to repeated tweaking over the years, and by no means just by me. Might we try to rewrite the current description to be a bit more prescriptive?  I may end up doing something that is (quite rightly) not encouraged as a rule, and send a "round robin" request for help around other people who have edited the article over the years, but we do need to get this silly fuss behind us both - the article is the thing! --Soundofmusicals (talk) 16:31, 25 May 2020 (UTC)


 * The article is the thing, that's why I'm trying to add something of value to it. to reiterate, it shouldn't be just one or the other -- there should be both a descriptive and prescriptive section, side by side. they are two paths that lead to the same place, complementing one another and providing a better explanation than any single method. And again, my best sources use the 'dual approach', laying down rules, and then demonstrating the decimal pattern, and I think this article should do the same.
 * if you'd like, I could tinker with my ruleset even further by providing examples as to which permutations are legal and illegal according to a rule, e.g. XXXIX is legal but XXXX is not; XCIX is legal, but IIX & IC are not, etc. Xcalibur (talk) 16:49, 25 May 2020 (UTC)
 * Not sure quite what you think you're talking about here. "XXXIX" is perfectly "legal" - it means 39. Two powers or decimal places here "XXX" = 30 (tens) and "IX" = 9 (units). Nothing whatever to do with "consecutive iterations of a symbol. "XXXX" is not "illegal" - there are no "laws" in that sense. It unambiguously means 40 - but we'd normally write "XL" - at least since before the construction of the colloseum. By all means give us yet another version with the odd example if you must (but do check they are "correct"! - sounds as if they might provide the "context" to make some of the more cryptic rules understandable anyway. How far down that road before you're left with something like the current description? But seriously, where are those "unanswered questions", what exactly are they, and do they form a finite set. I just want to see a criticism (constructive if possible) of the current article in those kinds of terms. Does the article even need reorganisation into a sort of FAQ (not in favour of that myself, really). --Soundofmusicals (talk) 17:11, 25 May 2020 (UTC)


 * Of course XXXIX is legal. XXXX is not, it's a variant that breaks the rules I've thoroughly sourced. the unanswered questions can be found strewn across this talk page and its archives. and yes, my ruleset is complete and logical, without any defects, and functions more efficiently than before.
 * one of the major complaints is that it's too technical/theoretical and difficult for beginners. hopefully, by combining it with examples, this final objection can be resolved.
 * Powers of ten are dealt with separately, ordered from greatest to smallest and from left to right.
 * so that MCMXCIX and CIV are legal, while IC and MXM are not.
 * Repeated 'tens' numerals (I, X, C, M) are added together, with up to three (3) permitted in sequence (there may be up to four (4) non-sequential repetitions per power).
 * so that XXXIX is legal, while XXXX is not.
 * ''but surely the number 39 has two powers - units and tens?? so how is this an example of "(4) non-sequential repetitions per power)"? That was the point of my last comment on this one, although it may have been a little unclear.
 * 'Fives' numerals (V, L, D) may only be added once per power (not repeated nor subtracted).
 * so that XLV and MDCCC are legal, while VL and MDCD are not.
 * Although "VL" is hardly a good example - since it has the same value a LV anyway.
 * Smaller numerals placed to the right are added, while smaller decimal numerals placed to the left are subtracted.
 * this should be self-explanatory. XVI is 16, IX is 9, MDCCLXXVI is 1776, MCMXCIX is 1999, etc.
 * Only one (1) decimal numeral may be subtracted per power, and this must be by 1/5 or 1/10 (i.e. the next lower decimal numeral).
 * so that IX, XLV, and MCMXCIX are legal, while IIX, VL, and IC are not.
 * Addition must be less than subtraction for any given numeral.
 * so that XIV, XLVII, and XCIX are legal, while IXV and IXI are not.
 * how's this for a first draft? Xcalibur (talk) 18:17, 25 May 2020 (UTC)


 * The examples are certainly an improvement! In fact - why not go just a little further and explain each example? getting more and more like a version of the current text? (only still LESS comprehensively and far less coherently) or am I imagining things! Get over the "legal" bit - no law has ever been passed about Roman numerals (not even a proclamation by a Roman emperor, so far as I know). Correctness, or at least following the convention, is desirable, and a description of the convention at some stage of the article is necessary, but this only needs to be done once!!! (Really) Saying exactly the same thing over again but in a different order and in other words isn't "another approach" - it's repetitious and redundant. --Soundofmusicals (talk) 19:13, 25 May 2020 (UTC)


 * BTW - have a look at my interleaved comments on the previous thread --Soundofmusicals (talk) 19:31, 25 May 2020 (UTC)


 * I can't explain each example, because detailed instruction would violate WP:NOTHOWTO. yes, there is a convention, but there are two different approaches: basic decimal pattern, and explicit rules. again, my best sources use the dual method of description and prescription, and I think that would work fine for the article. Xcalibur (talk) 19:44, 25 May 2020 (UTC)


 * There are topics where WP:NOTHOWTO represents a dangerous pitfall for an encyclopedia article - there are others where it is less of an issue. I don't honestly think you can avoid a certain degree of "instruction" with a topic like this one, although this remains a reason we prefer description and explanation (NOT in itself "instructional") with format rules which ARE, however you do them. Have a look, as I said, at my "interleaved" comments on the last thread (although I hadn't at that stage looked at any of your examples - some of which do clarify things a bit). I honestly don't see anywhere where the existing text (current article) is insufficiently 'explicit' - can you give an example of a place where it could be taken more than one way, or where anything is implied rather than spelled out? Why do you apparently need three rules to cover the very simple (and, in modern orthography especially, very limited) concept of "subtractivity"? We have said (preferably as rule 1. since all else depends on it) that the notation for a given digit ("place" or "power") - is separate and discrete. "IC" for 99 cuts across two "powers" and thus clearly breaks this rule, quite apart from any subtractivity consideration. There are just 6 meaningful subtractive numbers in modern notation - why not just say so? At least, wouldn't it be clearer if all the "subtractive rules" were integrated into one, in order to clearly make this point? This is in itself very useful background indeed for a little later in the article, when we go into the many (mostly but not entirely) historical NON-standard forms. Which, as I'm sure you would agree, is something else that the article needs to do. The "questions asked" about Roman numerals are almost incredibly varied - no "set of rules" (and no logical description either) could possibly forestall them all. The real answer to "why can't I..." may well be "well actually you can, but if you want your numeral to convey a particular number in a format people will recognise (rather than just look impressive) you had better stick to the standard forms - now this is the standard form for the number you are asking about, and this is how it is derived". To a question like "What is wrong with IIMM for 1998? - the answer is "nothing whatever, but MCMXCVIII is very much more usual, because the standard form is "M CM XC VIII" (without the gaps - the Romans didn't even put them between words). Each arabic digit has its own numeral". If we get (as we probably will) a supplementary question "But my way is shorter and I read it somewhere" then what can you do? All a bit like this discussion. Round in circles and never a resolution. --Soundofmusicals (talk) 21:07, 25 May 2020 (UTC)


 * I generally agree, and WP policies should always be applied with some flexibility. however, I had WP:NOTHOWTO brought up to me merely for proposing rules, which is an excessive interpretation. I think a rules section is fine, but a step-by-step guide to constructing RN would indeed go too far.
 * the 'powers of ten' rule is supposed to be redundant and intersect with the rest. again, there are two different ways of explaining this, which are complementary. yes, you can describe subtraction as 4s and 9s only, but that's only describing the decimal pattern, it's not explaining why that is the case. my rules make matters explicit, by limiting subtraction to a single decimal value that is 1/5 or 1/10, which coincides with the decimal pattern of subtractive 4s and 9s. we can have it both ways, and we should. on the contrary, my ruleset can indeed forestall any possible question about proper form, ending all circular discussions; it fills in all gaps with a coherent, logical system, which is why it's worthwhile, alongside a simpler overview of the pattern.
 * as for this discussion however, it seems you have a fundamental disagreement with a rules-based treatment of this subject, in spite of the RS supporting it, and various other benefits. you've denied that there's a universal convention for orthography, even though there demonstrably is. I think this discussion keeps going because we're meandering around, and not getting to the crux of why you object to my ruleset, in spite of it being well-sourced, well-constructed, and an original addition to the current article. Xcalibur (talk) 22:32, 25 May 2020 (UTC)
 * Citing facts by reference to good sources is one thing. Drawing from another source for something like a basic approach (say a heavily didactic one) that may well be appropriate to the source but NOT here is quite another. Among innumerable other pitfalls we could even run into copyright problems. Even good sources on a subject like this will vary tremendously. And please read this extract from the article:


 * There has never been an officially "binding", or universally accepted standard for Roman numerals. Usage in ancient Rome varied greatly and became thoroughly chaotic in medieval times. Even the post-renaissance restoration of a largely "classical" notation has failed to produce total consistency: variant forms are even defended by some modern writers as offering improved "flexibility". On the other hand, especially where a Roman numeral is considered a legally binding expression of a number, as in U.S. Copyright law (where an "incorrect" or ambiguous numeral may invalidate a copyright claim, or affect the termination date of the copyright period) it is usually desirable to strictly follow the usual modern standardized orthography which permits only one equivalent Roman numeral for any given value.
 * Do you really have an issue with this? We then go one, of course, to describe the usual standard forms, But "the existence of a universal convention for orthography" is still highly deniable - there really is no such thing - no law has been passed, no panel of experts has sat, it is all based on observation of what forms have usually been followed in relatively recent times. As for your praise of your wonderful set of rules - I have to say no. If one rule actually explins 90% of errors then it should be allowed to do so. If there remains anything else to be said - then say it. Then have the grace to finish. --Soundofmusicals (talk)


 * I disagree with much of this. there's nothing inappropriate about a didactic section, especially in combination with a basic description. I see no issues with copyright, especially when I'm paraphrasing multiple sources. And no, the sources don't vary, they're all completely consistent in describing a single universal standard. yes, there is one correct, legal way, as shown by the RS, and this is not deniable. the fact that there's been plenty of variation, especially in antiquity, doesn't take anything away from this. Xcalibur (talk) 12:36, 26 May 2020 (UTC)

Apostrophus
The Apostrophus table suggests 1500 is CIƆƆ while the Westerkerk picture seems to use CIƆIƆ (plus CXXX to give 1630) 17:58, 3 February 2020 (UTC)  — Preceding unsigned comment added by 165.225.81.40 (talk)


 * I wouldn't worry about this myself. --Soundofmusicals (talk) 23:25, 26 February 2020 (UTC)


 * This was a very poor answer of mine to a perfectly logical question! I have addressed the difficulty by changing the caption to the illustration, anyway. --Soundofmusicals (talk) 11:27, 31 May 2020 (UTC)

Scrap the rules?
Rulsets are a very 21C idea and we are attempting to force a 2 or 3 thousand-year old system into a modern mathematicians view of the universe. Would it not be much simpler to simply do the following and forget rules altogether? There are after all only a few possible components, not an indefinite number that readers need to construct.


 * it's more of a 20th century idea. the graph table is an interesting, efficient approach to showing the patterns. however, there's no need to scrap the rules, especially when they're very consistent and backed by RS. rather, we can discuss the same system in different ways, showing the basic pattern and the internal logic side by side. That's what my best sources do in fact, and I think it would work here also. Xcalibur (talk) 13:13, 26 May 2020 (UTC)
 * What graph? I just suggested the table above. Martin of Sheffield (talk) 13:25, 26 May 2020 (UTC)
 * I meant table, pardon my verbal slip. Xcalibur (talk) 13:33, 26 May 2020 (UTC)
 * This table is a *vast* improvement over the "rules". I still object to any "rules" where it takes more rules than there are actual patterns. Again even the references for the "rules" only have 3 rules each. About 3/4 of your "rules" can be replaced with "use the shortest pattern". But in the end, I do not want to see the system described twice. If "rules" are better then you must replace the current description with your "rules". If you don't think this is a good idea, then it means your "rules" are not a good idea.Spitzak (talk) 18:55, 26 May 2020 (UTC)


 * My references have more than 3 rules, they tend to compound them, even the NMSU source has 5 rules (not counting vinculums). my current ruleset has 7 rules, which is quite efficient. No, I don't have to replace the description, and that doesn't mean they're not a good idea. it's an alternate approach, which complements the basic description. again, my sources offer both rules and a basic description side-by-side, I see no reason why we shouldn't do this. I've also reworked the intro to tell the reader that the system can be described both ways consistently. Xcalibur (talk) 20:16, 26 May 2020 (UTC)


 * I agree totally about this table - thank you Martin of Sheffield!! it's a very neat way of expressing the real "information" conveyed by the proposed rules - and what's more consistent with the way we illustrate similar information elsewhere in the article. Never mind "replacing the rules" with it (since there are problems with the rules that Xcalibur continues to refuse to address) But it could well form a part of a simplified "description". Exactly where it would fit in, and whether it might be even better with the pattern of symbols running across ("landscape") rather than up and down, are the things we'd have to thrash out - in fact that's where this discussion needs to go. --Soundofmusicals (talk) 08:18, 28 May 2020 (UTC)


 * Try the above. Personally I would go for the tall rather than wide form, a lot of people use things like mobile phones to view WP and width can be an issue.  Ultimately though it's a personal choice. Martin of Sheffield (talk) 09:09, 28 May 2020 (UTC)


 * How about THIS - on second thoughts I prefer the columns, but in the ordinary "arithmetic order" (powers descending). With a suitable little note this just about covers all bases! I have also cut the "additive powers of 4" - these are no longer "conventional" in any real sense - only persisting (and then only in the form of ) on clocks and a few old coins (which we cover later anyway). Makes it just a little bit simpler. --Soundofmusicals (talk) 09:51, 28 May 2020 (UTC)
 * I was always in two minds about the alternative 4s. They are seen (clocks, Admiralty Arch) but maybe are best handled under "Variant forms".  I can see why you might prefer "arithmetic order", personally I think "complexity order" is easier to understand, but these are fine distinctions that a bit of concensus can decide. Martin of Sheffield (talk) 09:58, 28 May 2020 (UTC)


 * Yes, they are "variant forms" and belong there. Point is that Excalibur or Bigdan or whatever he calls himself has (rightly) underlined the importance of a simple, straightforward account of "current practice" - so we really don't want any "except when it isn't" stuff, at least not at this juncture. Off to my sandbox to work out a rewritten "description" section incorporating the table! I'd love to put this one to bed for good...--Soundofmusicals (talk) 10:19, 28 May 2020 (UTC)

"References"
What I "removed from a talk page" was not anything anyone said, but a list of "sources" not linked to anything specific - at least some, if not all of them are actually cited in the article. I suppose their removal was not actually "necessary" - but it cleaned up the mess a little and did no harm whatever. Never mind - let them stay if you think they do any good. Incidentally I have had a good old read of the talk page guidelines - might be a salutary exercise for you, too. --Soundofmusicals (talk) 19:43, 31 May 2020 (UTC)
 * the content I'm trying to add to the article is directly compiled from those sources. in any case, it's better not to make edits to previous discussion unless needed, since it can be confusing for anyone reading the backlog. the times I've done so, I added strikethroughs without deleting, and added notes (because I linked to a newer rendition below). or if I alter my text just after posting, I'll usually re-sign. Xcalibur (talk) 20:18, 31 May 2020 (UTC)

Rewritten description
I have been bold and rewritten our description section - formatting it as a set of rules! (this is what we were trying to goad Xcalibur into doing with his, i.e. the point of our "constructive criticism"). As I also repeatedly requested, it replaces, rather than duplicating, our previous "descriptive" text. As more than one of us has pointed out - no reason on earth to boil our cabbages twice. I have incorporating a version of Martin of Sheffield's table, which is neater than what we had before. I have reduced some of the redundancies that were bothering Spitzak - and finally I hope that Johnuniq will forgive any tendency towards undue didacticism, although a certain amount of this is inevitable with a topic like this one.

In fact one might describe this edit as a "reconciliation version". At least I hope it doesn't tread on any toes - and that the article (which is a frequently consulted, quoted, and even cited one) remains a good introduction to the subject. --Soundofmusicals (talk) 10:12, 29 May 2020 (UTC)


 * What sort of mutation have you done here? there's no reason to apply a 'rules' format without the actual rules. combining the structure of rules with the 'pattern description' is rather odd and pointless. the whole purpose of my rules section was to offer a second approach to understanding the system, using logical rules which, in character and substance, are quite different from the 'basic description' while leading to the same place. We should either have just the basic description, or that + a ruleset. and on this particular topic, I do think two approaches side-by-side is what's needed, and readers would understand just fine. I went in and fixed it, btw. Xcalibur (talk) 13:48, 29 May 2020 (UTC)


 * @Xcalibur Consensus (in this case plain "commonsensus" :)) says we need to keep our approach as plain and simple, and yes as "logical" as ever we can. The new "description" has been changed in light of the very rules you have been advocating for so long, and deliberately and systematically subsumes their "logic". I was, in fact concerned that you might have felt that after being so critical of your work for so long I had unfairly "stolen" your ideas!
 * Clearly, a "logical" system of rules must be based on the shape and form of the "pattern" and a definition of the pattern implies rules, whether these are formatted as such or not. Otherwise they would not "lead to the same place".
 * If Roman numerals were as complicated as many websites (and even printed material) imply, then they would never have survived as long as they did as the normal way people (most of them not otherwise literate) wrote numbers. If they were so fundamentally different to our present system then we might very well still be using them! As it was the changeover took centuries.
 * They are in fact a simple system - the difference between them and the system that replaced them (for normal purposes, anyway) boils down to the use of multiple (as opposed to single) symbols to represent a decimal digit, and the absence of a symbol representing a "place-keeping zero". Thus (for instance) the "Arabic" number "8" is represented by a set of 4 symbols - either "VIII", "LXXX", or "DCCC", depending on whether we are referring to 8, 80, or 800. I think we have probably hit on the simplest, most comprehensive way of expressing this notion by using the new table proposed by Martin of Sheffield.
 * The only other rule we need to be able to read any "conventional" Roman numeral, or express any integer between 1 and 3,999 in "ordinary" Roman numerals, is the one you define as:
 * Powers of ten are dealt with separately, ordered from greatest to smallest and from left to right.
 * This then is our other basic rule. Between them these two cover ALL the bases. Anything else will be supplementary - or a "correction of a common error".
 * For instance we do need to have a rule defining "subtractive" notation, if only so the term is not new to a reader of our "Variant forms" section. Really, a simple statement about "IV" and "IX" and their cognates in the tens and hundreds should be plenty. The "subtractive problem" is not caused by inadequate rules, but the false impression given by many writers that "modern orthography" (as opposed to the sometimes somewhat chaotic historical practice outlined under "Variant forms") allows ANY smaller number to precede (and be subtracted from) ANY larger number. I am a little uncertain about this rule - I have actually reworded it a little.
 * The new rule 6 brings the process of reading and composing a standard Roman numeral into focus - defining the operation of the table and including the old set of examples.
 * I have taken the liberty of restoring the numbers of the rules - this is something you might have done for your set: makes it much easier to refer to a particular rule - and I have also selected another section heading. I do hope you can (eventually?) become reconciled to this version. --Soundofmusicals (talk) 22:15, 29 May 2020 (UTC)


 * I understand all you said about RNs. normally, it's reasonable to seek a compromise, but it doesn't work in this context. there's no reason to format a basic description as rules, it should be straightforward prose. in fact, rules were never the point of this, the point is to describe Roman Numerals in a rigorous, mathematical way, it just happens that rules are the best way to do so. a basic intro to the pattern and a logical analysis are two different methods of describing the system, they really can't be combined effectively, and work best as complements. if my ruleset is added, it should only be after a basic decimal description, such as we already have. to reiterate, I don't want rules, I want a mathematical treatment, which needs to be formatted as rules. this is only making a mess of things and misrepresenting my entire concept. with that said, I appreciate the suggestion to use numbers, that can be another improvement. Xcalibur (talk) 22:47, 29 May 2020 (UTC)


 * In what specific way are any of the rules you propose actually more "rigourous" or "mathematical" (or in any other way more useful) than the section as rewritten? What new information do they actually add? Remember that the section is about what you have called "modern standard orthography" - i.e. what is in current use.
 * This is basically the very first question I asked when you first brought this up (quite q while ago now).
 * Nonetheless I admit the word "rule" is perhaps not what is required in this context - so I have eliminated it from the section altogether. I have restored the numbers, however, as they make it easier to reference an individual subsection without adding a series of fiddly headings. -Soundofmusicals (talk) 02:27, 30 May 2020 (UTC)


 * the purpose of my content is to describe the logic underpinning how RNs are expressed, in addition to the basic pattern. I've compiled this from RS, which have the dual approach of rules and basic pattern (notably Shaw and Seitz). these are two different ways of understanding the system, which complement each other and provide greater insight. I only chose a rules format because that's the best way to express the inner logic; this does not apply to the basic pattern, which should be prose + examples. for you to give the basic description a rules-based structure, with numbered paragraphs, is almost a cargo-cultist parody of what I'm doing here, no offense. to reiterate, the 'basic description' and 'inner logic' approaches are different in substance and work best as complements, as shown by RS. I hope that clears this up. Xcalibur (talk) 19:13, 30 May 2020 (UTC)


 * to summarize even further: I want to add content currently missing from the article, namely the logic behind the standard form of RN. this is based on RS, and should be added in addition to the basic description already there. You'll see the 'dual approach' of basic pattern and rules-based logic in the sources I've compiled. Xcalibur (talk) 01:51, 4 June 2020 (UTC)