Talk:IPv6 packet/Archive 1

Review
Please shout at me in a week if this isn't done. -- Eraserhead1 &lt;talk&gt; 23:28, 9 June 2010 (UTC)
 * count on it —— Da n do r iD (talk) 07:40, 10 June 2010 (UTC)
 * OK here's the review.


 * 1) References: I don't think there are enough references for B class - I can add a series of  tags if you want.
 * 2) Covers the topic: To the best of my knowledge yes.
 * 3) Defined structure: Yes.
 * 4) The article is reasonably well written: Yes.
 * 5) Some supporting materials would be nice like illustrations or an infobox.
 * 6) I'm not convinced the article is accessible enough - though it is a technical topic so that is difficult, see Make technical articles understandable.

-- Eraserhead1 &lt;talk&gt; 09:39, 13 June 2010 (UTC)
 * Thanks for the review, Eraserhead1. If it is not too much trouble you may add the  tags, or indicate here what you would like to have referenced. I can think of a suitable illustration, but don't know how to create it; I've seen a lot of SVG images, is that the standard?. —— Da n do r iD (talk) 18:44, 13 June 2010 (UTC)
 * SVG is the "standard" though personally I think its ridiculous (as its not well supported) and that PNG has significant benefits so I wouldn't worry about which you use. -- Eraserhead1 &lt;talk&gt; 20:01, 13 June 2010 (UTC)
 * I've added the citation needed's. -- Eraserhead1 &lt;talk&gt; 20:11, 13 June 2010 (UTC)


 * I am not satisfied with the second paragraph of the lead section. An IP(v6) packet may just as well be encapsulated within another network layer packet, such as AH or ESP when used in tunnel mode, or network tunneling protocols like MPLS (the latter may be referred to layer 2.5 instead of layer 3). Nageh (talk) 18:58, 27 July 2010 (UTC)

Caps
Two guidelines appear to be relevant to capatilization: Manual_of_Style_(capital_letters) and Manual_of_Style_(capital_letters). I'm more confident about the latter than the former. For tech terms, I'm not sure it is ever perfectly clear what's a proper noun and what's a phrase that just happens to get used frequently. In most cases you're not going to find a reference indicating one way or the other. The fact that it appears with a certain capitalization in a reference may just indicate that that reference was written to a different MOS. The general thrust of my reading of Manual of Style (capital letters) is that capital letters are used sparingly. Where I see inconsistency as I did with Path MTU and Path MTU discovery, I think it is appropriate to err towards lower case. --Kvng (talk) 02:03, 14 September 2010 (UTC)
 * It's not a matter of a style manual or which kind to use or opinion, but a matter of the intention of the originators of the term. The RFCs proclaim it as a proper noun--look at the Path MTU RFC--it is a technical aspect with a name that conveys a specific meaning, not a general class or category. That makes it a proper noun in the English language. If you name your invention Moon and you make it clear that is not just any moon, then you capitalize it, and that's what the RFCs do. The RFCs also use 'Path MTU Discovery' as a proper noun in some places, when they describe the specific standardized method to discover the Path MTU, and not just any alternate method that could potentially obtain a similar result. Language is a finely tuned instrument and you don't have to err, at least not in this case. Kbrose (talk) 04:17, 14 September 2010 (UTC)
 * Would you to add a new section for technical terms to Manual of Style (capital letters) to make this clear in the general case and clear to other editors. There's a lot of variation in how caps are used for technical terms on WP. --Kvng (talk) 13:47, 14 September 2010 (UTC)
 * The rules for technical terms are no different than the general rules of the English language. Problem is many people are just ignorant of them and don't try to take care of details or examine sources of definition. Another problem is that many writers in technical subject, even experts in their fields, speak other native languages where different rules apply, so by translation details get lost. Sometimes it can be rather ambiguous indeed, whether something is a proper noun or a common one, especially with these highly specific technical terms that aren't used anywhere else in common speech, but that's when the original sources often provide the definitive answer. In most cases one will find that the basic rules of English are observed there. In this case, the usage of Path MTU and Path MTU Discovery is rather prevalent in the original literature, although I can certainly think of many possible uses of these without capitalization. Yes, there is a lot of variation on these matters on WP, and it is completely aligned with variation in the quality of writing and command of subject matter, in fact caps issues are probably a minor issue. Kbrose (talk) 15:45, 14 September 2010 (UTC)
 * I don't feel any closer to understanding the general rules for capitalizing technical terms. I've started a discussion at Wikipedia_talk:Manual_of_Style_(capital_letters). --Kvng (talk) 19:55, 14 September 2010 (UTC)
 * WP article proper noun spells it out and there are references to guide books of using the English language. Further this link you cite makes it pretty clear too. Kbrose (talk) 20:39, 14 September 2010 (UTC)
 * Kbrose is right. Yet, I agree with Kvng that a section in Manual_of_Style_(capital_letters) on capitalization of technical terms would be helpful. Some RFCs and technical papers tend to consistently capitalize specific technical terms, no matter whether proper or common nouns. A few examples could help an unsure editor. Nageh (talk) 21:38, 14 September 2010 (UTC)
 * Thanks for your patience. Proper noun is helpful. I guess it comes down to whether you would be more likely to see/say "the Path MTU Discovery" or "a path MTU discovery." This one is not a no-brainer for me. Internet Protocol and IPv6 packet are more obvious. --Kvng (talk) 22:07, 14 September 2010 (UTC)

GiB vs. GB
Elsewhere on Wikipedia, GB is used rather than GiB. Which one should be used here (concerning the entry on jumbograms).Jasper Deng (talk) 03:53, 20 February 2011 (UTC)


 * Per WP:COMPUNITS, the binary prefixes are to be avoided. Rwessel (talk) 07:35, 20 February 2011 (UTC)
 * I hope you mean the IEC prefixes (kibi, mebi, gibi, etcetera) which are NOT to be used. This is how I read the WP:COMPUNITS.—— Da n do r iD (talk) 20:22, 21 February 2011 (UTC)

Endianness?
I think there should be a mention of the endianness of the IPv6 packets. Supermagle (talk) 08:52, 22 February 2011 (UTC)
 * Besides the packet header, it can be anything, but I'm not an expert.Jasper Deng (talk) 18:03, 22 February 2011 (UTC)
 * Why should IPv6, a L3 protocol, care about how a L2 protocol transmits bits and bytes ? Mro (talk) 20:11, 27 February 2011 (UTC)
 * I guess you could ask whether LSB or MSB comes first when transmitting the fields. IPv4 was very clear on Data Transmission Order, but for some reason IPv6 never reiterated this info. --Cybjit (talk) 20:46, 27 February 2011 (UTC)


 * Two issues are being confused here - the order in which bits/bytes/whatever are transmitted on the wire, and the order of bytes in fields in the IPv6 fields.  The former is basically irrelevant, so long as whatever IP-to-link-layer mapping layer exists makes the same sequence of bytes pop out at the other end as went in.  The byte order within IPv6 packets, at least for IETF defined protocols, is big endian, or network byte order.  Obviously a private protocol is allowed to do anything in chooses.


 * It *has* been commented that RFC2460 neglects to mention byte order, except indirectly via references to RFC791 (IPv4). OTOH, RFC1958 pretty much specifies that all standards should use the same bit and byte order (although it doesn't mention what that is, but there's clearly a current practice).  Some other related RFCs do specify things a bit more clearly, for example, RFC3542, while mainly defining a sockets API, also effectively specifies packet layouts for raw sockets, thus providing a backdoor specification of the byte order.   Basically the issue is more that everyone assumed that network byte order was already defined (and it is), and didn't really bother mentioning it further.  And on the extremely slim chance that anyone got it wrong, their IPv6 implementation would wholly fail to interoperate with any other hosts.  Rwessel (talk) 08:32, 28 February 2011 (UTC)