Wikipedia:WikiProject Chemistry/IRC discussions/12 Feb 2008

--- Log opened Tue Feb 12 10:35:19 EST 2008

10:38 -!- itub [n=tubert@lalo.chemie.unibas.ch] has joined #wikichem

10:38 -!- mode/#wikichem [+v itub] by ChanServ

11:02 -!- walkerma [n=chatzill@admin-151-108.potsdam.edu] has joined #wikichem

11:02 -!- Rifleman_82 [n=blahblah@wikipedia/Rifleman-82] has joined #wikichem

11:02 -!- mode/#wikichem [+v Rifleman_82] by ChanServ

11:02 &lt;+Rifleman_82&gt; hi all

11:02 &lt;+Rifleman_82&gt; sorry i'm late, just got home!

11:02 &lt;+Rifleman_82&gt; have you guys started?

11:02 &lt;walkerma&gt; Hello! You're not late

11:03 &lt;+Rifleman_82&gt; sorry about last week....

11:03 &lt;+Rifleman_82&gt; was out buying new year flowers... nurseries open 24 hours!

11:04 &lt;+itub&gt; hi

11:04 &lt;walkerma&gt; I was hoping PC would be here- I just emailed him

11:04 &lt;walkerma&gt; The agenda he suggested last week was: Choice and indexing of identifiers

11:05 &lt;walkerma&gt; Can we talk about this anyway? I think it's a fairly general topic that we need to keep working on.

11:06 &lt;+Rifleman_82&gt; sure

11:06 &lt;walkerma&gt; Or is there something you guys were talking about before I got here?

11:06 &lt;+itub&gt; nope

11:06 &lt;walkerma&gt; That needs finishing off?

11:07 &lt;walkerma&gt; OK: The first on PCs list was:

11:07 &lt;walkerma&gt; Which identifiers (InChI, CAS, etc) are the most important for us (already discussed to some extent)?

11:08 &lt;walkerma&gt; I would put that as: Which should we try and have in all Chemboxes?

11:08 &lt;walkerma&gt; And which should be optional?

11:08 &lt;+itub&gt; inchi and smiles should definitely be optional

11:08 &lt;walkerma&gt; And which should be hidden, and which should be in view?

11:08 &lt;+itub&gt; some compounds don't have a structure that can be handled by them

11:09 &lt;+itub&gt; inorganics and such

11:10 &lt;walkerma&gt; Yes, definitely, though I think IUPAC is working on that

11:10 &lt;walkerma&gt; I think we should make them required fields for organics, but perhaps hide them

11:11 &lt;+Rifleman_82&gt; i would disagree with hiding any fields

11:12 &lt;+Rifleman_82&gt; if the field exists it should be shown

11:12 &lt;+Rifleman_82&gt; if it's too cumbersome, we should move it to the data page

11:12 &lt;+Rifleman_82&gt; but whether something should be in the main chembox or on the data page, we should decide on

11:12 &lt;+Rifleman_82&gt; might be arbitrary, but we should be consistent

11:13 &lt;+Rifleman_82&gt; if there's something that notable such that it should be in the chembox, instead of the data page against our across the board consensus, it should be discussed in-text

11:14 &lt;+Rifleman_82&gt; what do you guys think?

11:14 &lt;walkerma&gt; The reason for hiding is not the issue of notability, it's the issue of (a) size and (b) use

11:14 &lt;+Beetstra&gt; Well, some things don't show nicely .. but need to be in the main-page, otherwise it does not get google-indexed

11:15 &lt;+Beetstra&gt; (a.o.)

11:15 &lt;+Rifleman_82&gt; a.o. ?

11:15 &lt;walkerma&gt; Beetstra: Right!

11:15 &lt;walkerma&gt; http://en.wikipedia.org/w/index.php?title=Quassin&oldid=188784759

11:15 &lt;+Beetstra&gt; amongst other things ..

11:15 &lt;+Rifleman_82&gt; ah thenotorious quassin

11:15 &lt;+Beetstra&gt; (google, yahoo, alexa, ..., name them)

11:16 &lt;+Rifleman_82&gt; the problem with hidden fields is that it is hidden to potential contributors -

11:16 &lt;+Beetstra&gt; There is a nice dilemma there, indeed

11:16 &lt;+Beetstra&gt; And I don't know if something can be made invisible, while it is still there in the html-output of the wikipage

11:16 &lt;walkerma&gt; If we limit hidden fields to being ONLY machine-readable ones, it's not a problem because

11:17 &lt;+Rifleman_82&gt; cut and pasting? i do it quite a bit

11:17 &lt;+Rifleman_82&gt; for searches

11:17 &lt;walkerma&gt; These can be generated by a machine - so passing contributors only need supply another form of input

11:17 &lt;walkerma&gt; Such as a structure or IUPAC name?

11:18 &lt;+Rifleman_82&gt; iupac name can be generated by machine too

11:18 -!- PC62|away is now known as Physchim62

11:18 &lt;+Rifleman_82&gt; and if smiles/inchi is there already, the structure can be generated too

11:18 &lt;+Physchim62&gt; HI? SORRY i4M LQTE

11:18 &lt;+Rifleman_82&gt; (circular problem?)

11:18 &lt;+Rifleman_82&gt; hi pc

11:18 &lt;+Rifleman_82&gt; we just started

11:19 * Beetstra hands Physchim62 an 'A' ..

11:19 &lt;+Beetstra&gt; (or two)

11:19 &lt;walkerma&gt; Hi, we're just discussing "To hide or not to hide"

11:19 &lt;walkerma&gt; Rifleman82: Harder to make a NICE structure from InChI/SMILES

11:20 &lt;+Rifleman_82&gt; surprisingly, it works well enough on chemsketch most of the time

11:20 &lt;walkerma&gt; Yes, ChemSketch is not too bad

11:20 &lt;+Rifleman_82&gt; as in, i copy smiles from pubchem into chemsketch and generate the structure to generate the image to uplaod

11:20 &lt;+Rifleman_82&gt; (jmol will solve the problem? :) )

11:21 &lt;walkerma&gt; As a reminder, the feedback from the listserve makes very good reading on this topic:

11:21 &lt;walkerma&gt; http://en.wikipedia.org/wiki/User:Walkerma/Sandbox5

11:21 &lt;+Rifleman_82&gt; pc sending you log of today

11:21 -!- mode/#wikichem [+o Rifleman_82] by ChanServ

11:22 -!- mode/#wikichem [+v walkerma] by Rifleman_82

11:22 -!- Rifleman_82 changed the topic of #wikichem to: *Next channel meeting: 19 Feb 2008; 1600h UTC* Welcome to the Wikipedia Chemistry IRC channel. Queries to User:Rifleman_82 or User:Beetstra.

11:22 &lt;+Physchim62&gt; Thnaks I've jusr read through my screen base

11:22 -!- mode/#wikichem [-o Rifleman_82] by Rifleman_82

11:22 &lt;+Rifleman_82&gt; oops forgot :)

11:22 &lt;+Physchim62&gt; no probs.

11:23 &lt;+walkerma&gt; Here is an interesting response to the question, Do you ever copy/paste such identifiers FROM Wikipedia into Google, etc, in order to do a search?

11:23 &lt;+Rifleman_82&gt; martin i promised to check on making text wrap arbitrarily, such as every 25 chars. tried and searched and i can't figure it out. sorry about that

11:23 &lt;+walkerma&gt; "This could be very useful, too. It doesn't matter if it's visible or not, a link "copy SMILES" would be enough. Thanks for asking."

11:23 &lt;+Physchim62&gt; Beetstra, "hidden" content still appears in HTML output (there is a lot that is hidden in HTML)

11:24 &lt;+itub&gt; the one I copy most often is CAS, to search in scifinder

11:24 &lt;+Beetstra&gt; Yes, it would be not too hard to do, actually

11:24 &lt;+walkerma&gt; Rifleman82: (Thanks, that's a shame it's so hard to do)

11:24 &lt;+Rifleman_82&gt; the possible solutions would require some developer support i think...

11:24 &lt;+Rifleman_82&gt; there are many other workarounds

11:24 &lt;+Beetstra&gt; But there are things that get lost if it does not get used .. I don't think shows up in the final html

11:24 &lt;+Rifleman_82&gt; but nothing easily usable

11:24 &lt;+Rifleman_82&gt; <!-- does show up

11:25 &lt;+Rifleman_82&gt; brb

11:25 &lt;+Physchim62&gt; Andrew's problem's seem to arise from a lack of browser compatibility in MediaWiki

11:25 &lt;+Physchim62&gt; for once, IE seems to wrap OK, but Mozilla is awful

11:25 &lt;+itub&gt; I just tried adding a <!-- and it didn't show up

11:26 &lt;+Physchim62&gt; :P

11:26 &lt;+Physchim62&gt; no, HTML comments are removed by MediaWiki

11:26 &lt;+Physchim62&gt; (not least for legal reasons)

11:27 &lt;+walkerma&gt; So would it be possible to have an entry that says "Copy SMILES"?

11:27 &lt;+Physchim62&gt; if you want to add "hidden" content, *please* put it in a table

11:28 &lt;+Physchim62&gt; in JavaScript, yes, but I personally don't know how to do it

11:28 &lt;+walkerma&gt; I wonder if Cacycle could do this?

11:28 &lt;+walkerma&gt; He's a whiz with Java

11:29 &lt;+Physchim62&gt; we would need to make a good politiocal case to other Wikipedians for including it in common.js

11:29 &lt;+walkerma&gt; Ah, yes, I see

11:29 &lt;+Physchim62&gt; I almost got lynched over class="InChI"

11:30 &lt;+itub&gt; I think copying to the clipboard is problematic

11:30 &lt;+Beetstra&gt; they are javaphobic at mediawiki .. partially understandable

11:30 &lt;+Rifleman_82&gt; pc: i see your point. i just checked out quassin on ie7

11:30 &lt;+itub&gt; depends on the browser, config, etc.

11:30 &lt;+Physchim62&gt; which is why I wanted input as to which classifiers wre the *most* important

11:30 &lt;+Beetstra&gt; I needed java once for something .. Brian killed me ..

11:30 &lt;+itub&gt; security restrictions normally prevent scripts from accessing the clipboard

11:31 &lt;+Rifleman_82&gt; heh.. maybe we can bug mozilla for the wrapping

11:31 &lt;+Physchim62&gt; OK, so its a no-starter then

11:31 &lt;+Physchim62&gt; RM, please do!

11:31 &lt;+Rifleman_82&gt; ok i'll go email them or something, and update on the actions column

11:32 &lt;+Physchim62&gt; unfortunately, my HTTP access is limited at the moment, so I cannot promise to do it myself

11:32 &lt;+itub&gt; it would at least be possible to have a script pop up a box where the user can copy manually ;-)

11:33 &lt;+Physchim62&gt; I have a gripe :P

11:33 &lt;+itub&gt; but after seeing the debate over the "tooltips" in the infoboxes I expect there will be some opposition

11:33 &lt;+Beetstra&gt; itub, that would also be solved by giving every chemical page a /data page ..

11:34 &lt;+Physchim62&gt; SMILES and InChI seem to have become necessary elements in Template:Chembox identifiers, and so there are many chemboxes which show these lines without any information to go in them

11:34 &lt;+Beetstra&gt; Then you can safely hide it in the page itself, and just put it in the datapage ..

11:34 &lt;+itub&gt; a data page is the simplest and most accessible option, I think. But also the one most likely to cause maintenance headaches

11:34 &lt;+Beetstra&gt; something like 4500 of them ..

11:35 &lt;+Physchim62&gt; data pages have never really taken off, IMHO, dispite the best intentions of all involved

11:35 &lt;+walkerma&gt; Physchim62: Karlhahn has done a lot with the common chemicals, thankfully

11:35 &lt;+Physchim62&gt; the naming format should be Acetic acid (data page)

11:35 &lt;+Beetstra&gt; They are sometimes nice .. but most of the time .. it is extra work, and the data fits in the page itself often enough

11:36 &lt;+itub&gt; in most cases there is very little data to add, and the data page is not visited as often which means that it is not updated as often

11:36 &lt;+walkerma&gt; But otherwise you're right

11:36 * Physchim62 agrees with itub

11:36 &lt;+walkerma&gt; The background with data pages: We had some pages where the chembox was getting as long as your arm

11:37 &lt;+Beetstra&gt; I think if you look how many pages transclude the chembox new module for the datapage .. you know how many there are

11:37 &lt;+Physchim62&gt; which it is starting to again

11:37 &lt;+walkerma&gt; Filled with the latent heat of vaporization at standard entropy on the planet Neptune, etc.

11:37 &lt;+itub&gt; something that could be added to every data page (although perhaps it's better to do some other way) are the 3d coordinates

11:38 &lt;+Physchim62&gt; itub, which 3D coordinates :P

11:38 &lt;+walkerma&gt; itub: How would that be done?

11:38 &lt;+itub&gt; I think it's been mentioned before, but I would really love if WP had 3d coords ;-)

11:38 &lt;+walkerma&gt; Jmol

11:38 &lt;+walkerma&gt; ?

11:38 &lt;+itub&gt; it could even be just a table

11:38 &lt;+Physchim62&gt; we can do that fairly easily, if the choice of coordinates is left to editorial discretion

11:39 &lt;+walkerma&gt; I think we need a standard format for it to be useful

11:39 &lt;+Physchim62&gt; and we worry about the problems later :P

11:39 &lt;+Physchim62&gt; we can talk to PMR about possible formats

11:40 &lt;+walkerma&gt; I think he will say CML

11:40 &lt;+itub&gt; probably

11:40 &lt;+Physchim62&gt; he has been active in that area

11:40 &lt;+itub&gt; I'd vote for XYZ, the most primitive format

11:40 &lt;+Physchim62&gt; well in that case he can ...

11:41 &lt;+Physchim62&gt; we would have war on our hands at the moment if we tried to introduce CML to WP

11:41 &lt;+Physchim62&gt; next year, maybe, if the developers love us

11:42 &lt;+walkerma&gt; OK, so how can we proceed with this? Is hiding InChIs etc (a) Just impossible or (b) Definitely possible, but in the future or (c) Don't know?

11:43 &lt;+walkerma&gt; (We can come back to 3D)

11:43 &lt;+itub&gt; look, the compchem wiki has jmol, cartesian coords, and internal coords: http://209.85.129.104/search?q=cache:8KxyH7a33WAJ:compchemwiki.org/index.php%3Ftitle%3DCaffeine+site:compchemwiki.org&hl=en&ct=clnk&cd=2

11:43 &lt;+Beetstra&gt; b .. just hide them, but in a way that they show up in the final html

11:43 &lt;+itub&gt; ok

11:43 &lt;+Physchim62&gt; hiding InChIs could be done by any administrator tonight... oh, you want them to be visible on request ;)

11:43 &lt;+Beetstra&gt; and as long as they are used in the chembox new with a 'InChI=' statement .. if we decide differently .. we can always use them different after that

11:44 &lt;+itub&gt; that can be done with stylesheets

11:44 &lt;+Beetstra&gt; i think that if you remove the info from template:chembox InChI, the line will be empty on-wiki ..

11:44 &lt;+walkerma&gt; We want a user to know they're there, we want Google et al to find them, and we want people to be able to cut/past

11:44 &lt;+walkerma&gt; e

11:45 &lt;+walkerma&gt; And we'd very much prefer to keep them on the main page, not the data page, if poss

11:45 &lt;+Physchim62&gt; if you remove the call to Template:Chembox InChI in Template:Chembox Identifers, all InChIs will disappear from cheboxes (good), but also from the HTML (BAD)

11:46 &lt;+itub&gt; something like the collapsible sections sometimes used in talk pages?

11:46 &lt;+Physchim62&gt; I would suggest a bot run to copy all InChIs into Template:InChI

11:46 &lt;+walkerma&gt; Physchim62: Yes

11:47 &lt;+Physchim62&gt; that way data is preserved both in wikitext and in the HTML outpt

11:47 &lt;+walkerma&gt; But how do we avoid a problem like at Beta-hydroxybutyryl-CoA?

11:47 &lt;+walkerma&gt; http://en.wikipedia.org/wiki/Beta-hydroxybutyryl-CoA

11:47 &lt;+Beetstra&gt; But Template:Chembox InChI contains the actual line .. if you make it ' .. it is just an appearing tag .. but you don't see anything in the chembox on the page ..

11:47 &lt;+Physchim62&gt; and certain InChIs can be retained, and others deleted, from the chembox proper

11:47 &lt;+walkerma&gt; (Look at it in FIrefox)

11:48 &lt;+Physchim62&gt; Martin, I've seen it :(

11:48 * Beetstra does not understand ..

11:48 &lt;+Beetstra&gt; What is wrong there ..

11:49 &lt;+walkerma&gt; Or even worse is http://en.wikipedia.org/w/index.php?title=Beta-hydroxybutyryl-CoA&oldid=187504473

11:49 &lt;+Rifleman_82&gt; whoa

11:49 * Beetstra repairs

11:49 &lt;+Beetstra&gt; or thinks he does :-(

11:50 &lt;+Physchim62&gt; Beetstra, Template:InChI needs a

as its first line, at least until we have soted out browser problems

11:51 &lt;+Physchim62&gt; that will push the problem to the very bottom of the page until we can find another solution

11:52 &lt;+walkerma&gt; Physchim62: Would it be possible to direct the user from the Chembox to the InChI table in some way?

11:52 &lt;+walkerma&gt; Or do we just add "See foot of page" in Chembox?

11:52 &lt;+Rifleman_82&gt; as a quick and dirty solution, why not we just disable smiles and inchi for the moment?

11:52 &lt;+Physchim62&gt; not in the current setup, but it's doable

11:53 &lt;+walkerma&gt; Rifleman82: Could we hold off on that until we are ready to upload all of Antony's SDF file?

11:53 &lt;+walkerma&gt; Hopefully that will be ready this month

11:53 &lt;+Rifleman_82&gt; walkerma: not sure what you mean

11:53 &lt;+Physchim62&gt; RM, that would be overkill

11:54 &lt;+Rifleman_82&gt; itub: compchemwiki looks great... if only they'd let us use it

11:54 &lt;+walkerma&gt; Antony (AKA ChemSpiderMan) will be giving us a validated set of ~6000 compounds

11:54 &lt;+Physchim62&gt; Beta-hydroxybutyryl-CoA was a test case, after all, we used it to see what problems we might have and it is showing them :/

11:54 &lt;+walkerma&gt; as an SDF file containing structures, IUPAC names, InChIs, InChIKeys

11:54 &lt;+itub&gt; We can produce our own coordinates too

11:55 &lt;+Rifleman_82&gt; i don't mean the coords

11:55 &lt;+Rifleman_82&gt; i mean the image

11:55 &lt;+walkerma&gt; Rifleman82: Do you understand what I mean now?

11:55 &lt;+Physchim62&gt; the image needs changing too, yes

11:55 &lt;+Beetstra&gt; WHAHAHAHA .. it had nothing to do with the chembox new

11:56 &lt;+Rifleman_82&gt; walkerma: hold off = don't touch smiles/inchi? or don't disable it in

11:56 &lt;+Rifleman_82&gt; i understand antony's work

11:56 * Beetstra slowly strangles Rifleman_82 and Physchim62 ..

11:56 &lt;+walkerma&gt; Yes

11:56 &lt;+Rifleman_82&gt; beetstra: the image?

11:56 &lt;+Beetstra&gt; -> http://en.wikipedia.org/w/index.php?title=Template:Chembox_Identifiers2&action=history

11:56 &lt;+Rifleman_82&gt; yes = ? which?

11:56 &lt;+Beetstra&gt; take note of the '2'

11:57 &lt;+Physchim62&gt; I have

11:57 &lt;+Rifleman_82&gt; beetstra: what about the image? what about chembox new?

11:57 &lt;+Beetstra&gt; That is why Beta-hydroxybutyryl-CoA displayed wrongly

11:57 &lt;+Physchim62&gt; Beetstra, wrong!

11:57 &lt;+Beetstra&gt; There is an error in the copied template ..

11:58 &lt;+Physchim62&gt; the separate template was created so that I could test the addition of InChI keys into chemboxes

11:58 &lt;+Physchim62&gt; go ahead

11:59 &lt;+Beetstra&gt; Ah, OK .. now I understand

11:59 &lt;+Beetstra&gt; But there was something else wrong in that

11:59 * Physchim62 loves it when the boss tells me exactly why my experiment has failed ;)

12:01 &lt;+walkerma&gt; What's wrong Dirk? We're on the edge of our seats.....!

12:03 &lt;+Physchim62&gt; while Dirk is searching for my mistake, can I ask the participants if they really want to see blank lines of SMILES and InChI in chemboxes?

12:03 &lt;+itub&gt; y don't like the blank lines

12:04 &lt;+walkerma&gt; Physchim62: I'd like to see either "Click to display/copy" or "At foot of page"

12:04 &lt;+walkerma&gt; We shouldn't just have blank lines

12:04 &lt;+walkerma&gt; But bear in mind that Antony will fill up most of the lines for us soon

12:05 &lt;+itub&gt; I just had trouble with an inorganic compound that had blank smiles and inchis because they don't really apply there, and the fields refused to go away...

12:06 &lt;+Rifleman_82&gt; i think it's fair to show the fields only when the field is defined

12:06 &lt;+walkerma&gt; Yes, we should definitely get rid of these for inorganics

12:08 &lt;+Beetstra&gt; I solved the mistake(s), missing |'s

12:08 &lt;+Physchim62&gt; so can that be added to "actions" to return those fields to their previous status as optional for dispaly?

12:09 * Physchim62 is sorry to order people around, but he can no longer do it himself

12:09 &lt;+walkerma&gt; Physchim62: I'd say yes

12:09 &lt;+walkerma&gt; Physchim62:We won't take offence!

12:10 &lt;+Rifleman_82&gt; pc: tell me how to do it and i will

12:10 &lt;+Physchim62&gt; Beetstra, can you restore the test versions of, say Beta-hydroxybutyryl-CoA to see if they work in Firefox?

12:10 &lt;+Rifleman_82&gt; best if you cut and paste the desired correction and message me

12:10 &lt;+Beetstra&gt; I did, Beta..CoA now uses the test-version .. like you left it

12:11 &lt;+Physchim62&gt; RM, I will be on HTTP once this discussion is over, I shall leave a message on the Talk page

12:11 &lt;+Beetstra&gt; i just repaired both templates

12:11 &lt;+Rifleman_82&gt; ok

12:11 &lt;+walkerma&gt; Beetstra:Thanks!

12:12 &lt;+Physchim62&gt; :) (and sorry if I messed up)

12:12 &lt;+Beetstra&gt; Not sure .. it may have been something older

12:13 &lt;+Physchim62&gt; it was probably me, I had to make the change to include InChIKey (but it worked OK on IE...)

12:13 &lt;+Beetstra&gt; Opera screwed it .. and apparently FireFox as well

12:14 &lt;+walkerma&gt; Time is getting on, and I'm thinking we should probably draw things to a close.

12:14 &lt;+Physchim62&gt; FireFox screwed it up, but I only realised that later ;)

12:14 &lt;+Physchim62&gt; a straw poll on identifiers?

12:14 &lt;+Physchim62&gt; CAS, obviously

12:15 &lt;+walkerma&gt; I knew that this meeting wouldn't magically solve all of our problems, but I think some progress has been made

12:15 &lt;+Physchim62&gt; InChI seems to have supprt for organics at least (+ InChIKey)

12:15 &lt;+walkerma&gt; Must haves: CAS, SMILES, InChI, InChIKey, though the last three could be hidden as needed, and should only be used with organics

12:16 &lt;+itub&gt; I would say "moleculars" rather than organics :)

12:16 &lt;+Physchim62&gt; I am sceptical about the utility of SMILES, but if others disagree...

12:16 &lt;+walkerma&gt; A 3D version in a standard format could be a nice option - I'd suggest JMol format (though I'm not certain)

12:16 &lt;+walkerma&gt; SMILES is still popular in the pharma industry I believe

12:17 &lt;+walkerma&gt; Different specialisations tend to like specific things

12:17 &lt;+itub&gt; it is popular, and I think will remain popular as it is the only line notation that is almost readable

12:17 &lt;+Physchim62&gt; Peter Murray-Rust has argued for PubChem in an opinion piece in this week's Nature: it wouldn't cause us any more problems than the others

12:18 &lt;+walkerma&gt; Yes, PubChem would be good too

12:18 &lt;+Physchim62&gt; OK, once again, I admit defeat on the SMILES question ;)

12:18 &lt;+Rifleman_82&gt; is there any way to transform InChI to SMILES?

12:18 &lt;+walkerma&gt; Yes, I think so

12:18 * Beetstra has to go .. will read up later

12:18 &lt;+Rifleman_82&gt; or vice versa?

12:18 &lt;+Physchim62&gt; RM, it could be done

12:18 &lt;+Physchim62&gt; vice versa is impossible

12:18 &lt;+Rifleman_82&gt; could wikipedia do it on the fly?

12:18 &lt;+Physchim62&gt; no

12:18 &lt;+Beetstra&gt; IMHO, support them all, but make a choice of which would be the ones to use

12:19 -!- Beetstra [n=djbeetst@Wikimedia/Beetstra] has quit ["Bye Bye"]

12:19 * Physchim62 agrees with Beetstra

12:19 -!- walkerma_ [n=chatzill@admin-151-108.potsdam.edu] has joined #wikichem

12:19 &lt;+Rifleman_82&gt; martin you didn't miss much

12:19 &lt;walkerma_&gt; (not sure what is happening here with my name!)

12:20 &lt;+itub&gt; it's possible both ways in general, but inchi supports some features that smiles doesn't

12:20 &lt;walkerma_&gt; And InChI is unambiguous and has only one format. SMILES has various formats and may be ambiguous

12:20 &lt;+Physchim62&gt; vice versa is possible in simple cases, which are exactly the cases where SMILES is preferred over InChI

12:21 &lt;+itub&gt; that is true

12:21 &lt;+itub&gt; smiles is nearly useless for simple google searches and such because there is no standard canonical smiles

12:22 &lt;+Physchim62&gt; Anthony claims that InChI is ambiguous, although I think he is counting angels on pinheads as far as WP is concerned

12:23 &lt;walkerma_&gt; http://www.chemspider.com/blog/does-inchi-account-for-tautomers.html

12:23 &lt;+Physchim62&gt; our need for unambiguity is not the same as that of ChemSpider

12:23 &lt;walkerma_&gt; Physchim62: I agree

12:24 &lt;walkerma_&gt; It seems to me that we not only need a "primary key" for each ChemBox structure - should we define that as the structure as drawn?

12:24 &lt;walkerma_&gt; Sorry, I got mixed up

12:24 &lt;walkerma_&gt; We NOT ONLY need a primary key for the article

12:25 &lt;walkerma_&gt; We ALSO NEED a primary key for each Chembox

12:25 &lt;walkerma_&gt; to define the structure which the data applies to

12:25 &lt;+Physchim62&gt; we *cannot have* a primary key for each chembox

12:25 &lt;walkerma_&gt; (or structures if more than one is given)

12:26 &lt;walkerma_&gt; Physchim62: Explain

12:26 &lt;walkerma_&gt; I accept this may be true for inorganics, I'm considering organics at the moment

12:27 &lt;walkerma_&gt; (Walkerma apologises to be repeating old discussions!)

12:27 &lt;+Physchim62&gt; glucose

12:28 &lt;walkerma_&gt; OK, but we need to know WHICH form of glucose has WHICH InChI, CAS# etc

12:28 &lt;+Physchim62&gt; apart from the fact that chemboxes are meant to be a part of an /article/, not a seperate chemiacla database

12:28 &lt;walkerma_&gt; Either we only display one form, or we define KeyA, KeyB, KeyC etc

12:28 &lt;+Physchim62&gt; disagree

12:29 &lt;+Physchim62&gt; We can have a unique ("primary") key for each *structure*, but not for each chembox

12:30 &lt;walkerma_&gt; OK, I can agree with that

12:30 &lt;+Rifleman_82&gt; unless you want to have multiple chemboxes for different forms?

12:30 &lt;walkerma_&gt; That's in effect what I meant to say

12:30 &lt;+Rifleman_82&gt; the multiple chemboxes may be abbreviated

12:30 &lt;+Rifleman_82&gt; formula, mw can be dispensed with

12:30 &lt;walkerma_&gt; Rifleman_82: That's what I had in mind

12:31 &lt;+Physchim62&gt; otherwise, we would have to have seperate articles for each enantiomer, each hydrate, etc

12:31 &lt;+Rifleman_82&gt; but once again, i wonder if we really need so much detail

12:32 &lt;+Rifleman_82&gt; i don't think we should attempt to replicate the CRC handbook, even though wp is already being used as such

12:32 &lt;walkerma_&gt; OK, I think we basically agree - so should we use the STRUCTURE as the main identifier, and generate the other things from that (except CAS which can't be done automatically?)

12:32 &lt;+Physchim62&gt; I wonder if it wouldn't be easier to have ad hoc tables of CAS numbers, etc for such cases

12:32 &lt;+Physchim62&gt; NONONONONO!

12:33 &lt;+Physchim62&gt; the main identifier is the *article name* - if we change that, we stop being an encyclopedia

12:34 &lt;+Rifleman_82&gt; agree with pc

12:34 &lt;+Physchim62&gt; we ca try to make things easier for those people who don't wish to call thing by their name

12:34 &lt;walkerma_&gt; Physchim62: OK, let's say I have an article Tartaric acid. The primary key for the page is (rightly) Tartaric acid. It has an InChI on the page. What does the InChI mean?

12:35 &lt;+Physchim62&gt; Martin, you're the organic chemist, you tell me! :P

12:35 &lt;walkerma_&gt; Does it represent the R,R, the S,S, or the meso form? Or unspecified?

12:35 &lt;+Physchim62&gt; we need to make sure that all three InChIs are included in the HTML file

12:35 &lt;walkerma_&gt; We MUST define what the InChI refers to, what the CAS refers to, etc

12:35 &lt;+itub&gt; it has to be the "generic" one

12:35 &lt;+Physchim62&gt; CAS is not usually a problem in these cases

12:35 &lt;+Rifleman_82&gt; hmmm

12:36 &lt;+Rifleman_82&gt; how about a new section/

12:36 &lt;walkerma_&gt; OK, and so now we have three InChIs - which one refers to which form?

12:36 &lt;+itub&gt; if the format doesn't support it or there is not a generic CAS, then there should be none

12:36 -!- walkerma [n=chatzill@admin-151-108.potsdam.edu] has quit [Read error: 110 (Connection timed out)]

12:36 &lt;+Physchim62&gt; as for the InChI, I suggest we leave that until next week when we discuss the "protonation problem"

12:36 &lt;+Rifleman_82&gt; so you have properties, hazards, related, etc. have one: isomers

12:36 &lt;+itub&gt; then there can be a table with all the isomers

12:36 &lt;+Rifleman_82&gt; yeah

12:37 &lt;walkerma_&gt; Physchim62: We can discuss the protonation problem next week, but let's say for now we have three InChIs. Which one goes with which structure?

12:37 &lt;+Rifleman_82&gt; there you can have pretty much freeform text cas smiles and inchi

12:37 &lt;+itub&gt; all the properties can also be dependent on the specific isomer

12:37 &lt;+Rifleman_82&gt; for each

12:37 &lt;+Rifleman_82&gt; properties?

12:37 &lt;+Rifleman_82&gt; mp perhaps?

12:38 &lt;+itub&gt; mp, hazards, solubility, etc.

12:38 &lt;+CheMoBot&gt; user:TJ Terry has edited monitored page Alprazolam (Reason for monitoring: 'happy spam target') - diff - (-347)- summary: /* Side effects */

12:38 &lt;+itub&gt; the molar mass doesn't depend on the isomer, but it can vary if you consider related forms such as hydrochlorides (especially common for drug-like molecules)

12:38 &lt;+Rifleman_82&gt; we'll skip the .HCl for the moment?

12:38 &lt;+itub&gt; sure

12:39 &lt;walkerma_&gt; Rifleman82: Yes, that's next week

12:39 &lt;+Rifleman_82&gt; hazards, solubility - we can keep it in the main box; for most isomers they shouldn't differ too far from each other

12:39 &lt;+Rifleman_82&gt; where they are significantly different, discuss in text/

12:39 &lt;+Physchim62&gt; walkerma_, you are assuming that people want to use WP to go from InChI to structure. I'm sorry to be uncivil, but that is ridiculous. We cannot give them that, nor do we pretend to. However there are several freeware packages which will do such a conversion

12:40 &lt;+itub&gt; I just noticed that http://en.wikipedia.org/wiki/Tartaric_acid already has a table lising the CAS numbers and other identifiers for each isomer!

12:40 &lt;+Physchim62&gt; people want to use WP to go from name to another identifier, or possibly from structure to name

12:41 &lt;walkerma_&gt; Not if they found the page initially from a Google search

12:41 &lt;walkerma_&gt; of the InChI or InChIKey

12:41 &lt;+Physchim62&gt; you hypothesis is that they searched for an InChI and then found the WP page! so they already know the InChI they want, we do not have to inform them of it

12:42 &lt;walkerma_&gt; But they need to know which structure goes with which InChI

12:42 &lt;walkerma_&gt; Here are three responses to my survey:

12:42 &lt;+Physchim62&gt; if they want a structure from an InChI then sorry, MediaWiki is not the best software solution to do this relatively simple transformation

12:42 &lt;walkerma_&gt; # I search multiple internet sources (e.g. ChemSpider, Wikipedia, search engines, online catalogs, etc) using various chemical identifiers (common names, IUPAC names, CAS number, SMILES, INCHI, etc). Since I didn't see your original posting, I don't have the full context for your questions.

12:42 &lt;walkerma_&gt; # YES

12:43 &lt;walkerma_&gt; # Yes and I would use it more if I knew that I would come up with more useful data, including Wikipedia entries. I copy SMILES or InChI strings from Chemdraw and paste them in internet searches.

12:43 &lt;+itub&gt; yes, whenever there is an identifier, it has to be clear what it means

12:43 &lt;+itub&gt; in ambiguous cases, an ambigous identifier should be used (for example, tartaric acid has a SMILES with no stereochemistry)

12:43 &lt;+Physchim62&gt; OK, so we eliminate all InChIs for possible ambiguity, is that what you want?

12:44 &lt;walkerma_&gt; I don't see us competing with ChemSpider - it's more a case that they are looking at a structure and they want to copy/paste the RIGHT InChI

12:44 &lt;+itub&gt; when detail is wanted, an extra table can be added, again as in tartaric acid

12:45 &lt;+itub&gt; what I wouldn't want is to have an article called tartaric acid with just one inchi which hapens to be the one for a randomly chosen isomer

12:45 &lt;walkerma_&gt; Tartaric acid does this fine, as far as it goes

12:45 &lt;walkerma_&gt; itub: Yes

12:46 &lt;walkerma_&gt; itub: This is unfortunately what we have in many cases

12:46 &lt;walkerma_&gt; Many of the problems Antony found were of this sort

12:46 &lt;walkerma_&gt; So which do we define as right - the Structure or the InChI? Or the IUPAC name?

12:46 &lt;walkerma_&gt; I would choose the structure, and derive everything from that

12:49 &lt;+itub&gt; we need to start from the title and use chemical sense... :)

12:50 &lt;+itub&gt; if the title is tartaric acid, one expects an article about a set of isomers

12:50 &lt;+itub&gt; if the title is DL-tartaric acid, one expects an article about a specific compound

12:52 &lt;+itub&gt; a more obvious example would be http://en.wikipedia.org/wiki/Xylene

12:52 &lt;+itub&gt; which again deals with an unspecified mixture of isomers, with its very own CAS number but no SMILES

12:53 &lt;+itub&gt; so, the first thing one needs to ask is "what is this ARTICLE about?"

12:54 &lt;walkerma_&gt; That sounds like a good starting point for next week's discussion

12:56 &lt;walkerma_&gt; Should we call it a day/night?

12:56 &lt;+Rifleman_82&gt; yeah sure

12:56 &lt;+itub&gt; sure

12:56 &lt;+Rifleman_82&gt; i won't be around next week, travelling

--- Log closed Wed Feb 13 12:56:23 EST 2008