Wikipedia:WikiProject Chemistry/IRC discussions/29 July 2008

--- Log opened Tue Jul 29 12:13:01 EDT 2008

12:13 &lt;+Physchim62&gt; yep, where's this discussion about Dirk's bot?

12:13 &lt;+Rifleman_82&gt; it's mired at BAG

12:13 &lt;walkerma&gt; http://en.wikipedia.org/wiki/Wikipedia:Bots/Requests_for_approval/CheMoBot

12:14 &lt;+Rifleman_82&gt; they're getting sidetracked into the same discussions we've had

12:14 &lt;walkerma&gt; I think the main problem is with people just misunderstanding what he's doing. Heck, I don't know if I know what he's doing!

12:16 &lt;+Physchim62&gt; no, I don't think it's "mired"

12:17 &lt;+Physchim62&gt; from a very quick glance at the discussion, it seems remarkably constructive, given the implications of what we're suggesting

12:17 &lt;+CheMoBot&gt; user:Gordonofcartoon has edited Chembox_new containing page Formic acid - http://en.wikipedia.org/w/index.php?title=Formic_acid&diff=228627591&oldid=228548258 (+423) - Summary: ' /* Safety */ not forgetting meths poisoning'

12:17 &lt;+CheMoBot&gt; Formic acid links: http://www.newton.dep.anl.gov/askasci/chem03/chem03561.htm

12:17 &lt;walkerma&gt; CBM is really good, and he seems to helped in guiding the process. He's writing our WP1.0 SelectionBot, he's a very smart guy

12:17 &lt;+Rifleman_82&gt; Chemobot help

12:17 &lt;+CheMoBot&gt; Hi Rifleman_82! Commands: wl search, bl search, ml search, status, boxlist, fieldlist, managers, channel, project, count, show, check

12:17 &lt;+Rifleman_82&gt; chemobot silent

12:17 &lt;+CheMoBot&gt; CheMoBot is silent (use 'CheMoBot report' to make me report again).

12:18 &lt;walkerma&gt; Wow, I wish my 3 year old worked like that!

12:18 &lt;+Physchim62&gt; I've always said that we will have reticence from the rest of the WP community for great changes, especially those which seem only to serve WikiChem

12:19 &lt;+Rifleman_82&gt; hahahaha

12:19 * Physchim62 rewrites Martin's daughter in Perl

12:19 &lt;walkerma&gt; Thanks PC! My impression, though, is that the people over at BAG do appreciate the value of what we're trying to do

12:20 * Physchim62 apologizes for the programming errors, but isn't any offspring just a beta-version?

12:20 &lt;walkerma&gt; Also, I think the default is normally to assume good faith from a responsible, seasoned Wikipedian like Beetstra

12:20 &lt;+Physchim62&gt; that's my impression as well, and they have expertise which we don't, not least in WikiPolitics

12:20 &lt;walkerma&gt; (Physchim62: I think the potty training was better on the old version)

12:21 * Physchim62 removes Function

12:22 &lt;walkerma&gt; Dirk's idea is to demonstrate the concept first by simply LOGGING problems, and very few people can argue against that

12:22 &lt;walkerma&gt; (on political grounds)

12:22 &lt;+Physchim62&gt; at some point or another, we have to submit our ideas to the wider community

12:24 &lt;walkerma&gt; Let's get it working first. We may or may not have flagged revisions - that would affect the debate a lot

12:24 &lt;+Physchim62&gt; and most of the wider community isn't interested in chemistry, even though they support the work we do to keep chemistry coverage on WP

12:25 &lt;walkerma&gt; I think you may find other people like Physics or Medicine adopting it first, or perhaps the really well-organised projects like MILHIST and CYCLONE

12:25 &lt;+Physchim62&gt; no, all of those projects run into the same kind of problems, I can assure you

12:26 &lt;+Rifleman_82&gt; what problems?

12:26 &lt;+Rifleman_82&gt; bot approval? validating data? autonomous operation?

12:26 &lt;+Rifleman_82&gt; or all of them and more?

12:27 &lt;+Physchim62&gt; two practical problems: on WP, you have to convince people; and its easy for someone to object

12:27 &lt;+Physchim62&gt; and, of course, the lower level problems you mention

12:28 &lt;+Rifleman_82&gt; seems like beetstra's trying to clear the bot approval +2 problems for everyone, assuming they piggyback his bot

12:28 &lt;+Rifleman_82&gt; as in, clear the problems for everyone who might potentially use the bot

12:28 &lt;+Rifleman_82&gt; which leaves... the politics

12:28 &lt;walkerma&gt; I think you need to do the classic salesman thing - make them believe that it's something their project NEEDS (or at least wants). And I don't think you'd have to work too hard to sell them on that!

12:29 &lt;+Rifleman_82&gt; heh alright

12:29 &lt;walkerma&gt; I'm sure I could talk with people like Kirill Loshkin, and Roger Davies, and if they liked the idea, it's very likely MILHIST would adopt it. They have over 500 members right there.

12:30 &lt;walkerma&gt; And I think over 100,000 articles

12:30 &lt;walkerma&gt; And once a group like MILHIST adopts it, a lot of people follow - MILHIST is often seen as a model of how things should be done

12:31 &lt;+Rifleman_82&gt; i thought we are the model...

12:31 &lt;walkerma&gt; No, we're just the back room boffins!

12:33 &lt;walkerma&gt; I think we're respected, but not as influential as some of the well-coordinated big projects - we don't have their level of organisation

12:33 &lt;walkerma&gt; So, Physchim62, do you think this bot approach is going to work?

12:34 &lt;+Physchim62&gt; walkerma, define "work" :P

12:35 &lt;walkerma&gt; Sorry, I failed my third year thermodynamics course!

12:35 &lt;walkerma&gt; However

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

12:36 &lt;+Rifleman_82&gt; work is the consumption of energy

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

12:36 &lt;walkerma&gt; I mean at a technical level - ignoring the politics

12:36 &lt;walkerma&gt; And also in terms of people keeping track of things,and setting up the pages, etc

12:37 &lt;walkerma&gt; people = us

12:40 &lt;+Physchim62&gt; at a political level, we need to get the idea of indices approved

12:40 &lt;walkerma&gt; indices = ?

12:41 &lt;walkerma&gt; Do you mean a page of data?

12:41 &lt;+Physchim62&gt; that we have a second copy of WP data in a more secure location

12:41 &lt;walkerma&gt; OK

12:43 &lt;+Physchim62&gt; "indices" could be CASRNs, or InChIs, or other ChemBox data, but we will get nowhere without permission to keep them (not in the main MW database structure, which would be refused, but publically available and frequently updated)

12:43 &lt;+Physchim62&gt; it is the automatic updating for which we need permission

12:44 &lt;walkerma&gt; Yes, I'd assumed that the auto-updating would be the discussion point

12:44 &lt;walkerma&gt; And Dirk has carefully avoided the need for that

12:44 &lt;walkerma&gt; so far

12:46 &lt;walkerma&gt; Do you want to comment more on the bot, or should we move on?

12:50 &lt;walkerma&gt; Now, as for substructure searching - do you think we should go ahead and do this? Have you had a chance to look at the Powerpoint?

12:52 &lt;walkerma&gt; Physchim62, Rifleman_82, are you there?

12:53 &lt;+Rifleman_82&gt; yup here

12:53 &lt;+Physchim62&gt; yes, I'm still here, I was just checking that nothing escaped me on the BAG discussion

12:53 &lt;walkerma&gt; Nothing ever escapes you, PC!

12:53 &lt;walkerma&gt; OK, ChemSpiderMan just needs to spend an hour or so on writing code, and we should have a page on ChemSpider where you can do the following search (shown as a three step process):

12:53 &lt;+Physchim62&gt; I can think of one or two young women...

12:54 &lt;walkerma&gt; 1. Draw a structure

12:54 &lt;walkerma&gt; 2. Click on "Search Wikipedia"

12:54 &lt;walkerma&gt; 3. You find the page on WP corresponding to that structure.

12:54 &lt;+Rifleman_82&gt; using chemspider as an external search engine to wikipedia

12:54 &lt;walkerma&gt; Exactly

12:55 &lt;+Physchim62&gt; yes, I knew CSM was very close to doing that.

12:55 &lt;+Rifleman_82&gt; i don't see wp growing a pop-up drawing window for substructure search...

12:55 &lt;+Rifleman_82&gt; so i guess that's the best we can do

12:55 &lt;+Physchim62&gt; there is simply NO WAY the devs are going to give us structure searching in MediaWiki, so off-site searching is our only alternative

12:57 &lt;walkerma&gt; Yes, you're right - but this feature could be VERY powerful and useful for us, and so this offsite version of the search seems to be (IMHO) the next best thing

12:57 &lt;+Rifleman_82&gt; chemspider's latest integration with wp has shown me that this sort of mashup can be useful

12:57 &lt;+Rifleman_82&gt; i initially was somewhat doubtful

12:57 &lt;+Physchim62&gt; I would still like us to continue our efforts to publish InChIKey and CASRN indexes.

12:58 &lt;walkerma&gt; Rifleman_82: I agree - the web is all about mashups and collaboration these days - we would all benefit from this, I think

12:58 &lt;+Rifleman_82&gt; i had my own discussions with tony, and he showed me how with free access & chemspider & wp, one with literally nothing to do could search literature, create chemicals articles, and create chemspider entries

12:58 &lt;+Physchim62&gt; our aim should be to have our data accessible in as many ways as is reasonably possible

12:59 &lt;+Rifleman_82&gt; and without any institutional access to journals!

12:59 &lt;+Rifleman_82&gt; you think the devs will let us use molfiles by default?

13:00 &lt;+Physchim62&gt; Rifleman_82, we would have to make a very strong case

13:00 &lt;walkerma&gt; Rifleman_82: Please explain what you mean

13:00 &lt;+Rifleman_82&gt; instead of puzzling about how best to depict a nontrivial molecule

13:00 &lt;+Rifleman_82&gt; you upload it as a .mol

13:01 &lt;+Rifleman_82&gt; and users can zoom in and out all and turn around all they want, in 3D

13:01 &lt;walkerma&gt; I didn't think molfiles were compatible with the Wikimedia software - or am I wrong?

13:01 &lt;+Rifleman_82&gt; and yeah, that solves those pesky problems with formula and molwts

13:01 &lt;+Physchim62&gt; walkerma, they're not at the minute

13:01 &lt;+Rifleman_82&gt; ah, some day

13:01 &lt;+Rifleman_82&gt; :)

13:02 &lt;walkerma&gt; I remember Bduke was spearheading a group trying to get Jmol integrated into the software

13:02 &lt;+Physchim62&gt; I think, for the moment, the devs will say NO, but who knows in three years' time...

13:03 &lt;+Rifleman_82&gt; i don't suppose the inchi can automagically generate the molecule on the fly?

13:03 &lt;+Physchim62&gt; in principle, yes

13:03 &lt;walkerma&gt; Rifleman_82, it can, but only if the necessary script is in place

13:03 &lt;walkerma&gt; (as I understand it)

13:03 &lt;+Physchim62&gt; and the scripts are not always wonderful

13:03 &lt;+Rifleman_82&gt; okay...

13:04 &lt;+Physchim62&gt; see my discussions a couple of months ago with Antony about natural products

13:04 &lt;+Rifleman_82&gt; ok anyway, don't let me hold sidetrack you

13:04 &lt;+Physchim62&gt; yes, I'm going to have to love you and leave you

13:05 &lt;walkerma&gt; With a few hours more work, ChemSpiderMan believes we can do SUBSTRUCTURE searching! I think that would be amazing

13:05 &lt;walkerma&gt; even if limited

13:05 &lt;walkerma&gt; Physchim62: Great to talk with you - and good luck with the new job

13:05 &lt;+Physchim62&gt; if we could link that into a bot, then we have a solution for classification of organic compounds

13:06 &lt;+Rifleman_82&gt; ah yeah, that's true

13:06 &lt;+Rifleman_82&gt; never thought about that

13:06 &lt;walkerma&gt; FYI: http://en.wikipedia.org/wiki/Wikipedia:Using_Jmol_to_display_molecular_models

13:06 &lt;+Physchim62&gt; Thanks, sorry for the silence, the new job is really hard, so hard that it's taking me to Venice tomorrow :P

13:07 &lt;walkerma&gt; Oh, how awful!

13:07 &lt;+Physchim62&gt; Venice, Italy, that is, the one with the gondolas

13:07 &lt;+Physchim62&gt; I know, the sacrifices I have to make... ;)

13:08 &lt;+Rifleman_82&gt; :)

13:08 &lt;walkerma&gt; So, should we ask ChemSpiderMan to go ahead and do this structure searching idea?

13:09 &lt;+Physchim62&gt; errm, yes!

13:09 -!- Physchim62 [n=Physchim@unaffiliated/physchim62] has quit ["What did you say this button does?"]

13:11 &lt;walkerma&gt; It could be set up to use InChIs or InChIKeys, and so of course it would require these to be in the target WP article - but once we start uploading our validated content we'll have many more articles containing InChIs and InChIKeys.

13:12 &lt;walkerma&gt; Rifleman_82: Do you support the idea of getting CSM to do this?

13:15 &lt;+Rifleman_82&gt; sorry, was distracted

13:15 &lt;+Rifleman_82&gt; yes, definitely

13:15 &lt;+Rifleman_82&gt; just to check - inchis are unique?

13:16 &lt;+Rifleman_82&gt; as in, absolutely utterly unique?

13:16 &lt;walkerma&gt; InChIs are utterly unique

13:16 &lt;+Rifleman_82&gt; i understand they should be, just want to make sure

13:16 &lt;+Rifleman_82&gt; okay

13:16 &lt;+Rifleman_82&gt; must be problem with my software then

13:16 &lt;+Rifleman_82&gt; it kept flipping cis trans bonds

13:17 &lt;walkerma&gt; CSM explained that often it's because of the way you draw something

13:17 &lt;+Rifleman_82&gt; i just took the inchi from pubchem

13:17 &lt;walkerma&gt; It may LOOK like you drew the methyl in front, but unless you use a wedge the computer won't understand that

13:17 &lt;walkerma&gt; Aha - Pubchem is horribly unreliable

13:17 &lt;+Rifleman_82&gt; ah okay

13:18 &lt;walkerma&gt; InChI is layered - you can have "stereochem unspecified" versions, or you can have "this bond must be cis"

13:18 &lt;walkerma&gt; Whoever generated the InChI for pubchem must've left the stereochem unspecified

13:19 &lt;+Rifleman_82&gt; oic

13:19 &lt;walkerma&gt; We have the exact same problem on WP, according to CSM

13:19 &lt;+Rifleman_82&gt; okay then

13:20 &lt;+Rifleman_82&gt; half the time we're copying the InChIs from someone else

13:20 &lt;walkerma&gt; (the problem with the InChIs we have on WP)

13:20 &lt;+Rifleman_82&gt; but yeah, he should go ahead with structure searching

13:20 &lt;walkerma&gt; I think the validation process will remove a lot of those problem InChIs

13:21 &lt;+Rifleman_82&gt; okay

13:21 &lt;+Rifleman_82&gt; we're depending very heavily on tony's work

13:21 &lt;walkerma&gt; I saw in watchlist the other day, some anon edited an InChI, and someone from WP:Pharm immediately reverted it saying, "No, Pubchem says this!" Maybe the anon was right!

13:22 &lt;walkerma&gt; Rifleman_82: Yes, we are, and of course Tony may move onto other things at some point

13:23 &lt;walkerma&gt; However, I think (a) his motives are entirely honourable and (b) we need to grab these opportunities while we have them

13:24 &lt;+Rifleman_82&gt; if this thing takes off, more incentive for him to carry on

13:24 &lt;walkerma&gt; and (c) it's very possible that ChemSpider could become a major player in the world of chemical information

13:24 &lt;+Rifleman_82&gt; maybe advertizing will pay enough for him to live?

13:24 &lt;+Rifleman_82&gt; yeah

13:24 &lt;+Rifleman_82&gt; advertising

13:24 &lt;+Rifleman_82&gt; sheesh

13:25 &lt;+Rifleman_82&gt; was quite disappointed nobody commented for or against putting chemspider IDs on the chemboxes

13:25 &lt;walkerma&gt; It will take a long time before he'll get rich of advertising, but he has a lively business in consulting. I'm not surprised- look at the guy's experience and expertise!

13:27 &lt;walkerma&gt; Regarding putting ChemSpiderIDs on the chemboxes, I will comment on that - I was just immersed in writing the fall 2008 lab manual!

--- Log closed Tue Jul 29 13:27:03 EDT 2008