Talk:Graph canonization

Recent bad page move

I just undid the move from "graph canonization" to "graph canonicalization" performed by Blaisorblade (talk · contribs), who left a speedy deletion tag for an "implausible typo" that led to the old name being deleted by JohnCD (talk · contribs), breaking multiple links from the article and book name spaces. If such a move is to take place, it needs discussion. In any case the speedy deletion was a mistake since that type of speedy deletion is only allowed for recently created names, this article has been here under "canonization" since 2008, and speedy deletions of redirects that are being used in other links should not be performed until the links are fixed. But I would argue that the move itself was also a mistake. Searching in Google scholar, I find about 186 hits for "graph canonization" and only 21 for "graph canonicalization". Based on that, I believe that canonization is the standard term for this operation, and is the name we should keep it under. —David Eppstein (talk) 01:02, 3 February 2011 (UTC)

Re the broken links - my bad, I usually check "What links here" and find nothing, I don't know why I didn't on this occasion when it would have found something. Re "canonization", concur with DE that this is preferable. Any proposal to change it should go via WP:RM. JohnCD (talk) 10:58, 3 February 2011 (UTC)

Hi, my apologies if I misunderstood the rules. Let me explain why I acted on good faith, and I found no mention of your restriction on misnomers being recent. The text which I read mentioned as a possible reason:

"Recently created redirects from implausible typos or misnomers.[...]This criterion also applies to redirects created as a result of a page move of pages created with an implausible title." [[1]]

My current opinion is that the original (and current) title is a common enough misnomer, so that I wouldn't propose again removing the redirect.

Additionally, I expected links to the redirect to be fixed by a bot (http://en.wikipedia.org/wiki/Help:Moving_a_page), before speedy deletion was applied. I should have at least commented on that.

About the name and Google Scholar entries, the same searches without quotes gives the reverse results, with 19200 hits for graph canonicalization and 2330 for graph canonization. However, I based my results on the same query on plain Google, giving ~6 millions against ~14500 results.

Anyway, my biggest reason for the move is that canonization refers to creation of saints, while the discussed operation refers to "canonical form", and the correct derived form is canonicalization, even if also "canonical" comes from the same Greek root; even if canonization were the standard term used for graphs, unlike in other domains, I'd still call that an incorrect formation of a word, and the hit counts I mentioned made me believe that this feeling was shared. Depending on the exact Google query, one can also get numbers to support your thesis.

I'll try to follow up through WP:RM as suggested, linking to this discussion. Thanks! --Blaisorblade (talk) 09:37, 4 February 2011 (UTC)

I certainly believe you both acted in good faith; that doesn't prevent occasional mistakes, though. I'd think the fact that two of the references have "canonization" in their title, and none have "canonizalization", might also be seen as evidence that canonization is the more standard term. For technical subjects like this one, Google scholar and more precise queries are likely to be more accurate; plain Google without quotes is going to return far too many hits that have nothing to do with the subject. Also you need to be careful in the searches because of Google's ability to return imprecise matches: unless you put it in double-quotes, a search for "canonicalization" will return a lot of hits that are actually for the different word "canonical". As to whether the term that actually is being used is the same as the term that in some ideal world should be used, that's irrelevant. —David Eppstein (talk) 16:26, 4 February 2011 (UTC)

Canonization vs. Canonicalization

There seems to be a weird mismatch between Canonicalization, Canonization_(disambiguation) and this page's name. The other pages explicitly say that "canonicalization" is not to be confused with "canonization", while this page's name is canonization and it seems the above move was reversed quite a while ago. As a non-native english speaker i'm confused. Which one is correct? If both are possible, then maybe the other two pages should be slightly adapted to not generate the impression that this is wrong. Joern (talk) 14:02, 14 May 2015 (UTC)

Maybe your questions would be answered if you read the talk page section immediately above the one you just created? —David Eppstein (talk) 15:46, 14 May 2015 (UTC)

I agree with David that both canonization and canonicalization appear to be used interchangeably by graph theorists, especially non-native speakers. Cheminformaticians, on the other hand, appear to be far stricter on their usage. All three of the cheminformatics citations use canonicalization, and on google searching for "smiles canonicalization" and "inchi canonicalization" both report significantly more hits than "smiles canonization" or "inchi canonization" respectively. Additionally, wiktionary doesn't believe there's any ambiguity between canonization and canonicalization. Surely the fact that there exists a canonization_(disambiguation) but there isn't a canonicalization_(disambiguation) page suggests one might be more appropriate than the other in technical/scientific communication? BiomolecularGraphics4All (talk) 14:21, 30 January 2016 (UTC)

We're not here to judge whether one might be more appropriate for researchers to use, but only to record what they actually do use. Currently we have two references with "canonization" in the title and two with "canonicalization" in the title. Perhaps this is a disciplinary issue, with mathematicians and theoretical computer scientists using "canonization" and chemists using "canonicalization"? In any case since both terms are in use we should record both in our article. "Canonicalization" sounds unnecessarily sesquipedalian to my sensibilities, but again that's just opinion. —David Eppstein (talk) 18:31, 30 January 2016 (UTC)

Canonicalization of Unlabeled Bipartite Graphs

This is a different problem, but related. Seems to me (biased of course) like it is worth at least a link? Find a canonical representation for graphs that lead to the same dynamics in the sense of ion channel kinetics. It is solved but there are still some open questions. See http://www.pnas.org/content/102/18/6326.abstract. -WJB 97.123.137.129 (talk) 22:00, 4 January 2017 (UTC)