Wiktionary:Unresolved issues/Anagrams

Previous Discussions: Beer_parlour_archive/2009/August

Anagrams
I'd like to propose some changes to this. Not only does this totally ignore foreign languages, why not allow diacritics? In Scrabble diacritics are always ignored, because there are no tiles that have them! Only a few languages used diacritics for Scrabble - French and Romanian don't, nor does Italian, and Spanish only has ñ, nothing else. Mglovesfun (talk) 15:13, 29 August 2009 (UTC)


 * I'm not responding to your specific points here, but since (accents aside) the definition of an anagram is so algorithmically measurable, I'd like to see this automated. I realise it's a lot of work for something of no use to most readers, though. In the same way that the alphabetical index is periodically generated, perhaps someone could look into a process that determines which (newly added or deleted) words are anagrams of others...? Equinox ◑ 15:39, 29 August 2009 (UTC)


 * In French, . Certainly adding anagrams by hand is pretty futile, yeah. On fr.wikt we consider that eéêèëEÉ (etc.) are all e and iîïiI (etc.) are all just i, in terms of anagrams that is. So pâté is a perfectly acceptable anagram of tape. Mglovesfun (talk) 16:06, 29 August 2009 (UTC)


 * It would make sense to not add anagrams directly to entries, either to have a category for each one, Category:English anagrams of aenv, or (I prefer) a template that is included on each entry, . It makes sense, in this case, for the title to be included in the template, as that way the edit link will point to the correct place. I have created that template and added it to the two anagrams death and hated. While creating these templates and adding them to entries with a bot is very doable, it is not totally trivial and we need to work out (probably on a per-language basis) what to do with diacritics, punctuation (and clicks), how to deal with Mapudungun (and any other language that has two seperate writing systems using the same set of letters), what to do with multi-glyph letters (does the Hungarian cs just count as c + s), and whether the phrases created must be dictionary entries (i.e. should "the da", and "Ed hat" also appear on . Given the large number of entries that may have anagrams, it seems to me that we should ammend the WT:ELE example of a vertical list for them and have a horizontal list instead (I doubt there are many words for which this list is huge). I might start having a go at doing this with User:Conrad.Bot, given that I already have the word lists, though I will wait for further comments and a VOTE before editing in earnest. Conrad.Irwin 16:06, 30 August 2009 (UTC)
 * About diacritics and multi-gryph letters for anagrams, no general rule can be defined: in French, diacritics are traditionally ignored for anagram purposes, but the tradition might be different in other languages.
 * About phrases, I would include everything included in the Wiktionary, e.g. bien sûr but not Ed hat. I would make an exception for famous anagrams (e.g. un veto corse la finira for Révolution française, but without linking this anagram sentence, of course). Lmaltier 16:19, 30 August 2009 (UTC)


 * Other things to not include are misspelling entries, and (presumably) entries for Abbreviations, Acronyms and Initialisms? Conrad.Irwin 17:14, 30 August 2009 (UTC)
 * I agree for misspellings, of course. But why not abbreviations, acronyms...? If the reader is not interested, he can skip them... Lmaltier 17:19, 30 August 2009 (UTC)


 * No reason really, I just don't count them as "real words" :). The other issue is for alternatives like cafe and café where it is not clear from the entries pages which spelling is preferred. It would seem strange to list them as "anagrams" of each other, but then maybe we'd want to do that for linking purposes. For pages like co-operate and cooperate, the bot can detect the template and not include the alternatives when they use the same letters in the same order. Conrad.Irwin 17:17, 31 August 2009 (UTC)
 * Or links s.v. =Alternative spellings/forms=? In any event, I like the idea of automating this. &#x200b;—  msh210  ℠  18:21, 31 August 2009 (UTC)
 * Anagrams imply a different order. café is not an anagram of cafe. Lmaltier 18:26, 31 August 2009 (UTC)


 * Ok, how do (death, Death, hated),   (cafe, café, face) and  (eft, EFT, FET) look?  Conrad.Irwin 01:06, 4 September 2009 (UTC)

At fr.wikt, we use a single line when several words are the "same" anagram. An example in fr:écran: Lmaltier 05:27, 5 September 2009 (UTC)
 * ancre, ancré
 * caner
 * carne, carné
 * cerna
 * crâne, crâné
 * créna
 * encra
 * nacre, nacré
 * rance


 * I've now done that, the only issue remaining with my implementation (for English anyway) is that it lists theatres of war and theaters of war as anagrams (it excludes theater of war and theatre of war as they are clearly marked as alternatives). I presume this isn't too much of a problem, and the system has the ability to be manually overwritten (using the templates and  in the created templates). Any other suggestions? Conrad.Irwin 23:10, 10 September 2009 (UTC)
 * Why considering that theatre and theater are not anagrams (the letters are the same, in a different order)? Is the rule traditional for English anagrams? Lmaltier 21:25, 12 September 2009 (UTC)


 * They are anagrams, even though they are different spellings of the same word. I think we should certainly have them listed as such. Equinox ◑ 21:44, 12 September 2009 (UTC)


 * Since this is, hopefully, going to be added to the majority of English entries, I'd prefer switching to a horizontal layout. We could use parenthesis to group words differing only by diacritics. So Lmaltier's example would be " ". The word left before the parenthesis could be the one that would come first in an alphabetical sort. This would be the form w/o diacritics, if one exists. Is there much support for this? It would eventually take a vote to change the WT:ELE. --Bequw → ¢ • τ 19:03, 12 September 2009 (UTC)
 * Ok, I'll just include them then, there are around 75000 templates I could create, so yes they'd be on most entries. The format could be changed to the one you describe just by editing . Conrad.Irwin 22:08, 12 September 2009 (UTC)


 * Please indicate your opinion at Votes. Conrad.Irwin 22:25, 13 September 2009 (UTC)


 * This has now started. If you notice problems, please let me know ASAP. It will take it about a week to do English. Conrad.Irwin 14:50, 26 September 2009 (UTC)

Wiktionary:Votes/bt-2009-09/User:Conrad.Bot to do anagrams - Vote passes.

Beer_parlour

Anagram layout
As noted by Bequw, the anagrams section of tesla was fairly monstrous, I've updated it to put everything on the same line, which uses about 5% of the amount of screen space. As this change goes against the example given in WT:ELE (though not against the text), I thought I'd ask your permission before updating Conrad.Bot to use the new horizontal format henceforth (old anagrams sections won't be updated unless they need anagrams added or removed). Are there any objections? Conrad.Irwin 02:28, 17 December 2009 (UTC)
 * The "eveything on the same line" link above should be to http://en.wiktionary.org/?oldid=8078582, not what it links to now. &#x200b;— msh210 ℠ 02:59, 17 December 2009 (UTC)
 * I prefer the one-line way. It does seem to violate ELE, but I, for one, am willing to overlook that, or change ELE. &#x200b;— msh210 ℠ 02:59, 17 December 2009 (UTC)
 * I think they should be kept vertical unless it's longer than five or so. We have the same situation we the synonyms/antonyms lists, that they look a little ridiculous when they get really long. I don't think there should be a bot running in violation of ELE, so I think it might be a good idea to hold a vote to modify ELE. --Yair rand 03:11, 17 December 2009 (UTC)
 * From memory, synonyms are supposed to be in horizontal lists, one per sense. I did think about the "up to five" idea, but it seems fairly arbitrary; a word with five synonyms has five lines, a word with six has one; if the word with six was still on five lines it would feel more consistent, but then it wouldn't make much sense. Conrad.Irwin 13:34, 17 December 2009 (UTC)


 * Maybe the vote can trim the Anagram section of the ELE a bit, and have the majority at Anagrams, that way if we want to make minor layout changes a vote won't be required. --Bequw → ¢ • τ 17:03, 17 December 2009 (UTC)


 * What if we did like this for large sections?:


 * least
 * leats
 * salet
 * setal
 * slate
 * stale
 * steal
 * stela
 * taels
 * tales
 * teals


 * The one-line proposal is fine too. I'm not fussy about the arbitrariness of what defines a "long section", personally. —Internoob (Disc.•Cont.) 03:24, 18 December 2009 (UTC)


 * I don't think that's worthwhile. I would prefer we be fully consistent, rather than collapsing some sections and not others, and the vast majority of anagram sections aren't long.  We could use visible columns, if needed. --EncycloPetey 04:21, 19 December 2009 (UTC)


 * least
 * leats
 * salet
 * setal
 * slate
 * slate


 * stale
 * steal
 * stela
 * taels
 * tales
 * teals


 * : least, leats, salet, setal, slate, stale, steal, stela, taels, tales, teals


 * I am still under the impression that all on one line is more aestetic, would people prefer multiple columns? Conrad.Irwin 12:58, 19 December 2009 (UTC)
 * Using one line for all anagrams entries is a possibility with me too, as long as we do it consistently for all entries. --EncycloPetey 13:06, 19 December 2009 (UTC)

Started Votes/pl-2009-12/Modify anagram section of ELE. --Bequw → ¢ • τ 23:33, 19 December 2009 (UTC)


 * I prefer all on one line, as opposed to the two in the vote, but I'm not overly fussed. Conrad.Irwin 23:35, 19 December 2009 (UTC)
 * How would you prefer to show the alphagram (the only reason that I listed two lines instead of one). --Bequw → ¢ • τ 01:13, 23 December 2009 (UTC)


 * As above: Conrad.Irwin 11:33, 23 December 2009 (UTC)


 * : least, leats, salet, setal, slate, stale, steal, stela, taels, tales, teals

Votes/pl-2009-12/Modify anagram section of ELE - Vote passes.

This issue was resolved following the completion of Votes/pl-2009-12/Modify anagram section of ELE, on January 17, 2010.