Talk:OmegaWiki decisions on its usage

No Redirects[edit]

I quite understand this policy, but it is really reflecting a heuristic within Wiktionary itself rather than a database design issue. In other words, redirects are a mechanism for implementing alternative headwords. What is MUCH more important is whether the Ultimate Wiktionary will cater for alternative headwords to cater for languages like Japanese where there is often more than one valid orthographical form of a word. For example, aikido can be written: 合気道, 合氣道 or あいきどう. All are valid; none are mis-spellings. A person literate in Japanese would recognize them as being the same thing.

There are two ways to implement redirects; either for technical reasons or for reasons of displaying alternate spellings. I am very much opposed to using redirects for alternate spellings. A stub could suffice to indicate alternate spellings. The problem is that expectations are that redirects are valid spelling. As you should only have redirects for exclusively for one reason, I am having redirects at all.

"redirects" and "stubs" are a wiktionary concept; not a database design or content issue. I'm not saying we need either; I'm suggesting that the database allow for multiple orthographical forms. --JimBreen 05:48, 1 August 2005 (UTC)[reply]

Redirects are a Mediawiki concept stubs are a Wikipedia concept. Stubs have their place in old style wiktionary and can indicate an alternate orthographical form. Redirects are bad news. GerardM 06:57, 1 August 2005 (UTC)[reply]

If the Ultimate Wiktionary were to be restricted to a single orthographical form per entry, it would be very difficult for it to cater for Japanese properly. The kana version of a word (e.g. the あいきどう above) could at a pinch be regarded as the pronunciation, but the pronunciation of the kana form can be context-dependent. Something like もう is pronounced one way if it appears as a verb-ending and another way in other situations. I would go so far as to say that a one-orthographcal-form-per-entry design would rule out the Ultimate Wiktionary being really usable with Japanese, and thus it would not be "ultimate" at all. --JimBreen 04:43, 31 July 2005 (UTC)[reply]

You can have the three versions that you point out and Romanji too. GerardM 05:20, 1 August 2005 (UTC)[reply]

In the database? Within the entry? Good. End of debate.

PS: "Romanji" a mis-spelling. The word is "romaji" or "rômaji". It's the Western alphabet transcription of ローマ字. The "roma" or "rôma" is from Italian; not English. --JimBreen 05:48, 1 August 2005 (UTC)[reply]

In the database, yes. In the entry, no. Will they be linked, when someone links them. GerardM 06:58, 1 August 2005 (UTC)[reply]

You've lost me. How can something be in the database but not in the entry? Do you mean the alternative forms will not be viewable in the WWW display of the UW? Or that you can't get to an entry using any of the alternatives? If so it will be close to useless for some languages. It will have problems with English, that's for sure. Can you enlighten me as to how you will handle "centre" and "center" or "accoutrement" and "accouterment" or "sulphur" and "sulfur"? Surely you are not going to put each in its own entry. --JimBreen 08:26, 2 August 2005 (UTC)[reply]

They can when they are destinct entries. They will be their own entry. From a database POV that makes excellent sense. Accoutrement and accouterment are two words with their own history, the idea that they are "the same" is ignoring the difference between the two.

What difference? There is none. "Accouterment" is simply the way some but not all Americans spell "accoutrement". Ditto for centre/center; the difference is purely orthographical. JimBreen 23:34, 2 August 2005 (UTC)[reply]

One spelling is from the French the other is a variation on the English word that is coming from the French. One was used for a first time in 14** and the second in 19** or something. There is more to a word than just what it means. GerardM 05:49, 3 August 2005 (UTC)[reply]

Yes, within a language they may be alternate ways of spelling. The orthography can however be indentified. The database design is quite specific about this; you start with a Language, then you have the Spelling that requires a Lanuguage then you get a Word that requires a Spelling and lastly you get the Meaning which requires a Word. Alternate spellings are related through Relation and the relation is defined by WordRelation. The WordRelation defines if only words of the same language can be targeted or not. GerardM 08:47, 2 August 2005 (UTC)[reply]

I'm sorry, but much of this is simply not making sense to me. It seems you are trying to enforce levels of prescription and rigidity that simply do not apply in natural languages. If you are saying that words like centre and center or 生け花 and 生花 have to have their own entries because of an abstract rule that "a Word [...] requires a Spelling", I think an essential aspect of the UW is being steered in a fundamentally silly direction. It might suit some European languages where there are high levels of prescription, but there are a lot of languages where it will cause nothing but confusion and difficulty. --JimBreen 23:34, 2 August 2005 (UTC)[reply]

You are wrong. As the different spellings of a word are treated equally, no value judgment has passed. The distinct spellings need to be identified as such. Now given the current practices in Wiktionary you often find that one word is chosen as the "main" spelling and the others are either redirected or a stub is created to point to the main word for more information. So in my opinion you are barking up the wrong tree. GerardM 05:49, 3 August 2005 (UTC)[reply]

OK. Just to focus on a single example in one language, can you state how the Japanese word 合気道/合氣道/あいきどう (three valid forms of the one word) would (a) be recorded in the UW database, and (b) appear in a subsequent wiktionary WWW system. If I'm "barking up the wrong tree", it's because the handling of multiple valid forms is not described clearly. --JimBreen 22:22, 3 August 2005 (UTC)[reply]