Wiktionary talk:Entry layout explained/archive 2002

Definition from Wiktionary, the free dictionary
Jump to: navigation, search

The current layout won't work as the etymology/thesaurus/quotation section could be different for each subdefinition.

One way to get around this is to use disambiguation like Wikipedia, or just by numbering separate articles so we would have dog(1), dog(2), dog(3) with the article dog pointing to all of these. --Imran 11:09 Dec 13, 2002 (UTC)

So, the basic definitions would be left on the main page but the etymology and usage examples would be on diambiguation pages?
Yes --Imran
That should work but for the most common usage, or words with very concrete etymology, should the entry remain on the main article? -- Tobes
Etymology could vary for different meaning of a word, I can't think of any English examples off the top of my head, but in Arabic the word for philosopher is sufi from the Greek suf (Wisdom), but the Arabic word for mystic is also sufi but it comes from an early arabic word rather than from the Greek root. So I think anything but the basic def. should be off in its own page. --Imran 00:48 Dec 14, 2002 (UTC)

I think that we should go for some form of standarized markup system for the individual sections (etymology, spellings, et al) as a dictionary will neccessarily be more standardized than an encyclopedia, it would also allow on the fly conversions of formatting.--Imran 11:09 Dec 13, 2002 (UTC)

Are you suggesting using something like XML (or a wiki equivalent), or simply standardizing entries using the existing syntax? XML is theoretically more attractive, especially when dealing with large numbers of entries, but it's not very practical for having random people create entries. If we stick to a simple template with headers and ordered lists, it should be fairly straightforward to pick out the individual sections with Perl or some other data mining tool if the need arises. -- Merphant
As long as we follow a standard it shouldn't matter how we do it. --Imran 00:48 Dec 14, 2002 (UTC)
I think at the moment it would be good to evolve a comprehensive template using the existing syntax and stick to it. Maybe we could put a link to the template specification in the edit page to make sure that it is followed. Perhaps later we could simplify things by modifying the edit page to provide different boxes for noun,verb,synonym,translation etc? Arvindn 11:43 Dec 20, 2002 (UTC)

Moving most of the Template article here leaving only the currently proposed template.

The ultimate goal of this page is to create a standard Wiktionary entry template that is flexible enough to handle multiple definitions, etymologies, synonyms, etc. and can thus be used for all Wiktionary entries. Since Wiktionary is so new, most entries tend to evolve on their own anyway, but please use this page as a guide. If you see something in a Wiktionary entry that works better than what is here, please incorporate it into this article.

Part of speech, e.g. noun

[Syllable1-Syllable2] pronunciation1 / pronunciation2

1. Definition text goes here.
Synonyms: synonym1, synonym2, synonym3.
Antonyms: antonym1, antonym2, antonym3.
2. Another definition, still a noun.
Synonyms: etc.

A different part of speech, e.g. verb

[Syllable1-Syllable2] pronunciation3 (if necessary)

3. Defintion for the verb; synonyms, as above.

== Etymology == Etymology details go here

== Translations ==

  1. Word (language code)-Complications with translation...
  2. Word (language code)

Thoughts on what to include, copied from "Wiktionary:Wiktionary Definition Standards" (defunct):

A definition template could consist of the following:

  • the word, including:
    • Word split up into syllables.
    • alternate spellings
    • plural formation (for nouns)
    • past tense and past participle (for verbs)
    • comparative (for adjectives)
  • derivation or etymology
  • identification of usual part of speech (see the Wikipeda article on many parts of speech of the
  • definitions (one of two orders)
  1. most common meaning to least common
  2. first use to most recent use
  • usage examples, where needed, with each definition
  • synonyms
  • antonyms
  • other related words or relationships


My suggestions

  • Inclusions of the translations at the same level that synonyms (it is more easy for words with a lot of meanings and various categories)
  • Inclusion next to the translations, remarques for translatiosn (e.g : false friends)

* compound noun (e.g : trade union in trade), phrasal verbs, proverbs at the end of the template. * it is not really to be include in the template : languages should be classed alphabeticaly --Youssefsan

I agree with the first two point (esp. the first) but I'm not sure how useful the third will be. Anyway, I've hacked your vision of point one at dog. --Maveric149

One thing that has to be notd is that some words are spelt teh same in diffrent langauges. So possibly a ---- is needed and then a repeate of teh template? -fonzy

Another thing to consider is that words that are spelt the same in a language do not necessarily have the same etymology. For example, the Dutch word riem can mean (1) belt, (2) oar or (3) ream. (1) derives from Old English reoma, (2) from Latin remus and (3) from Arabic rizma. We should think about how to differentiate between them. We could group the meanings of a word which have the same etymology, but it's not really clear to me how to do that in an uncomplicated way. Ideas? D.D. 11:46 Dec 14, 2002 (UTC)

See what I've said at [[Talk:Wiktionary:Entry layout explained]] --Imran 12:03 Dec 14, 2002 (UTC)

Ok, thanks. I must have overlooked it :-) D.D. 12:09 Dec 14, 2002 (UTC)

See also portmanteau and talk:portmanteau for the following:

AFAIK, fr: and nl: do not have a separate adjective for portmanteau in the linguistic sense. Mot-valise and mengwoord are both nouns and mean portmanteau word. To write "nl: mengwoord (adj)" and "fr: mot-valise (adj)" is not correct. I've tried to clarify that after nl:. Any comments? D.D. 12:35 Dec 14, 2002 (UTC)

Wouldn't it be easier in this respect to number the meanings through without starting again at a new word class? D.D. 12:40 Dec 14, 2002 (UTC)

What about flexional forms in languages with flexion? Or every langauge should develope its own template? I suggest to put it right beneath the spelling, unless flexional forms depend on definition (like in some rare cases in Polish). --Youandme 22:12 Dec 14, 2002 (UTC)

Copied mainly from the talk page with some mods,

The current layout of different defintions of the same word won't work as the pronunciation/etymology/thesaurus/quotation and basically every section could be different for each alternative definition.

So I propose a system of disambiguation, numbering seperate articles so we would have dog (noun-1), dog (noun-2), dog (verb-1) with the article dog pointing to all of these including a short definition of each one. --Imran 23:56 Dec 14, 2002 (UTC)

This seems reasonable and would lead to better and easier to follow entries with clear and easy to use ==Synonyms==, ==Antonyms==, ==Etymology== and ==Translations== sections. But I am worried about long-term maintenance of the resulting articles. Isn't there a hierarchy of most common to least common definitions in any definition list? Suppose a person makes an entry with three definitions in summary on it. Each of these link to [word (noun-1)], [word (noun-2)] and [word (verb-1)]. But then somebody else comes along and sees that the ordering of the nouns is backwards and that there are three more common defintions for the verb that should go before verb 1. Thus pages would need to be moved. Then somebody else comes along in 10 years and corrects the ordering of the list again to reflect new usage (placing noun 3 in the number one spot). Such is the case when you hard code hierarchical information in page titles. A better system is needed but I can't think of one right now. But you do have the right idea. --mav
Maybe we should abandon the hierarchy by commoness, as in many cases it's going to be rather arbitary anyway and is likely to lead to pointless arguments and editwars.
I thought about wikipedia style disabmbiguation, but for some words that could cause problems, for instance differentiating between the two meanings of nonhumanoid (as non-humanoid and nonhuman-oid, two similar but different meanings).
I think perhaps we should arbitarily number them with out noun/verb/etc as someone pointed out a word can be more than one of these--Imran 00:28 Dec 15, 2002 (UTC)

I've taken my proposed system and tried it out on Portmanteau, I've also used a standarized layout on each of the definitions which hopefully is easy to understand/use while at the same time being logical/simple enough to be automatically parsed for information by software. What do people think --Imran 11:43 Dec 16, 2002 (UTC)

IMHO, numbered entries like that would be a navigation and maintenance nightmare. Arbitrary numbering is extremely unsatisfying, and meaningful ordering is asking for people to try to change the order if they think it's not correct... Better would be categorical disambiguation: Portmanteau (linguistics), Portmanteau (baggage), Portmanteau (collaboration) if you simply must separate them that aggressively. I suspect though that in most cases it's of little help to separate them that much, or even counterproductive -- you have multiple copies of pronunciation keys and etymologies to keep in sync, and it's more work as a reader to browse if you've found yourself at the wrong one or just want an overview of the various uses of a word. --Brion 11:57 Dec 16, 2002 (UTC)
On the first point, arbitary number is pretty much the only way to do it and is pretty much the standard way for most dictionaries, and we might as well have a rule saying if the articles already there then there is no point in moving it. On the second as I said above wikipedia style disambiguation would be nearly impossible for words like nonhumanoid which have two very similar but different meanings. On the third, two words that are spelt the same way can be pronounced differently for different definitions and have different etymologies. On the fourth for an overview you can just look at the main article, Portmanteau in this case, which will give you the basic definition. If you can come up with a better way to deal with these problems please do, as I agree this system of break-up isn't perfect, but it's the best way I can think of to deal with these problems. --Imran 12:10 Dec 16, 2002 (UTC)
You gave the answer yourself with the example... Nonhumanoid (non-humanoid) and Nonhumanoid (nonhuman-oid) are your disambiguations. --Brion
It's not an entirely satifactory solution, as some of the disambiguations will be on topic, others a break up of a word as you give, and others by the tense of the word (for instance the word read), I don't think either solution is satisfactory but I'll be happy to go along with whatever is most popular. --Imran 21:13 Dec 16, 2002 (UTC)
An alternative idea, how about using the date of the earliest quotation for disambiguation ?, that way we have a standardized form of disambiguation across all articles, it doesn't impose an artificial ordering on the main word page and contains some information about the word in question. Perhaps also with language for non-english words, so for instance we would have Pope Julius (1521) or Klein (1530-de)--Imran 11:47 Dec 17, 2002 (UTC)

A proposed template for articles:

An article should include one main definition, with optional sub-definition for words that derive from the main word. So for instance the article Walk would have the main definition of walk, but also have sub-definitions for walking, walked, walker, etc.

An article should have some of these sections (in this order):

==Alternative spellings==
==Derived words==

(language) should be the full form ISO name of that language in italics, so it would be (english), (french), etc.

As most of the rest of the sections could contain different entries for the sub-definitions then as to the definition the word being defined should be stated in each section being preceded by a semi-colon (;).

So for instance for Walk and Walked under definition would be,

;Walk (verb)
: Move using feet from one location to another at a relaxed rate.
;Walker (verb)
: Someone who walks.

Which comes out as,

Walk (verb)
Move using feet from one location to another at a relaxed rate.
Walker (verb)
Someone who walks.

Pretty much the same applies for all the sections.

The pronunciation should be laid out like the following,

:(UK) SAMPA:/pO:t'm{nt@u/, IPA:/pɔ:t'mantəʊ/. 

With the country {(UK), (US), (AUS), et al} first, followed by the pronunciation system (SAMPA/IPA) a colon and then the pronunciation between the slashes.

For quotations instead of using a ":" to indent the section use a bullet point "*" before the date/reference and put the quotation on the next line preceeded by a "*:". See Portmanteau, 1. for an example.

--Imran 12:05 Dec 17, 2002 (UTC)

Good work, but can you try the template out on something, somethig which an show teh full extent of the template.

I've made an example with Leap it still needs a bit of filling out (especially the pronunciation as I don't know how to do the IPA/SAMPA stuff), but most of the principles are there. One of the things I was keeping in mind while making the template was to make something which would be fairly easy for a computer to parse so that if sometime in the future we wanted to change template it could be automated fairly easily. --Imran 15:39 Dec 17, 2002 (UTC)
Does anyone object to the above format, if not I'll start converting articles to it ?
Why not try doing a version of abate (from the 1913 Webster's) or dog (based loosely on a structure inspired by WordNet)?

The format suggested in portmanteau is very undesirable. All the major definitions of a word need to be on the same page unless it is just too complicated. Fred Bauder 17:29 Dec 18, 2002 (UTC)

All the basic definitions are on the same page, it's only the more detailed pages that are seperate. Could you clarify what you mean ? --Imran 22:03 Dec 18, 2002 (UTC)

The basic definitions are there but very stripped down. I want to see the whole thing on one page (at least for simple words like portmanteau). A word like the would, of course, require several pages. I also don't like the way the page is spaced with so many empty lines. I like the format in dog and pumpkin much better. Fred Bauder 01:27 Dec 19, 2002 (UTC)

Are you reffering to the fact that the definitions given on the portmanteau page are shorter then the definitions on the linked pages, if so that's only because I wrote that page in a hurry, I have no objection to disambiguation page listing full definitions as given on the disambiguated pages.
As discussed further up on this page we can't really have anythin but the definitions sharing a page, as virtually every sections could be different and misleading. For instance if someone wanted to call a german person Pumpkin under definition 4, and used the suggested translation Kürbis it's likely to cause some confusion...
On the issue of spacing thats due to the way wikipedia is processing the various headings and not due to artificially added blank lines, it shouldn't be too hard to change the code so a users preferences can change the number of empty lines --Imran

I'm not happy with having article(quotedate) to me it would look like the date the word came about.

We could always have it as the date of the earliest known quote, in which case it would be the earliest attestation (known to us).

Also you can have a simple ---- then the word again. That would be good in my opinion.

Which bit are you talking about (or demonstrate on the portmanteau page) ? --Imran 21:09 Dec 20, 2002 (UTC)

yes, but i think doing that is too messy. I prefer having all meanings of teh word on one page. -fonzy

I've no real problem about them being on the same page but there are a few reasons for seperating them, (1) Shorter pages would make reading/editing articles easier especially for people with slower internet connections (2) It allows direct linking to individual definitions. (3) Casual browsers have a tendency not to scroll a webpage so might look at the first definition and leave without noticing the other defs. (4) The majority of people will probably only want defs which should be grouped together. --Imran 22:22 Dec 20, 2002 (UTC)

After looking more deeply into the Portmanteau material, I became more than ever convinced that a premature splitting of articles is the wrong way to go. There is no doubt that some articles will need splitting when they get too long. By that time the basis for splitting will become obvious in many cases. Articles that are less than 32K in length seem to function quite well in the software as long as they are not just continuous masses of links to other articles. There's no rush. In Portmanteau (1500) I questioned the date of the first usage. From what I could find out it should be 1584. If my findings are corroborated that would necessitate a change in the article title to Portmanteau (1584). If that's the case I expect that similar problems will arise more frequently than we want. That will mean more work for everybody, which is great ..... if we're getting paid.

There is little doubt that Wiktionary's formats will need to be more strongly hierarchized than Wikipedia's. Given the very positive multilingual visions that many of us have, it would be tragic if those visions were shattered because of an irretrievably inappropriate, inadequate or inflexible structure.

I plan to propose hierarchical guidelines which will reflect much of the discussion which has taken place. Eclecticology 06:49 Dec 24, 2002 (UTC) See also User:Eclecticology/Vision

Further to the above I will be suggesting an order for topics about a word. This follows the principle that the more general topics and those least likely to fork should come first. Thus I would provisionally propose the following order

  1. Alternative and obsolete spellings
  2. Etymology
  3. Pronunciation
  4. Part of Speech
  5. Definition
  6. Quotations
  7. Derived words and phrases
  8. Synonyms
  9. Translations
  10. References

Detailed reasons for choosing this order will be developed on the vision page noted above. Eclecticology 22:24 Dec 26, 2002 (UTC)