Talk:Modular Wiktionary

From Meta, a Wikimedia project coordination wiki

Are you sure that definitions are best attached to individual words and not concepts? Also, I don't see any reason that a word and its definition have to be in the same language. What if you're making a dictionary for a language with very few native speakers ("Hopi" is the one they always talk about on the listservs) and it's easier to define the words in another language? What if you're defining a word that doesn't have a translation in another language? How will speakers of that language know what your word means (beyond approximation through the concept map). DanKeshet 19:02, 13 Mar 2005 (UTC)

Nevermind regarding the first; it would be chaos if definitions weren't attached to individual words. I still think words should be allowed to be defined in any language, though. We won't need to have an English-language definition for "gato" that says "cat", because it will show up under translations. But if we wanted to have an English-language definition for some word that doesn't have a direct translation, that's great. DanKeshet

What about having a word in English that has no text (or the non-English text) but does have a definition, connected to the same concept as the word in the other language? User:Somervillain

In the ultimate wiktionary, definitions will be language specific. They do not translate. Definitions are linked to each meaning of a word. These are concepts. When an English definition is missing, it is missing and someone is needed to add it. A definition defines a concept and as such it also defines words in other languages. GerardM 01:24, 21 Mar 2005 (UTC)
See comments below, but the way that the Modular Wiktionary works is that one concept maps to many sub-entries (senses of words) in specific languages. Each sub-entry has a language-specific definition and is linked to a (language-independent) concept.
A word in a language of a particular wordtype may have several meanings and each meaning maps to a concept, so I have the impression that there is not much different between modular and ultimate wiktionary. GerardM 06:37, 22 Mar 2005 (UTC)
Yes and no. Yes, our visions are really similar. That is why we consider Modular Wiktionary to be a specification for an implementation of the Ultimate Wiktionary. Regarding differences: 1)Unless you have more ideas that haven't been posted on-wiki, I don't have a full understanding of how you propose to implement ultimate wiktionary, so I can't really speak to whether there are differences. The "Word Forms/Paradigms" area doesn't have an analogue in the original ultimate wiktionary proposal and is implemented significantly different than the Vortaro proposal. The relationships between Concepts we propose seems different than the ones you propose, but I'm not sure, because I haven't seen a full specification.
I think I speak for Somervillain when I say that the reason we chose to write this on a separate page, and not just edit "Tables for Wiktionary" is because we needed a sandbox where we could edit without stomping on what was already there. The reason we write it at all is that we believe this is a good implementation of the Ultimate Wiktionary, and we hope to help get these ideas implemented in the Ultimate Wiktionary. DanKeshet 17:14, 24 Mar 2005 (UTC)

Examples[edit]

The examples are no longer valid, because they are all examples of what things will look like after Paradigms are introduced, but the page is now in roadmap format. We should show what examples would look like after each step of the process, and perhaps do a little speculation on how to get things from here to there for each step. After that's done, there's still more to do on this page: mockups, etc. But I would recommend that when the examples are in good shape, we link this page from The ultimate Wiktionary and Tables for Wiktionary. DanKeshet

The examples are only marginally bad, and I think it would add a lot of redundancy to repeat them below. Perhaps just a note that explains that the two offending attributes ("paradigm" and "concept") would actually only be added after those modules are in place. And I agree we should link it from those pages. User:Somervillain

Initial comments on modular wiktionary[edit]

You state that it should be based on the current wiktionary. There is no ONE current wiktionary. One of the major points of the ultimate wiktionary is that it will bring all the current projects together. Technically the Ultimate wiktionary will be more like the nl: or it: wiktionary than any other as these are the more structured wiktionaries.

Ultimate wiktionary will be different from Wordnet because Wordnet is centred on the English language.

There will be no "subdomains". In the user preferences a user can state what his primary language is and, what other languages he/she wants to include in the output. We hope to implement this last feature, it is however not essential to the ultimate wiktionary.

When a word is defined as a translation from one language, a placeholder will be created. This means the existence is recorded, maybe a gender will also be added, but the definition will be initially only available in the language of the word for which it is a translation (if there is a definition).

One thing that we WILL implement first is the GEMET functionality. This will give us things like hierarchy. This will also be available to wiktionary words. eg wife includes female girl includes female or toddler includes child etc. These relations are language dependent. They will NOT automatically inherit from one language to the next. There are languages where there is one word for girl while other languages have many words..

I do agree that there will be a huge culture shock when we get the ultimate wiktionary. The first iteration will be basic and it will be refined over time. There is a definite need to have a XML interface for import and export. I hope we will find people interested in programming these. GerardM 01:17, 21 Mar 2005 (UTC)

Hi Gerard,
Thanks for your thoughts. I think that we are pretty much on the same wavelength. We have conceived of the Modular Wiktionary as an implementation of the ultimate Wiktionary and agree with most of your ideas. I hope that our replies below will clarify our ideas.
  • We agree that all words and definitions in all languages will be in a single, unified database.
  • We agree that interface language will be selectable
  • Re: subdomains: our idea was that subdomains as we know them (separate wiktionaries) will disappear, but en.wiktionary.org could remain as a shortcut to mean "use the English interface to the one Wiktionary".
  • We agree that the languages that a person is interested in, both for definitions and translations, will be selectable.
  • We agree re: WordNet, that wiktionary will be different, because all languages will be in a single, unified database.
  • Re: GEMET and WordNet: GEMET has a hierarchy for organizing words. WordNet has some similar kinds of relationships, but also has other kinds of relationships that GEMET does not have: for example, direct antonyms (hot vs. cold). GEMET's hierarchies make sense for what they're using them for, but WordNet has some relationships that are more useful for classifying adjectives and verbs. Probably, we want to evaluate the types of relationships that both have, then come up with our own set.
  • In the system we propose, a "translation" is when two sub-entries in different languages (say, the "first sense" of the English word "bank" and the "second sense" of the Spanish word "banco") map to the same concept. When you enter a new "translation" through the interface, what actually happens is an entry in the new language is created that is mapped to the same concept. This will have the effect as your "placeholder" definition.
  • re: "refined over time": Our idea for the modular wiktionary was that, independently, each of the steps we outline is a great improvement over what we currently have; together, they form the way toward the ultimate Wiktionary.
  • re: XML import/export: yes, we are in total agreement.
We hope that these points clarify our ideas. Let's continue this discussion toward the Ultimate Wiktionary! :) Peace, DanKeshet 15:48, 21 Mar 2005 (UTC), Somervillain (composed jointly)
I think you created the database in Access ? If so I can send you a database that has something in it like what I envision. It is not finished. To a large extend I would like to learn the lessons from the GEMET part before we decide what the database should look like. Polyglot has his vortero you have an ERD and I have another. What we need is something basic that is stucturaly correct with lods of room for improvement. It should not be overly complicated.. GerardM 23:57, 24 Mar 2005 (UTC)
Only User:Somervillain has regular use of Access. If you send it to him, he can take a screenshot and send it to me. --DanKeshet
We did make our example Relationship Diagram in Access. It would be great to look at your ideas for the database. You can post it, or email it to me (my email is indicated on my User page). BTW, the goal for Modular Wiktionary is to implement something very basic (the Main Entries module), then add the other modules as they are perfected and as time allows. So, again we are in total agreement about implementing something simple first, then building on it. Thanks! Somervillain 16:16, 25 Mar 2005 (UTC)
This template will point from the discussion pages of all the different proposals for a single Wiktionary DataBase to the one page where all discussion on the subject of a single Wiktionary Database is conducted, to create a discussion of that purpose, rather than of each proposal separately. User:Aliter