Many people brought up the need for "wordlists" on localize.drupal.org in the past. These would be (maybe not so) mini dictionaries per translation team, so people coming in would be able to quickly accommodate themselves with the terminology used, without the need to infer that from other translations. The basic form of this would be a table with English and suggested translation words. While this can just be a full HTML input format node (to let you put in a table), ideally we'd reuse this data to actually help translators. Magic can happen if we can list words and suggested translations from the wordlist when you are about to translate some text. Especially if we can limit the displayed words to those which appear in the string (and maybe somehow let you look for others if you have something similar in the string which is not exactly in the dictionary yet).

The simple full HTML node input (see http://drupal.org/node/570408) and a full intelligent wordlist (see http://drupal.org/node/263794) are two ends and we need to define our needs somewhere to have a useful system for everyone. So what would be your team needs, ideas for a useable user interface for this wordlist? What would be too limiting (English/translated text input one-on-one?) or too powerful (full HTML to figure out yourself?).

I know some teams recently migrating to localize.drupal.org had wordlists before, so building on the existing experience would be useful to move forward.

Comments

The Swedish translation team has a wiki page with an HTML table on g.d.o: http://groups.drupal.org/node/6097

I has its obvious limitations. A dream solution would:

* make it easy to find/read and edit terms
* make it possible to discuss terms
* have an edit history
* have integration with other languages (Scandinavian languages could share ideas etc.)

I'm thinking some kind of node type for terms with translations, and then table views for different languages...

On the wiki page we also have some general guidelines. I might be a good idea to have a space for that kind of language specific content too.

We used to have this glossary on the Spanish site, though the module didn't survive past upgrades...

Word = Node title
Some field for translation, was custom coded, could be cck
Some listing pages to create a dicionary like view, was custom too, could be views module now

Then we can have full threaded discussions for each word. I think this could be done with almost zero coding.

It is important that 'translations' allow comments so you can have a discussion in the same place.

We (the Portuguese team) have a simple list http://localize.drupal.org/node/56 where we put the terms we have agreed upon.

Our process generally is discussing the word and me, the admin, then adds the word to our Word List.

It should be kept in mind that one English word could have many Portuguese translations depending on the context, so if a suggestion is given it should be able to list the options.

I think ideally something like this Microsoft page would be useful:

http://www.microsoft.com/language/en/us/search.mspx

As there are 2 versions of Portuguese (Brazil and Portugal) and other latin languages, I also believe it would be useful to also list the chosen translations in related languages. And like zoo33 said, a space for discussion and guidelines would also be importante (the space for guidelines is actually fundamental).

Listing both the approved translations and the used translations of the different Drupal modules.

I am one of the leads of the Greek l10n team. Before moving to l.d.o., we used BerliOS as a svn repository and a web site to store our wordlist and some basic instructions for our users. You can checkout the site there, where you can also find our wordlist. I used Docuwiki to build these simple pages and from up to now, it served us pretty well.

Regarding the available content type in l.d.o., we think it's rather limited for our needs. We don't have basic HTML tags available, plus we would need to make the wordlist (regardless of how it will be implemented) sticky. Searching through the terms would be very nice, plus a history capability, as zoo33 also said.

These are our first thoughts.

Hi all,
I'm one of the mantainer of the italian translation.
For a long time we used glossary module for a wordlist, but recently we switched to taxonomy+views to render a lightweight glossary with search functions.

you can watch it at
http://www.drupalitalia.org/glossary

If you need it i'll post the view code.

The Norwegian community uses a word list based on CCK and views. We have one node per English word, with the option of multiple fields for multiple translations of the same word (for instance access as verb and as noun).

We also have room for further description about a word. For instance, the entry for title mentions different places it's used, and has a "see also" link to subject.

We have comments, so we can have discussions per word. Some of those can go on for some time and be quite long.

We also use taxonomy to add the tag "word under discussion" when a discussion is open, and no consensus has been reached. So far that tag has been added/removed by an admin, but it would be nice if anyone commenting on a word could add the tag (like the status in the drupal.org issue queue).

Our wordlist is at http://drupalnorge.no/ordliste/engelsk-norsk , and will most likely stay there until there's sufficient functionality to move it here.

I'm pretty biased here as I was the one who came up with the implementation of the norwegian wordlist. I am not taking credit for the content though... It's of course a community effort. IMHO, common words and phrases used in Drupal will have lots of advantages if they are separate nodes. As zirvap mentioned, each word/phrase can have its own discussion (comments), taxonomy, etc. We had very long discussions on "feed" and "theme" for example. Discussions were very much alive especially the time when the wordlist is growing fast. Since I am also the maintainer of the Filipino localization project, this implementation of wordlist is also being used in the Filipino localization server, though contributions there aren't so active.

A full intelligent wordlist like where the discussion on http://drupal.org/node/263794 is going would be nice.

"There are some teams who implemented a wordlist but I'm not sure they integrated that with the editing interface."
I decided to come up with a wordlist after I decided to contribute to the community my 100% norwegian localization of Drupal 5.0. There wasn't a localization server module at that time, if I remember correctly. So the wordlist was set-up at at http://drupalnorge.no before the localization server at lokalisering.drupalnorge.no came along. So the two ended separate, unfortunately. We just had to come up with a solution so translators can easily search for standardized words. What we have is a Firefox and IE search plugin (http://drupalnorge.no/drupalnorgeordlistesok.xml) which just simply searches drupalnorge.no using the advanced search in core (/search/node/type:word+{searchTerms}). If I remember correctly, we also had a bookmarklet doing the same thing, but that bookmarklet just vanished in the air.

A full-blown integration would be nice.. Personally, this is my ideal wordlist: Each and every word/phrase is its own node so it can be discussed. Translations are custom fields where a user can also specify the part of speech (noun, verb, etc). The node body is the description/explanation of localizations - when to use the word or to which module a synonym is more appropriate to use, the particular translation doesn't fit with module A so it is advisable to use another translation so it also doesn't clash with module B, etc. Status of words e.g. Under discussion/ Approved etc can be changed by a user. On the editing pages, only the approved standardized words show up in the editing interface, may it be on a block/pane/tooltip or automagically translated in the field. Probably on the team pages, it would also be nice if the wordlist is also integrated to existing translations and there's a way of listing words that are translated differently so existing translations can be controlled for standardization.

What is important, IMHO, is to use standardized words on common Drupal words and expressions. May it be magic or old-fashioned, I think what is important is that translators have a way of easily checking the words, not solely because it will be easier to translate, but more importantly, the wordlist serves as a guide for consistency in translations and localization. Citing the Drupal handbook page Translation Guidelines:
"It is easier to understand Drupal if translators manage to be consistent with the use of words. If you translate a word with different synonyms throughout Drupal, it is difficult for people to understand that you are talking about the same functionality."

Seems like a dictionary module (demo) could be a good start for word list engine.

Our requirements
The list itself has to be in front of translator`s eyes, so block at right will be nice
jQuery based input field like already did in l10_client with auto-complete list
thats all

fields

english(original)
--module - translation variants
--module2 - translation variant
--module3 - translation variants

Our team wordlist
http://localize.drupal.org/node/186
many of them "imported" from Russian list http://drupaler.ru/forum/61

We (the Russian team) have a forum topic http://drupaler.ru/forum/61 and links for discussion of each termins

We, the catalan team, are using a wordlist by softcatalà on http://www.softcatala.org/recull.html which is also used for other catalan translation teams of many other OS projects. My experience with that is that an html table is enough. In fact, we will always use the sofcatalà resource as it is an standar already, but we might use a wordlist in l.d.o top add some drupal words. What I would find veeeery useful is having context attached to words, which I believe is something you are working on already, so I would like to have it integrated with this wordlist.

My idea was:

- to have "word" and "rule" node type at localize.drupal.org
- simple view to view the list
- since every word is a node, we can easily discuss the words

More complex set up can utilize some workflow in order to mark words as accepted/rejected

We realized we needed a strong (ideally community
discussed and approved) set of rules, that people would follow.

On the LDO pages we started drafting wikis about typical words (a sort
of a dictionary), things to do and avoid (rules). But the process and
discussion about it has been rather chaotic.

So we started thinking that it'd help to have a list of words, rules
and so on, where we could have discussion about each of them
and possibly some process to have them confirmed as valid.

That's why we started making the web - http://preklady.drupal.cz, where
these node types appear (or will) :

Word:
- english version, czech proposals, czech confirmed version
- status (proposal, selected, confirmed), note

Rule:
- name, description, DOs, DONTs
- status (proposal, selected, confirmed), note

Module:
- owner, translators, subscriptions

Each of these has coments enabled and the idea is they just be pushed
through the process basicaly by the dicsussion. Probably we could even
mimic the d.o. project issues functionality.

So, what is the best practice to maintain a translation wordlist? Is it some local website with custom content type + views or is there any hope to host wordlists at L.D.O.?

Turkish glossary is a simple wiki page.
And Gábor , is this topic still on your agenda?

French team put a full glossary site (also used to provide guideline for contributors) that permits to vote on term translation suggestions and discuss via comments.

Then the glossary is exported in JSON and can be viewed while translating on l10n_client and localize.drupal.org via browser extensions, giving this:

Having this directly in l.d.o would be nice, unfortunately we have to wait (or better: contribute!) for D7 version.

Reviews and comments appreciated!