Thảo luận Thành viên:Laurent Bouvier/heureux

I notice that there are cases where you can have the correct hyperlink to the vietnamese word, but other cases not.

For example, the first meaning Sung sướng is correct (not sung sướng); but hạnh phúc is not (should be hạnh phúc).

It looks to me like you have a bug: the final (compound) word in the list of the alternative (compound) words for a meaning is "broked-up-hyperlinked". (a small note on vietnamese, much of vietnamese words are actually compound words, like all the words in the bug, hạnh phúc, gặp vận, mỹ mãn, thích đáng, tốt lành).Trần Thế Trung 13:10, ngày 11 tháng 6 năm 2006 (UTC)

Second attempt

A lot of improvement have been done! Bravo! There is still a problem of wikification:

người có hạnh phúc should be người có hạnh phúc

Here hạnh phúc is a compound word (meaning heureux) while có hạnh is not a vietnamese word. Trần Thế Trung 08:31, ngày 12 tháng 6 năm 2006 (UTC)

Actually, I have có hạnh in my dictionnary, which is based on the um of the Free Vietnamese Dictionary Project and that's why he has done this weird assignment.

Check this out có hạnh > vertuous.

If you have any idea of how we can improve that ...

Laurent Bouvier 08:04, ngày 16 tháng 6 năm 2006 (UTC)

Sorry that I have not check out this item in Vietnamese->French dictionary in FVDP. Actually, this item is not in Vietnamese-Vietnamese dictionary, nor in Vietnamese->English dictionary. Anyway, this can be quite problematic, when we have 2 choices "[có hạnh] phúc" and "có [hạnh phúc]".

In this particular case, it is easy as "có hạnh" is almost inexistent. One solution that can solve for this kind of problem is that for the DB of Vietnamese compound word, you can take only the intersection of Vietnamese-Vietnamese, Vietnamese->French and Vietnamese->English. This makes sure that any word in the DB exists in all 3 dictionaries and there is high chance that it has widely recognized meaning.

Then there is still a case when both "[có hạnh]" and "[hạnh phúc]" are widely used. This reminds me of some form of AI, and as I am not convinced of the efficiency of the current AI state, I suggest leaving un wikified. So in short, the algorithm is that, whenever the bot find more than 1 way of wikification, it leaves it un-wikified. Probably, MXN or David can have better ideas.Trần Thế Trung 14:06, ngày 16 tháng 6 năm 2006 (UTC)

I have no better ideas. This particular example (có hạnh phúc) could be solved by checking whether there is a verb in the sentence (the dictionaries can help you with that), but then not all definitions have a verb and I very much doubt that all similar problems would be solved this way. So leaving unwikified seems to be the best option to me. David 14:30, ngày 16 tháng 6 năm 2006 (UTC)