http://tatoeba.org/fre/sentences/show/1546330


# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

Ah, on dirait que le moteur de recherche prend en compte les majuscules. J’avais recherché « ça le fait » avant de créer cette phrase sans trouver la tienne…


# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

L'explication est toujours la même : Tatoeba, c'est fait par des anglophiles pour des anglophones. Donc avec les lettres anglaises, ça fonctionne, mais avec toutes les autres lettres, ça ne marche pas...Tatoeba est incapable de reconnaître que Ç est la majuscule de ç.
Merci le monde borgne unipolaire !


# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

Franchement, abstiens-toi de parler de ce que tu ne connais pas et de tirer des conclusions hâtives. Le problème ne vient probablement pas de Tatoeba mais plutôt du moteur de recherche Sphinx qui est un logiciel tiers sur lequel Tatoeba s’appuie.

Et puis le problème est beaucoup plus compliqué que ça. Ce n’est pas Tatoeba ni Sphinx qui sont anglophiles, mais tout l’univers informatique en général (le logiciel, le matériel, les outils…) qui prend ses racines directement aux États-Unis et qui est anglais dans sa conception même. Si tu veux une explication, il te faut plutôt remonter à l’invention de l’ASCII en 1963. Et je dois dire que si un logiciel tente de contrebalancer cette tendance, sans pour autant y arriver encore à cent pour cent, c’est bien Tatoeba.

Étant moi-même développeur, et un peu impliqué dans le développement de Tatoeba, je prends ton commentaire de manière assez personnelle, d’où cette réponse un peu salée. Je suis tout sauf l’anglophile que tu décris. Et franchement, plutôt que de maugréer à tout va contre ceux qui font fonctionner Tatoeba, tu pourrais peut-être adopter une approche un peu plus constructive en rapportant les problèmes afin qu’ils puissent être corrigés. Et sans juger les gens. (Pour ce problème précis des majuscules, j’ai déjà rapporté l’information donc c’est bon.)


# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

>Le problème ne vient probablement pas de Tatoeba mais plutôt du moteur de recherche Sphinx qui est un logiciel tiers sur lequel Tatoeba s’appuie.

ben c'est un choix de Tatoeba d'employer un moteur de recherche qui n'est pas internationalisé !!

>Ce n’est pas Tatoeba ni Sphinx qui sont anglophiles, mais tout l’univers informatique en général (le logiciel, le matériel, les outils…) qui prend ses racines directement aux États-Unis et qui est anglais dans sa conception même. Si tu veux une explication, il te faut plutôt remonter à l’invention de l’ASCII en 1963.

Oui eh bien je ne le sais que trop, parce que je suis né avant toi, et ce mépris des autres cultures n'a que trop duré !
Moi aussi je développe depuis des décennies et j'utilise des briques qui respectent le multi-culturalisme. Elles existent, il suffit de choisir les bonnes plateformes, et pas de laisser ce choix systématiquement à des Étasuniens qui nous dictent constamment tout ce qu'on doit faire, à leur manière bornée.
Les anglophones devraient être interdits de développement de services internationaux, parce qu'ils sont incompétents en matière d'internationalisation et que c'est le cadet de leurs soucis !

>Et je dois dire que si un logiciel tente de contrebalancer cette tendance, sans pour autant y arriver encore à cent pour cent, c’est bien Tatoeba.

La preuve ? La voici : cherche "Tatoeba" sur Google.fr (j'ai bien dit ".fr" !!!)
=> TOUTES les pages indexées en première page sont en anglais !!!
Tu peux aussi faire le test sur le Google japonais...
C'est un vrai scandale que je n'ai cessé de dénoncer, parce que vous ne prenez même pas la peine de faire indexer correctement les pages dans les autres langues. Le SEO, c'est un métier !
Et le résultat, on le connaît ! Tatoeba n'attire que les anglophones ou les singes anglophiles et dissuade tous les autres !

>tu pourrais peut-être adopter une approche un peu plus constructive en rapportant les problèmes afin qu’ils puissent être corrigés

Je n'ai cessé de le faire depuis 4 ans, et j'ai l'impression, la plupart du temps, de pisser dans un violon.
Tu n'as qu'à consulter l'historique du mur et tu verras que j'ai fait des tonnes de suggestions.
Seules quelques unes ont été prises en compte, après que j'ai longuement pleuré, dont, merci, dernièrement, l'ordre de tri des noms de langues. Enfin !


# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

I'm sorry that I can't use french well enough to write this comment in french, but anyway.... I'd just like to say that no one contributing to Tatoeba is paid to do it, they volunteer their time and effort for no return. Calling them anglophile monkeys and making it seem like there's a grand imperialist anglophone conspiracy going on gets you no respect and gets nothing done. What gets you respect and gets things done is contributing patches and tackling problems, the last person who had the time and knowledge to get this sort of thing done was sysko and he's too busy now. Sphinx was and still is one of the best search indexing software out there and handles multiple languages just fine. It's free and open source software. You're not entitled to anything. If you find a problem that bothers you so much, you don't get to yell at it. You tackle the problem and send a patch. If you think sphinx isn't good enough, help us move to something better, most of us just don't have time for this. If you think SEO is terrible, contribute a few patches to the html markup that tatoeba generates, I promise that you'll find people to help you with this and your work will show up on the site in the next update. You can yell at this for another 10 yrs and nothing will get done and you just end up alienating the few people that are still sending us patches.


# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

Let me first tell you you shouldn't mingle in other people's conversations when you don't master the language they use, because you misinterpret everything.

So I'll reply to every single point you mention so agressively (but I'm used to it, from someone who went crawling into my personal blog to extract irrelevant out of context quotes...)

>no one contributing to Tatoeba is paid to do it, they volunteer their time and effort for no return

I perfectly know that. I initially proposed to help, including financially, and was snubbed by Trang at the time.

>Calling them anglophile monkeys

That's a distortion of what I wrote (your specialty). I said Tatoeba attracts anglophile monkeys, not that the people developing Tatoeba are anglophile monkeys, although this doesn't exclude that...
I already explained, long ago (but no one took heed, as usual, or even took the pain to just try understanding the problem) the vicious circle Tatoeba is in, like so many non-correctly-internationalised websites, leaving the pervading illusion that the world entirely speaks English.

So here I go again :

When you search for Tatoeba on any search engine (try Google) and you're from any non-english speaking country, what you get is a result page that is almost only in English. Try it on google.de / .fr / ...
So if you're a person interested in languages, but you don't speak English, you don't even understand that Tatoeba is about languages and translation and you move to something else.
The result is that most people who subscribe to Tatoeba are English speakers, getting the illusion that most people interested in languages speak English (that's the victims of that illusion that I call the anglophile monkeys). Which not only is a big stupidity, but is an open sore for Tatoeba because it misses so many talents of people who speak so many languages but...no English ! Yes, they do exist ! The vast majority of bilingual or multi-lingual people on Earth actually don't speak English. They speak Arabic and Wolof, Nahuatl and Spanish, Hindi and Malayalam, Chinese and Tibetan, Kazakh and Russian, Berber and French, Hungarian and Slovak, Ukrainian and Romanian, Swedish and Finnish, Portugese and Tagalog...because that's how the world ACTUALLY IS, outside of Hollywood and California, I mean...

Of course, anglophile monkeys are unaware of the problem, or they don't care, so they never solve it, although search engines have taken the pain to provide tools to address the problem :
https://support.google.com/webm.../answer/189077

In order for Google to know which page relates to which languages, and index them accordingly for each national domain, Tatoeba pages should contain these :

<link rel="alternate" hreflang="es-ES" href="http://tatoeba.org/spa/" />
<link rel="alternate" hreflang="de-DE" href="http://tatoeba.org/deu/" />
<link rel="alternate" hreflang="fr-FR" href="http://tatoeba.org/fre/" />
<link rel="alternate" hreflang="pt-BR" href="http://tatoeba.org/pt_BR/" />
<link rel="alternate" hreflang="ja" href="http://tatoeba.org/jpn/" />
...

>the last person who had the time and knowledge to get this sort of thing done was sysko and he's too busy now.

I know, and sysko was a very dedicated and open-minded person and did wonderfully and I thanked him for this many times.


>making it seem like there's a grand imperialist anglophone conspiracy going on gets you no respect

It's not only a political conspiracy, which is well enough documented, it is also the result of complete neglect and disrespect for other cultures on the part of English native programmers in general, and a self-immolation on the part of non-English native programmers who are always so eager to wallow in that self-disrespect for the greatest worship of their masters. When the master craps, the slave always licks it with delight.


# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

as for the search engine, has anybody ever considered http://lucene.apache.org/solr/ ?
It's very powerful and features a full UTF-8 support (which apparently works only for certain languages and limitations in Sphinx)


# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

So I'm out there to distort your every word, tatoeba has anglophile monkeys afterall, and is stuck in a "viscious circle" of "big stupidity" for a long time because "no one took heed" of your heroic defense of linguistic diversity and your proposals for support were "snubbed". And of course there's an anglophone conspiracy that we're allowing to extend to tatoeba as we "wallow in self-disrespect" and lick the master's crap with delight.

I don't doubt that you're a nice and great person and we'd have a great time if I ever met you and we had some coffee together one day, but this isn't the kind of language you want to be using in an international project and especially while defending ideals like linguistic diversity that I also happen to support and I'm pretty sure everyone else on tatoeba does. You hurt your own cause and rub everyone else the wrong way, you probably ruined someone's day right now, be it gillux, or someone else who's reading this.

I'm not very sure why such a simple SEO hasn't been implemented yet, but maybe, just maybe, it has to do with the website not having much active development for a while and there not being any ticket describing in detail what needs to be done and very little to do with anglophile monkeys and their conspiracies. I'm not familiar with php or cakephp's templates or the base template tatoeba uses to generate the head section, and don't really have the time at the moment or patience to sink some time into learning this, but I imagine you need to query the database for our supported languages (or store them in a config file somewhere), convert the iso codes from the one that tatoeba uses to the one that google is expecting, and loop in the template and generate the html needed. But again, this isn't gonna fix itself and if you have the time and know how to get this done, send us a patch. I'm more aware of this issue now that you mentioned it (thank you) and I'll try to take it into consideration for pytoeba since I'm in the middle of writing the interface at the moment (does that make me an anglophile monkey, I wonder?).

As for solr, I'm not sure either why it hasn't been considered, but I'd guess it's a RAM hog and we didn't have the resources to run it, or sysko just likes c++ projects and thought sphinx was fine for our use-case. Anyway, if someone goes through the trouble of writing a configuration file (autogenerating it with a script) for tatoeba's use, integrating it into the current codebase, and testing it on a dev machine. I'll be happy to take the time to set it up on the live server myself.


# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

>I imagine you need to query the database for our supported languages (or store them in a config file somewhere), convert the iso codes from the one that tatoeba uses to the one that google is expecting, and loop in the template and generate the html needed

That's in effect what should be done in an ideal world.

But you have no ressources to program that little loop and I don't know the architecture of your application and I'm not familiar with php, which is antechrist to me.

But here is my suggestion for a quick-(and not so dirty) solution :
You surely have a layout HTML page. So you could just insert the hreflang HTML lines I suggested above in its <head> section.
At the moment, Tatoeba features 20 locales. That's just 20 lines to insert, taking care for each to match the correct Google language code to the correct Tatoeba choice of url extension (example : /jpn for "ja")
In some cases, the locale corresponding to a language variant doesn't exist yet on Tatoeba. For instance, there is no Canadian French locale.
In this case, the country extension of the language code should be omitted so for French, it should be :
<link rel="alternate" hreflang="fr" href="http://tatoeba.org/fre/"; />
and not
<link rel="alternate" hreflang="fr-FR" href="http://tatoeba.org/fre/"; />
otherwise, francophones outside France would not retrieve results in French on their national version of the search engine.

I happen to have a strong experience in multilingual user interfaces and here is the best advice I can give to someone who works on a multilingual/multicultural user interface :

NEVER test on the English UI if it is your code base (or Russian, or French...if that is your code base), ALWAYS test on another locale ! You have no idea how many faults you find this way !
Consider English as a locale(s) as well, because it is ! Nobody so far has asked for a different UK-English locale and US-English locale (or even Indian-English locale) on Tatoeba, but that might be necessary someday... It happened to me that field names (in business applications) defined in the UK where not understood - or even worse, misinterpreted - by US users...



# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

@gillux

en testant, tu peux constater que Ç/ç ne sont pas les seuls caractères qui ne sont pas correctement associés par Sphinx.
C'est aussi le cas de À/à, É/é, È/è
Par exemple, si tu cherches « À la tienne Etienne », tu trouves quelque chose, mais si tu cherches « à la tienne Etienne », tu ne trouves rien...

J'ai fait un peu de recherche concernant ce problème Sphinx qui n'associe pas certaines lettres à leurs majuscules et j'ai trouvé ce lien : http://sphinxsearch.com/forum/view.html?id=11073

Je ne sais pas si c'est toujours d'actualité, mais si je comprends bien, il y a moyen de contourner le problème en tripatouillant la charset_table du fichier de configuration de Sphinx et d'ajouter:

À->à,Ç->ç,É->é, È->è ...

Évidemment, idéalement, il faudrait le faire pour toutes les langues.
Bizarrement, des tables existent pour certaines langues : http://sphinxsearch.com/wiki/do...charset_tables

Mais pas pour le français...
Avez-vous connaissance que des utilisateurs d'autres langues sur Tatoeba se soient plaints de ce problème ? Dans ce cas, la solution est là, à portée...


# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

In case anyone wants to continue the discussion, please go here: http://tatoeba.org/eng/wall/sho...#message_20121
So that we keep the comments for things related to the sentence.
Thanks.


# --------------------------------------------------------------------------------
# This comment was copied from #3410618 when duplicate sentences were merged.
# --------------------------------------------------------------------------------

Duplicates of this sentence have been deleted:
x #3410618