For one of the websites that I am building (corpqna.com) , it turns out
that I have a very high rate of international users (>50%). I am looking
to do some inexpensive translation services (maybe even leveraging my users
to do this if possible)
markup some html element with a "translate" attribute. Then, there would
be some sort of widget that lets users offer translations. I would have a
console that can accept, reject or modify those translations. It would
automatically translate elements if available.
I\'d prefer to do it this way, vs a "back-end" approach which would be more
expensive and time consuming.
Does such a thing exist?
I almost chimed in earlier but didn\'t. To translate text resources on a CMS
with user generated content (as I believe we\'re talking about), is really
about localization not internationalization. In addition, I think that when
most developers see i18n they think of internationalizing static assets,
not user generated content. (I.e. the text on buttons, or words on a menu,
or your terms of service vs content like tweets.)
Web frameworks typically handle the overhead of i18n/l10n, so as long as
the developer builds a dictionary for each language, there is no hassle. I
believe you\'re still at square one for translation services, but just
wanted to provide some notes on your use of jargon.
For a solution, you\'ll need to figure out where your users are coming from,
assume language localizations (India might be tough due to dialects), or
just let the user select the language.
For user based content translation, simply 2x text fields on a form (1
mutable-the translation box, 1 not-the original answer) and 2 selects
stating the from and to-languages. Whenever a change is made to the
original, all the translations need to be marked as dirty. It\'s up to you
whether to display old translations.
Depending on the volume of translations/importance of localization, you
have a great opportunity to make a play for software as a service. This
might be an interesting revenue source to consider.
If you decide to do it the old fashioned way and need translators, I have some fiends who do it professionally for major organizations and the CIA which is what I\'d actually recommend (I also used to run a chinese /english translation co in china). Ever read a Chinese website that was translated to English automatically? Pretty bad.
But yes, if you could figure out a way to do it dynamically like you\'re looking for, I imagine its worth a lot to a lot of people.
Sent from my iPhone
On Jan 7, 2013, at 12:33 PM, Devin Fee <devin....@gmail.com> wrote:
I was hoping for something a bit more magic than that. I have to believe
that someone has built this service. If not, someone should. Because I
would pay for it! :)
ie, I don\'t want to write lots of i18n code that has nothing to do with my
Well Google translate widget is magical alright:
https://translate.google.com/manager/website/. Too magical perhaps. It\'ll
translate the whole page and the resulting translation will be quite bad.
(Robots are not yet translating as well as humans).
Translating web specially tagged html elements on the front end sounds like
an interesting idea but in practice I can see some of caveats with it:
1) Server generated messages may be visible only under specific conditions
and may be hard to catch.
2) Because translation is decoupled from your server side code, keeping
your translations up to date becomes even trickier (unless the text is
translated on automatically by robots).
3) You still need to tag your messages for translation on the server side
(in your code or templates) which is not that much different from
triggering the whole i18n machinery on the server side.
So my take on this is, roll up your sleeves and do it properly, i.e., on
the server side.
On Mon, Jan 7, 2013 at 12:02 PM, Douglas Tarr <douglas.t...@gmail.com>wrote: