Back to LanguageTool Homepage - Privacy - Imprint

Token's translations

(Novram) #1


I think it would be a good thing to translate each token in every language proposed and spellchecked by LanguageTool.

I really think, that it will increase the number of users. And surely the users who would need it most…

What do you think?


(Daniel Naber) #2

By “tokens”, do you mean the messages shown to the users? That would be a lot of translations, as there are e.g. more than 2000 messages just for English.

(Novram) #3

Yes that’s exactly what I mean. Perhaps by using Transifex, and put some links and use the “developers note” to explain some particular words?

(Konstantin Ladutenko) #4

I think it is a bit strange if someone writes something in some language cannot read the message from LT in the same language… In Russian part of LT there are a lot of links to external web resources to clarify the case, they are all in Russian too…

(Novram) #5

Of course, it is a bit strange.
But what do you do for example French or Germans learning Russian and want to have a quality corrector?
It’s a shame not to offer them this service.
I think it’s cut off a part of potentials user.
By doing this you will discover some new errors, that you have never thought in, if the users decided to registered.
The “tokens” could be translated, and then having a message “for going further: external links”.

(Ruud Baars) #6

LT is not a learner’s tool exactly. But you do have a point. Nevertheless, there are far too many situations to translate from any language to any other.
An intermediate alternative might be to translate all of them to a generally known language, e.g. English.
But to be honest, I would consider translating all Dutch rule’s messages to English too much work as well.
It is not so strange to expect some knowledge of the language you are writing in.

And unfortunately, automatic translation is far from reliable still.

Don’t forget, this is not a large software company with millions to invest; it is a community, working for free (mostly).

(Lodewijk Arie van Brienen) #7

perhaps we should first modularize the messages. (would greatly reduce the amount of actual translation needed, and would greatly help with keeping the messages a coherent whole)

(Ruud Baars) #8

Maybe this is possible for part of the messages, but not for all. But it should cover java rules as well as xml rules, so there is programming involved.