On Thursday 27 September 2012 15:32:45 Tony Su wrote:
Everything you suggest is possible and I'm of course willing to setup any demo people may be interested in.
The main thing is I would want some kind of concensus from "The Powers that Be" there would be some kind of future if it really solves the problem satisfactorly, no one wants to waste time on something that can't happen.
Dunno if anyone else here gets skull crushing visions from 'The Powers that Be', I sure don't - I don't think there is anyone who can answer that question I am afraid. However if the demo works it will be up to the website owners to accept merge requests to enable this functionality. They will only do that if they feel it's worth it - as they will have to maintain it in the future.
As they don't seem like feeling responding here, I have no idea if they will. That is in part why I suggested to create a demo site of say open- build-service.org - if that works satisfactory, the owners of that site (that'd be Adrian & the rest of the OBS team, I suppose) might like it and be willing to adopt it. If they do, well, in dutch we say "once one sheep has crossed the dam, others will follow".
The reason why I used the term "shepherding" is because I am always interested in building something that can live independently of myself-- I'm not looking to make myself indispensible, from the first day of anything I do I'm looking to bring on others who want to build the same thing.
Unless someone does it, it isn't gonna happen. We've got shepherders plenty, but no sheep interested in doing it until someone has shown it's worth doing. That's what I suggested doing... If you've implemented this and it is accepted by the website admins, that means they are convinced of the value. At that point, it WILL live on independent of you because they have to maintain it and websites which don't implement it might get it just to fall in line.
Also, awhile back I started penciling out what mass Translation for openSUSE might become. Yes, all things start small. But, if this becomes important to more people in openSUSE, I'd like to involve anyone who wants this to enhance what they do and avoid being unable to deliver.
It's about the first step - unless it's shown to work on a site and provides some benefit, it won't go anywhere, I'm afraid. That's why I suggested to Just Do It.
On Sep 27, 2012 2:38 AM, "Jos Poortvliet" email@example.com wrote:
On Wednesday 26 September 2012 15:18:09 Tony Su wrote:
Hello Jan (and whoever else receives this, I'm not subscribed to all the mail-lists on CC)
Yes, it's quite possible that the initial translations might not be "good enough" -- and ultimately since machine translations today still cannot usually provide better than word for word literal word substitution, "good enough" is probably best defined as understandable although not with the smooth idiomatic linguistic structures that can best be provided by a human being.
What machine translation can provide is the ability to get the proper meaning across, to communicate an idea properly. And, if human resources aren't available, this is better than no communication at all.
As for accuracy... Particularly for short, "standard expressions" that crop up again and again in the types of documents we produce, Web-based translations provide a means for anyone to submit an improvement or correction. Assuming that Google or Microsoft or whoever is used as the Translation Partner properly evaluates, accepts and implements suggestions for future use of the same expression, we should expect that within rather short order future documents should be translated extremely well.
If there is any interest in openSUSE/SUSE to investiggate the capabilities of this technology, a project should be designated that can properly evaluate whether machine translation is worthless or promising and if desired I am willing to shepherd it.
'shepherd' or 'do' ;-)
I wouldn't know what would be needed to actually TEST this out - but you're right that there are plenty of pages not translated in plenty of languages.
Quite a few of our sites are in github, maybe you can set up a test version with a translation system of, say, openbuildservice.org: fork https://github.com/openSUSE/o-b-s.org and add the translation system, then run it somewhere so ppl can check it out. If it's better than what we have (and from your comments I take it it will be) you can just make a merge request to the github repo and the maintainers get it up. And done, one down, a dozen to go :D
Then there is the wiki. How do we support our translators with this, can google translate be helpful for that? For example, maybe it is possible to have an auto-translate run over our wiki pages so all pages get translations in say the basic 25 languages or so. Then people can edit as things used to are...
Is that possible? Is there a mediawiki tool which can crawl our en.opensuse.org wiki and, for pages that have no de.opensuse.org, fr.opensuse.org etc etc equivalents, create and fill them? If you 'just' manage to do that, our wiki has become far more accessible to non-native speakers...
On Sun, Sep 23, 2012 at 1:46 AM, Jan Engelhardt firstname.lastname@example.org
On Tuesday 2012-09-11 23:20, Tony Su wrote:
Both are free and only take seconds then ask for a native speaker review to clean up any idioms and colloquials.
The time to weed out the bugs of automatic translation is close to doing a non-automated, more targeted translation. Especially the farther east you go on the globe (Japanese TL with Google is pretty much unusable in either direction) and/or dealing with highly-technical words (and fillers) - which the announcement is in no way short of, like "Call For Papers", "to keynote", "to kick off", "workshop", "track", "session", "usability expert", and (obviousisms like) "speakers talking".