Locale updates and downloads

German Locale in stock :G :D

Thanks for all the hardwork Domi.

Looks good Domi but what's with the .rar files? Maybe we need the repo better organized, I'll come with an example soon.

Edit:
Right now in the original repo to download, update/change something you will be downloading ALL locales which is absolutely unnecessary and since the master repo contains the latest code committed seems you had to add .rar for the stable releases which is also unnecessary.

I've setup a repo at https://github.com/JoiNNN/PHPFusion-Locales as an example, here we have:
- a branch for each locale, so we won't download all locales when trying to change something. The users who update the locale will work on their own locale branch.
- releases for each stable version, after a release is made code can be committed to the branches, so the branches will always have the latest locale updates while using the releases for stable locale thus no need for a dev branch.

It looks really great!
But there are no counter on them archives.
That's why I added the files and linked to a raw download from our downloads to the stable Archives.
We should probably put em in a new branch called stable, but we will need to look at the structure for Github again when 8 will take it´s place there.
Perhaps we all can agree on a standard at that time.

We'll link from this site directly to release links such as https://github.com/JoiNNN/PHPFusion-L....02.07.zip keeping counter stats and users who want to download form Github will find easier stable releases.

So basically your saying. There are no designated translators anymore. Any one can mess up the translations.

As (seems to be) former translator who spend a lot of time on translations this is an unpleasant surprise.

@DOMI most of the locales marked as 7.02.07 are actually missing the locale lines added in this version. They need fixing or not linked as .07 ready.
All NSS should be required to publish a valid .07 ready locale by now, so at least we can update the Github repo if they are not going to.

I created a spreadsheet of languages in relation to our 34. It contains the top 34 by population of native speaking people, as well as some other useful data.

This is a summary using 2007 statistics:

[ulist=disc]of 34 PHPFusion languages, 13 are in the TOP 34 by Native Speakers.[/ulist][ulist=disc]of 34 PHPFusion languages, 31 are in the Google Translate Database.[/ulist]
[ulist=disc]PHPFusion has 7 in the TOP 13 by population of Native Speakers
These 7 languages count for 27.70% of the world population.
The 6 that we do not have count for 23.30% of the world population.
These languages include Chinese, Bengali, Japanese, Punjabi, Javanese, and Wu.[/ulist]
(I use top 13 because 50% of the world speak these 13 languages, and the remaining 50% of the world speak the remaining 6000+ languages, yes I said 6000)

[ulist=disc]Total PHPFusion language coverage by population equals 35.99%
If we had Just the top 13, plus the others we currently have, we would be at 59.29% world population coverage.[/ulist]
[ulist=disc]If we take the TOP 34 by population of native speaking, cross that with google translate, and figure out which we do not have PHPFusion language files for, we get a number equal to 30.85%. (current plus top 34 in google would give us almost 67%)[/ulist]
It may be possible to bolster our numbers by using google translate. This may attract people who speak the language fluently. Enough so that they want to help improve where translate did not work correctly.

I have attached the spreadsheet for your viewing pleasure.

Enjoy, B

I do not see how that is much different from the way it is done now, except the path.
As I mentioned above, we can look on the structure more close when it is time to add v8 structure. But this is how it is going to be done.
A open archive and a file for download that are linked to from here.



From the download here, most of em are also listed 7.02.xx
But yeah it is a shame that so many ignores the update, alto it have little impact on the everyday use of most sites.



We are in this as open source, we need to be open even with these files.
Anyone can request an update of these, yes.
We do not need to accept all of these requests and updates.
Locales is one of these things we need to be careful on and where we would need to double check. esp if it is a language that have a dedicated translator here.
Nothing prevents designated translators to remain designated.
We have the stable archived file as the download just for these reasons among others as mentioned above.
If a trusted translator comes with a new file we will just update the stable with that and the archive with his work.
Let´s not exaggerate or even imply that there need to be a negative impact out of this. I only see possibilities here.

@BrandonBlack
Very interesting read, thank you !
I hope we get a Chinese translation soon.
I am reluctant to Google translate at this stage.

I would like to remind everyone that our former Locale archive contained 15 translations. Most of the updates where done in 2011 to these.
Our new archive have a total 34 translations where many have been updated recently.

I concur with Domi about the Google Translate approach. I speak and write three languages. If the input is flawed, the output will be garbage. Many of us may have access to the Microsoft Translator Engine in Office 2010 and later (which is reputed to be better than Google), same goes there. If you have a badly written sentence in English, the result in Swedish, or any other language I would presume, will require a total rewrite and you might have done that from start.

As for locales, well, that is of course shorter sentences and even single words so it may not be that difficult, but you always have to keep context in mind. That means translation and corrections on the fly and corroboration and verification of a correct understanding and placement in the interface.

That is why I have direct file access implemented on my NSS, to try to catch the errors immediately as I see them by editing the locale on the fly and review. But I can not catch all, even if a translation is in place it may be an incorrect wording for that exact context. So we also rely heavily on reporting.

This artificially intelligent translation of languages issue should be approached no differently than the way we approach the different levels of skill pertaining to programming on this site.

Beginning programmers are allowed to come here, as I have done, and post their code to your forum, no matter how rough, out of wack, incorrectly formatted, etc, etc. Same for intermediate, advanced, and professional. So too should this be the "standard operating procedure" of people translating language files.

An artificially intelligent translation, in my eyes, is no different than a person beginning on the road of programming. Context is usually horrible, placement of language specific elements are wrong, however legible, among other things that differentiate a beginner from an advance user or professional.

The point here is to "spark" the motivation of someone simply by starting. We should be after "attraction" over "accuracy" in the beginning. To circumvent any confusion as to translation legitimacy we could create a branch of translations that include versions done through artificial intelligence and thoroughly mark them as so. Call them, officially unofficial if you want.

(A little off the subject, I lived in Okinawa Japan for over 4 years and always came across their Japanese to English translations, and they were so wrong, but in a cute way that it always made us laugh. Many times I would strike up a conversation with them to help explain the context or format of their translation better, so they could better promote their product. Similarly, and on the note of "attraction", we want to attract people who are helpful to the community. I am a strong believer in "If you build it, they will come". So if you build an incorrect translation, you will attract people of that language who are helpful. It is a simple formula actually.)

Another point I would like to make on this subject is the fact that, as it stands now, in order to get what you consider a legitimate translation, there is a requirement for someone to be bilingual. However, if we do a rough translation, mark them as so and package them for download, we have essentially removed the bilingual necessity. Now, someone who only knows the language can come in, make corrections, and with a few intelligently placed lines directing them toward a submission area or forum thread, they can post what they have created for the fusion world to use.

PS. I have dibs on Latin.

Yes, it is a possible path.

I am also considering the possibility of short sentences and single words actually might not be that bad as Richard mentioned.


It is usually true and one of my absolute favorite quotes.
However some constructions can be so bad that "they" prefer to run.

Some AI Translations might be just that bad.

Have you seen these mass translated production pages about foods or remedy's etc etc?
If you have , you will know what I talk about.

The translations on these sites just do not do em justice and have more of a negative effect.

If we do not have a decent reliable control of the translation after it is done, it is just hard to know how people will react when using it.
Mind you it can and will probably also differ from Lang to Lang.

Let´s give it a while and we can take up the subject after v8 is out if we seem to still be lacking some key translations.

Sounds good Domi, to be readdressed after v8. B