Closed Bug 1225068 Opened 10 years ago Closed 10 years ago

Move all Perisan l10n projects to pontoon

Categories

(Mozilla Localizations :: fa / Persian, defect)

Product:
Mozilla Localizations
Component:
fa / Persian
Type:
defect
Priority:
Not set
Severity:
normal

Tracking

(Not tracked)

Status:
RESOLVED WORKSFORME

People

(Reporter: amir_farsi, Unassigned)

Details

User Agent: Mozilla/5.0 (Windows NT 10.0; Win64; x64; rv:44.0) Gecko/20100101 Firefox/44.0 Build ID: 20151113004115 Steps to reproduce: Hi there. Now Mozilla have multiple systems for localization management and maintaining. One of them is on Mozilla.locomotion.org and another one is on verbatim.mozilla.org Actual results: Having multiple systems makes separation and wants for register and work with multiple systems for localization teams. Also pontoon is better system for localization and it's visual. Visualization in pontoon is new idea around localization systems and made it more user friendly. Expected results: Now, Mozilla should merge all of projects in locamotion and verbatim to Pontoon. Then close locamotion and verbatim permanently.
Hi Amir, Thank you for being a big fan of Pontoon. We agree that it's a great tool for translating and localizing Mozilla projects. Our approach to tools is one that allows each l10n team to decide which tool they prefer to use to localize Mozilla projects. Many love Pootle (mozilla.locamotion.org) and some prefer to only use a text editor. For us, it's important to allow each to work how they are most comfortable and in they way they feel is most efficient. That being said, we're going to continue to keep Pootle available. If this bug is more specifically for the Persian team's l10n projects, then I suggest CC'ing Sara, Reza, and Arash in this bug and having a discussion about whether this move is what the whole Persian l10n team wants before we proceed. Thanks, Jeff
Adding the rest of the Persian team here, should Amir wish to have a tool migration discussion.
Summary: Move all localization projects to pontoon → Move all Perisan l10n projects to pontoon
Hi Amir, Yes Pontoon is a great tool but as we had some discussions earlier with Arash and Reza about keeping Firefox, Firefox OS, Firefox for android, and Web parts in Pootle for now, till we make sure that Pontoon can provide everything that we currently have in Pootle (mozilla.locamotion.org), specially the terminology of the projects and suggestions based on previous work. We can have more discussion about it for sure!
Another thing to note is the quality checks that Pootle offers, both the direct quality flags and the user rights definition. I know Pontoon is catching up there, but hasn't arrived that same level yet.
Adding to Sara's comment, Pootle has a nice feature which categorizes strings based on priority. Pontoon doesn't provide this right now, and we rely on it for projects. So I can't actually see any advantages here for Pontoon right now, except that it has a more up-to-date interface.
When you all feel you've come to a conclusion, feel free to close this bug.
Hi there. I think because pontoon is owned by Mozilla, it's in hands of Mozilla. If we reported any problem with pontoon it not requires Mozilla ask from other groups for Fix this problems. Pootle is good, but have some of problems with languages like Persian. I reported them in BugZilla, but non of them fixed! List of them is here: https://bugzilla.mozilla.org/show_bug.cgi?id=1093496 https://bugzilla.mozilla.org/show_bug.cgi?id=944858 https://bugzilla.mozilla.org/show_bug.cgi?id=944780 https://bugzilla.mozilla.org/show_bug.cgi?id=944773 https://bugzilla.mozilla.org/show_bug.cgi?id=1220390 Also in response of you that said pootle can use previous translations, i want say pontoon also have this feature. In addition, pontoon's machine translation is very good. When i want translate a string pontoon make multiple suggestions from machine translation from multiple sources. I used pontoon for many times. I used pontoon's machine translation, it was very good. Also one of pootle problems is pootle showing the phrases that should be translate, and if you don't want translate that string you should click on next and next again to see that phrase which you want to translate; but in Pontoon you can see all of phrases which you should translate in a scroll. It's very important feature for Agile translation. I translated pontoon's UI very faster than pootle by this feature. Because first i translated those phrases which i can translate. Pootle is like a computer based test system which you should go next if you can answer the current question, but pontoon is like paper based test system, which you can first answer those questions which are easier, then go to other questions which you should try about them. Second is better. Also those features that pontoon haven't yet and Arash and Sara said them, can be included in it by Mozilla. But I couldn't request those features which i want from pootle because it wasn't Mozilla's product. Thanks.
Hi Amir, I've gone through the bugs that you list above and have either poked Dwayne for a status update or resolved them depending on the status. For Pootle issues, the best place to file issues is https://github.com/translate/ or make sure to assign a bugzilla bug to Dwayne Bailey. You're always welcome to request features in Pootle. Consideration of your feature request isn't determined by who owns the software, but by the technical and time cost to adding it to the tool. With both bugs and features, Pootle and Pontoon are both open projects that you can contribute fixes and developments too without waiting on someone to provide the fix for you. Dwayne & the Pootle team have standing contracts from Mozilla to develop the locamotion software according to Mozilla's needs (Verbatim is being replaced soon). The Pontoon team consists of one person, Matjaz. Either way, any feature request would need to be balanced according to the cost and goals of a particular release. Now what both projects can do better at is being more transparent about their feature roadmaps. This would indicate whether your feature will be taken up during a particular release. This is something that I can bring up with both Dwayne and Matjaz at All Hands.
I think best way for solving this problem is that Mozilla make all localization projects accessible from Pontoon and Locamotion at same time. Really, pontoon or locamotion(pootle) will send translated strings in a file to product source(Like github), also pontoon and pootle giving original strings of a project from a source. We can connect pontoon and pootle to that same source. You can think about a database which a Desktop app and a Web App are connected to it. Or you can think about a database source on a server which many users with different softwares can connect to it. Also it's like an email server which Outlook, Thunderbird Opera mail can connect to it. Every user can use one of them for connect to that mail server and use email service. If this happened, Everyone in a localization team can use Pootle or Pontoon. Yes, it think this solution can make end all of discussions about tools.
I thnik matjaz can do it
Flags: needinfo?(m)
(In reply to Amir Farsi from comment #9) > We can connect pontoon and pootle to that same source. This is totally not practical (both in terms of programming and team work) and eventually makes more issues. There should be only one tool. It's right that both tools are using one source, but imagine that I make a change in Pootle and you make a change on the same string on Pontoon. Now we have conflict here, which translation should be the one to be submitted on repo? And this is the simplest case, knowing how git, Pootle and Pontoone work, this is not doable. Also we don't want to change the whole system of Mozilla translation process here on this Bug. We just wanted to discuss if the Persian projects should be migrated to Pontoon or not. Right now majority of community prefer the current projects stays on Pootle and we keep other parts (Marketplace, Firefox accounts, firefox hello) on Pontoon, but lets keep an eye on Pontoon.
As Arash mentioned Pontoon and Pootle both update the same repository. But as far as I know currently Pootle do not get updated automatically(real time) if the repository changes through file push or other tools. I think Ponton is also similar. The untranslated strings get updated only if the changes occur in themselves or manually. So if you translate something in Pontoon I may translate it again in Pootle and it would be a problem.
Hi again. I think it's a very good technical discussion that having one translation server with multiple clients/tools(Maybe web based or Desktop or Mobile) which can connect to it. As your response, i want say if we have online translation/localization server which contains the original and translated strings, we can develop a protocol for communicate with that translation server. Also this server and clients should be have a sync system. For example, now when one of team members changed a translation string on pootle, all of other team members will see it, and they won't send any translation again. This can be develop by a sync system. As result, when you send a translated string to translation sever via pootle, if anyone connected to translation server with pontoon this string will show to it. Now an Airplane reservation system also working like it. If a person reserved a place in airplane, another user will see it as reserved. If a person is reserving a place, that place will lock for other users to end of his/her reservation, if ha/she cancelled reservation, that place will unlock again. This mechanism is also on Database servers like MySQL for records which are editing. Also it's in email. Now if you connect to your email via Thunderbird and remove an email, Thunderbird will sync with mail server. Then if you connect to mail server via outlook/web clieny(Like native Gmail client), you will see that email is moved to trash. The idea of a translation server that is separated from translation tool and these things can be sync via a unified communication protocol can be as new product of Mozilla. Yes, Mozilla translation/localization server!
Also i want say we can develop a three layer system. Layer 1: Github/Programming repository Layer 2: Translation Server Layer 3: Translation Tool/Client We will not require to communicate from Translation tool to Programming repository. Translation Client(Layer 3 can be vary from Pootle, Pontoon or Desktop Client or Mobile App) and it will sync with Translation Server(Layer 2) via unified Translation sync protocol. Translation server should be online and real time(It's important because that problem which sara and Arash described in comment 12 and comment 11). Yes. It will be stricture next generation of localization systems can be develop by Mozilla!
Dwayne, what's your comment about implementing three layer localization system on pootle?
Flags: needinfo?(dwayne)
Dwayne and Matjaz, please read comment 9 to cooment 16.
Hi Amir and Persian team, On the first point of moving. Lets let the whole team decide what works best for them. And that seems to have been discussed. On some of the other points raised: As Jeff pointing out the best place to report issues on Pootle is always going to be https://github.com/translate/pootle/issues - I try to troll through bugs on Bugzilla, but realise we get similar bugs from LibreOffice, etc and others who use Pootle. So its always best to report this upstream. And as Jeff says we actually are able to prioritise and work on features. Sorry about any that I did look at but which slipped under the radar, I'll try harder to move such ones onto Github in future and do a quick review of those items after I respond to this. Our current milestones are mapped out here https://github.com/translate/pootle/milestones and I'll take the hint of Jeff and make things a little more clearer at a high level. Please note that anyone can contribute to Pootle, its an open project. We also work very hard to prioritise Mozilla related issues.
Flags: needinfo?(dwayne)
Hi Amir, With regards the three tier system you propose. That sounds a bit like Pootle ;). No seriously, Evernote uses Pootle exactly like that. They use Serge to handle comms between layer 1 and 2, Pootle in 2, and Pootle to give and receive work that would have been done by 3. But I'm sure that isn't exactly what you mean. I don't think there is inherently anything wrong with the approach. A team that did try this approach is Solas, but they seem to have had limited success. My suspicion is that the effort of trying to define interfaces takes so much time that making a good localisation tool slips out of sight. In terms of ideas and bugs that are being worked on or thought about in Pootle that would address some of your ideas, this might help a bit: * Expanding import export or upload download. This works well now and has got better as it can handle what to do when there are conflicting changes. We have looked at how to expand this to talk to other services. Talking to any tool would be possible. Then Pootle is essentially acting as your Layer 2 and as long as tools are able to import/export with Pootle they can play a Layer 3 role. * Pootle FS - this will land in the next iteration of Pootle. This essentially allows us to further abstract ourselves from Layer 1 in a safe way. In our first iteration Pootle is able to talk quite well with Git and handle process any conflicts that might occur between Git and Pootle changes. But more generally. At Translate we wrote Pootle because a need existed. We wrote all our tools because of a need. Our aim is to ensure that there is a great set of open source localisation tools. Pontoon already uses amaGama which is a Translate tool. The checks in the Translate Toolkit I'd love to see reused by other tools instead of people rewriting checks that can't then be reused. And I'd love someone to pick up our Virtaal tool which is to me still one of the best offline localisation editors.
Thanks Dwarne. Dear friends, look here http://translatehouse.org/products.html and http://translatehouse.org/about.html all of Pootle and amaGama(Pontoon) are products of translate.org.za and these aren't product of Mozilla. Pontoon's software(amaGama) also is product of translate.org.za or translatehouse.org not product of Mozilla.
Hi Dwayne. Is translate toolkit product like a translation server(Layer 2) which i described? (http://toolkit.translatehouse.org/) I think that system which i want is ready to use, if Mozilla wants and use them: Really i think: Layer 1: GitHub Layer 2: Translate Toolkit Layer 3: Pootle or Pontoon or Virtaal Is it correct?
Flags: needinfo?(dwayne)
Hi folks, We should move this topic out of Pontoon component, because it's not so mouch about Pontoon developement as is about Mozilla Localization is general (and Persian). At Mozilla, we let each team decide what tool they use. Amir, your locale uses Pootle for product localization and let's respect that decision. You have to admit that prioritizing strings is a nice feature to have and it's not available in Pontoon yet. The only project I would still encourage you to consider moving to Pontoon at this point is Mozilla.org (Web parts on Pootle). You get all the strings exposed on Pootle, but only the ones on the web dashboard are needed for your locale: https://l10n.mozilla-community.org/webdashboard/?locale=fa. Pontoon exposes only the relevant ones, so you never translate strings that aren't even required. But again, this decision is up to your team. I also need to correct a few things that have been said about Pontoon: 1. Pontoon supports terminology & suggestions based on previous work. It gives you suggestions from Pontoon, Mozilla projects (Transvision), open source projects (amaGama) and Microsoft projects. 2. Pontoon uses quality checks, reused from Translate Toolkit. 3. User permissions in Pontoon are basically the same as in Verbatim. So we have team managers who manage permissions for their team. We no longer use Mozillians API like we did in the past, because it proved to be a joke. 4. "The Pontoon team consists of one person, Matjaz." This one made me laugh. :-) We only started working on Pontoon as a full blown localization tool last year and trust me, we wouldn't be able to have a replacement for Verbatim in such a short period of time. We're a team of three people. And dozens of contributors have contributed code to Pontoon. Oh, and the best way to learn about our roadmap and priorities is to filter the bug list by Priority: https://bugzilla.mozilla.org/buglist.cgi?product=Webtools&component=Pontoon&resolution=---&list_id=12694285
Component: Pontoon → fa / Persian
Flags: needinfo?(m)
Product: Webtools → Mozilla Localizations
Version: Trunk → unspecified
Sorry for that last one, Matjaz. I didn't realize that Pontoon had more resources added to it. I remember there being interest in Whistler, but didn't learn more after that. That's what I get for not being as involved in Pontoon anymore, I suppose :-\ Amir, there's been mention of the same system you're desribing in the dev-l10n mailing list just today, but with the repositories serving as the translation/localization server. Soon, Pontoon and Pootle will sync with both source and target locale repositories so that localizers can contribute to l10n projects from any tool they please and have their changes synced across tools/platforms.
Thank you everyone for providing this great piece of information and sorry Matjaz for our misunderstandings. I have just requested Mozillians to be moved to Pontoon in order to solve the problem that Matjaz mentioned. All other active team members also agreed on moving Mozillians from Pootle to Pontoon. Thanks again :)
Hi every One, i can just say this, Let's go to the Pontoon. :)
Webparts is now part of Pontoon. Closing. Thanks everyone involved and sorry for too many messages.
Status: UNCONFIRMED → RESOLVED
Closed: 10 years ago
Resolution: --- → WORKSFORME
Thanks from all of you. Finally, after long time discussions, Persian team moved webparts(On locamotion) to and all of most of projects which was on localize.mozilla.org to Pontoon. Also As Jeff described in comment 22, Mozilla and Translate house will make a sync system for Pontoon and Pootle. Thanks again, specially from Jeff, Dwayne, Matjaz.
Resolution: WORKSFORME → INVALID
Flags: needinfo?(dwayne)
Resolution: INVALID → WORKSFORME
You need to log in before you can comment on or make changes to this bug.