One day with LXDE
The LXDE project noticed its Romanian translation is pretty much dead in the water and considered dropping it along with a few other unmaintained translation, in a clean-up effort. rsandu ignited the alarm and in an effort spearheaded by ajoian and alexxed, all the strings got translated in less than 2 days, now there is still needed a validation from alexxed and upstream submission. It was awesome to see so many people lending a hand, including a couple of people at their first translation contribution (trizen, you kick ass!).
All this community effort was impressive and motivated me to an experiment: spend an entire day day with LXDE and evaluate what I like and dislike about it, evaluate how I could deal with it as the primary desktop. Is not my first experience with it, I toyed with LXDE in the past, but usually after 10 minutes I went back to the comfort of a desktop I am used to, but beyond the community enthusiasm, it was a long reaching study, if something goes wrong in the GNOME land, is useful for me to know what the alternatives are.
Installing is fast and easy, just:
yum groupinstall lxde-desktopWhat I like about it:
- the desktop is small and fast. Really fast
- it includes a few basic applications, which are also small and fast: image viewer, text editor and so on
- with a bit of customization, it was not hard to bring it to a familiar 2-panels layout
- it has in place all the known paradigms: panels, buttons, menus. It is not distrupting
- is painful to use a file manager without gvfs, I need to transfer files over ssh/scp on a daily basis
- customizing the LXDE panels is hard and limited
- the size of fonts and icons is ugly and hard to adjust
- no drag & drop adding shotcuts to the panel
- icons on the desktop are always auto-arranged, they move all the time and you can't place them where you want
- no integrated screen capture with the PrintScreen, you have to use external applications
I can see me using LXDE as my main desktop? Maybe on one of those underpowered ARM tablets and smartbooks... but if the computer does have the power to run my apps (I would NOT pay money for a computer unable to run GIMP and Inkscape), then it should be pretty much able to run also a normal GNOME desktop.
From an usability and customizability point of view I find it much behind a normal GNOME 2.X setup, but compared with GNOME Shell (as previewed in Fedora 12) it would be less of a productivity drop and a lot less annoying. However, in the case worst nightmares come true, LXDE may not be the best exist choice. Will spend another day sometime with Xfce and report the findings.
Posts
I find Xfce already a *tremendous* improvement usability-wise (probably because of following the GNOME HIG to some extent) - compare e.g. Thunar to PCManFM.
ReplyDelete@Vincent: today is my "One day with Xfce" day
ReplyDeleteYou can install sshfs and mount things like so:
ReplyDeletesshfs me@remote.server.com:/home/me /mnt/remoteserver
and then you can have seamless invisible ssh file transfers with any tool.
(/mnt/remoteserver has to exist; I chown these mount points to root:fuse and chmod them 775.)
I had KDE break on me. I needed a window system in a hurry, I tried several. Fluxbox is very minimal and very quick to setup. I now have it on all my boxes as a good fallback.
ReplyDeleteThanks for this review, Nicu.
ReplyDeleteFirst the good news: PCManFM has been completely rewritten and will be released in the first quarter of 2010 depending on the feedack from the testers. It now has many things you miss such as free placement of the deskop icons and virtual filesystem support for both remote and local filesystems.
And now the bad news: Romanian translations are not fully done. I'm sure ajoian and alexxed will finish them soon, but it's hard to work with them because they refuse to work with LXDE's pootle and insist on using their own tool called narro. I'm not saying that narro is bad, but it adds additional overhead for us and slows down the whole translation process. It is more work for our translation coordinator Martin and it makes it harder for new Romanian translators to join because they need to work with narro which is not documented anywhere. There is no way of dealing with conflicts between pootle and narro and even if we found a technical solution for this, we couldn't do it because there are no Romanian translators working upstream on LXDE. This means we might be forced to remove Romanian from our pootle - and this is just what Martin wanted to avoid when he wrote his blog post. (Note: We never wanted to drop Romanian translations from LXDE, only from pootle).
Martin and me tried hard to convince ajoian and alexxed, but they didn't give much about our arguments. Nicu, please tell your friends if they want to contribute to LXDE, they should contribute upstream instead of making yet another downstream project.
On the gvfs thing: You can mount remote locations with gvfs-mount and then access them through ~/.gvfs. Once they are mounted, you can access them with PCManfm too and if you mount them with gigolo, it will automatically fire up PCManFM.
ReplyDeleteCristoph, I don't understand the problem with Narro and Pootle: it is just a tool to produce .po files, if those are valid, what's the problem?
ReplyDeleteIs easier for Romanian translators to use Narro, since it is already used for translating Firefox, Thunderbird, Seamonkey, OpenOffice.org, Fedora, openSUSE, GNOME, KDE, GIMP, Inkscape, Wesnoth and more. It is taking advantage of an existing base of translators and a large database of existing string suggestions. The application is easy to use, with a short and smooth learning curve.
The news about PCManFM are awesome, it looks like I will give it a try soon.
PS: I see in the stats the translation is 83% done, the rest is translated but not yet validated.
Nicu, the problems with naaro are:
ReplyDelete1. Pootle or narrow are not just for creating po files but also for submitting them to upstream SVN. This is only possible with pootle.
2. The po files from narrow are submitted through the bugracker instead of being generated in pootle. This is more work for us: Check the tracker for new translations, commit them manually to SVN, close the bugs in the tracker and so on. Maybe narro is less work for the Romanian translators, but it is more work for us, the LXDE upstream.
3. There is no way of resolving conflicts between pootle and narrow, nether technically not from a linguistic POV.
4. Because we have no way of resolving conflicts between pootle and narro, we might have to remove Romanian from pootle. This means that we also might loose new Romanian translators because they don't know where to contribute. narro is not documented in the LXDE wiki.
5. By using a tool of their own, the Romanian translators separate themselves from the community of LXDE translators. It is hard to reach them. In pootle we have all users with their email addresses, we can manage privileges and and so on.
6. Fedora has a policy to follow upstream. So if two Fedora contributors want to contribute to LXDE, they should do this upstream.
@Christoph: I think as a free software project, you should first provide ways for translators to give you exactly what you need, that is po files. Having an interface to translate is an extra, but should not be a requirement.
ReplyDeletePersonally I'll refuse to spend 2 days translating in your pootle instance what's already translated in our software instance (reused work). I'd rather volunteer for spending 2 minutes (time required to read a bug that I'm notified via email, test a po and commit it to SVN) multiplied by 70 languages, which leads to 140 minutes instead of two days of duplicate work, I already help other translators in this way in different projects.
So in my opinion, whatever way you allow me to give you my work as a file in a free format (as is gettext) is acceptable. Making me use a software is not.
But that's just my 2 cents, maybe there will be a translator that would spend time to use Pootle to translate from scratch and hope for the best.
The thing is, I contribute to translations only from time to time and I did this so far for multiple projects (OpenOffice.org, Firefox, Thunderbird, GIMP, Inkscape, Fedora, LXDE - of course, not many strings but something). I would definitely DO NOT do that if I would have to use a different website and/or software for each of those projects and maintain tons of different accounts.
ReplyDeleteWe had also the idea that the best thing for FLOSS translation is the harmonization, is the best for the user if he find the same familiar terms in all the applications This is why was formed a community of translators who work cross-projects and cross-distros.
ReplyDeleteAlexandru, nobody forces you to use a particular software. I don't mind if you do the translations with poedit, gnome-translate, pootle, narro or whatever. All we care about is the workflow. It is not acceptable that LXDE upstream has more work which will slow down the overall translation process for all languages just to have the Romanian translators have less work.
ReplyDeletePlease be serious: You wont need to translate all your stuff in pootle again and you wont need two more days, you can import your PO files. Honestly, I don't really like pootle ether, but before judging about it, you should be familiar with it's functionality.
So how about that: Get yourself a pootle account and import the po files yourself. This is not more work for you than submitting them through the SF.net tracker, but it is much less work for us. Additionally, we have a central user management and centralized stats of all languages.
Sounds reasonable? If so, please get yourself an account, add Romanian in your settings, tell me the account name and I will be happy to give you all the necessary privileges to upload your translations.
Thank you Christoph.
ReplyDeleteI had an account and I just saw today that now I can upload too, which I did and we're now at 100%.