Remlab.net
Hosted projects
External projects
Communities
- The VideoLAN project
- SalutFinlande.net
- Maemo Software
You are using IPv4.
You are using IPv4.
Welcome to Remlab.net! This is the home page of Rémi Denis-Courmont.
I am currently working as a Linux kernel and system software engineer for Nokia in Helsinki, Finland. I am also one of the core developers of the VLC media player at the VideoLAN project, the original developer of Miredo, the Teredo stack for Linux & BSD, and the maintainer of the minisap server and ndisc6 IPv6 utilities.
The Atom feed is here.
What is the link between X11, VLC and OpenSSH? Click here to continue.
The official VLC 1.0.0 binaries were unleashed only a few hours ago, and we already have feedback that it behaves rather badly on Windows Vista. This is not entirely unexpected, but nevertheless a bit disappointing given how long the 1.0.0 final release was held blocked by the sole Windows port. But is VLC really that bad on Windows? And why?
A quick look at the forums (which are not available right now), the bug tracker, or even the development mailing lists, makes it look very much like there are many more problems on Windows. Lets face it, this is in great part true. For instance, there are problems with certain Windows-specific parts such as the DirectX video outputs and BDA input (DVB for Windows).
But there are actually quite a lot of bugs on Linux too... The trick is, Linux users mostly, and rightfully, file bugs against their distribution - and the distribution people are supposed to forward them to us if appropriate. You can find plenty of unfixed bugs on Ubuntu's Launchpad or on the Debian BTS, although relatively few of them are forwarded to the VideoLAN team (I will not blame the distros - they tend to be too busy, as we do). On top of that, some Linux bugs are blamed onto users not compiling the software properly - sometimes unfairly. That is an unlikely occurence on Windows.
And then, we have all those tricky problems with certain files, or inputs, etc. Many of them are filed against the Windows port. Probably, a fair share of those bugs affect other platforms too. Yet as they are more Windows users, they are found to occur on Windows by the laws of statistics.
Anyway, the question of how to address the recurrent issues on Windows
remain wide open.
The lack of Windows developer is at the core of the problem.
There are few open-source zealots who care about Windows and are likely
to start developing VLC for Windows.
And as time goes, the issues pile up, which might be discouraging.
Clearly, begging for help does not work.
Taking users hostage by not releasing Windows binaries, in hope of
attracting attention, does not work either.
And waiting for a miracle... well...
A short discussion on IRC today reminded me
that the former two Windows VLC developers were both paid.
Indeed, Gildas Bazin stopped working on VLC when he changed jobs
and his new employer would not have him work on VLC during work time.
He wrote the largest parts of the Windows port, that remain to this day,
until he left the project in May 2007.
Then, Damien Fouilleul, best known for the VLC ActiveX plugin,
left in February 2008 due to disagreements
with my admittedly authoritative decision making
(to my own defense, I need to mention that
those who expressed their opinion agreed with me).
So maybe it all comes down to a company paying some Win32 and DirectX capable C programmer to work on VLC. I am not very optimistic given with the short-term world economic outlook, but if you do have big pockets, your help will be welcome.
After the longest feature freeze stabilization period in the project's history, VLC 1.0.0-final is coming at last.
I think the freeze period was way too long, and as boringly usual,
the Windows port seems like the main culprit.
This delay is not quite a surprise.
Indeed, we had similar issues with 0.9.0.
Also, the Windows port has the most infinitesimal developer per users
ratio by far, since it has -supposedly- the most users, yet practically
no developer at all.
In other words, every one wants VLC to work on Windows,
but nobody is willing and capable of delivering.
The (not very scientific) developer statistics of mine
for the last month are pretty telling:
In that respect, I consider the mere survival of the Windows port to be a great success, and the fact that it actually mostly works nears the miracle. Of course, some Windows users will complain that their critical blocker bugs are not fixed anyway... (never would they accept switchimg OSes though)
Lets them not spoil the party! VLC will never be perfect, as won't any large piece of software. But VLC 1.0.0 will not be the VideoLAN project's variant of KDE 4.0. VLC 0.9 was probably our KDE 4.0 already. VLC 1.0 adds relatively few new features and ground breaking changes, and fixes the most annoying regressions from 0.8.6 to the 0.9 series. Note that the embedded video support is back and at last, stable.
News archive: 2008.
A few of my open-source projects are available here:
Some technical personal opinion (or factual) articles:
