Mplayer Revisited

Joe Barr writes "It's been two years since I first wrote about Mplayer. Maybe the fury of the developers/community reaction to the fact that I dared to criticize them for their treatment of users kept me away. Whatever. Now Mplayer has a pre1 version of release 1.0 out there and it's time for another look." Newsforge and Slashdot are both part of OSDN.

Comments

Re:Of course you were criticised!

[Anonymous Coward]

"I know it's illegal, but could you add DVD Support?"
"Why won't it Play DVD's?"
"I can't watch my DVD"

That's enough to make me cynical.

Spot on!

[idiotnot]

I've managed to compile and successfully run GNUMach and GNU/Hurd from CVS. I know my way around building code. But mplayer is still a pain in the ass. Seriously. And when I read the forums, I didn't dare ask a question. The developers' attitudes represent one of the most valid criticisms of the Free Software world -- support is fleeting.

As for using the software, it works pretty well, and has steadily improved. But I don't build it anymore -- I use the unofficial debian packages, and they work pretty much flawlessly.

Don't flame the devs

[arvindn]

Doesn't the author understand how the Linux/OSS community works, or what? Its not the devs' job to make shiny installation druids that you can click through. That's what distros are for. If you want to compile software, be prepared to do your homework. If not wait for the .deb to become available or subscribe to RedHat network etc.

Gimme a break.

Re:Don't flame the devs

[pgrdsl]

Once upon a time, many years ago, free software was wild and untamed. When you downloaded a tar-file, it may, or may not, unpack in ".". Sometimes it was a shar(k) and you had to hope it didn't eat any of your files when you ran it.

And then, once you had your nice shiny sources, you could compile it. After, of course, hand editing the Makefile. Oh, and the other one. And, damn, that silly config file. And then you would fix all the compilation problems because it had been developed on a different version of Unix, and used strdup.

But, in those days, men were real men, and hackers were real hackers.

People complained, whinged, sent patches, and things improved.

And then, miracle of miracle, automated configure and build scripts came. There was the perl one, which asked lots of questions that nobody ever knew the answers to. Then there was the GNU configure scripts, which tried out things and found what worked.

And, yea!, verily, time was saved all round the world. Things started to work. Porting to other platforms became simpler. Installation was tamed, and things went were you expected them to.

What I'm trying to get at is: the same argument about "be prepared to do your homework" was used years ago pre-autoconf. Nobody would even think of going back to hand-editing all those Makefiles.

It doesn't take a vast amount of effort to get a sane build and installation process, and the amount of time it saves everyone (including the developers themselves) is massive.

With distros it is less of an issue for mere mortals, but the benefit any open source project will get from being easy to configure and install is that developers who are willing to chase bugs will do so - because it takes no pain to build.

Re:Of course you were criticised!

[curne]

Much amused.

But I cannot say I agree. I find it refreshing that a development team develops a high end program that requires some seriousness of people to use. It is becoming a widespread myth that free software developers are little Tele-tubbie-happy people just sitting on their asses coding for hundreds of idiots that luckilly flock to their mailing lists.

MPlayer is a fantastic program (along with other fantastic media programs running on Linux & Co) so many users want it to work for them. And I think that the MPlayer core team acknowledge that but when you for time number 796 get an email reading 'I problem compiling, Please help!!! Is it bug?' with no log or dump... well the coding gets sour. So I can understand that criticism is difficult to take. Especially when it seems as unfounded as the first review.

Who is this doofus?

[Ih8sG8s]

This guy's article is self congratulatory and self seeking. I don't care what run-ins he's had with the developers in the past.

A terrible review where he actually admits to not really checking it out fully, but still manages to come to the self-affirming conclusion that he was right all along, and takes the opportunity to take a personal jab at the project.

The only thing I learnt from this article is that the writer is bitter, and lacks tact.

Barr and bias

[0x0d0a]

Won't this seem daunting to the end user (labelled automatically as stupid), having two different applications, with individual libraries, for doing the exact same thing.

No. Xine should be installed on systems intended for non-techie end users. Mplayer is not a particularly great choice for non-techies. A'rpi is very much opposed to the idea of binary distributions (since it means that things may run slightly slower on a given system), and Mplayer can support so many things that to set up everything required for full support during a build can take a long time. It's less bad than building GNOME or KDE, but it's definitely not an "rpm -Uvh" either.

That being said, I use mplayer exclusively, and love it to death. It's keyboard controllable, can be used without one of those godawful "fake media player" UIs, is faster than anything else in existence, and has support for just about every interface and codec under the sun (that Open Source folks can get their hands on or reverse engineer). Those of you not familiar with Barr and A'rpi (the lead mplayer developer for a long time) should be aware that the two intensely dislike each other, and have flamed each other for ages. Regardless of how good Barr is in most areas (and this review seems pretty reasonable, saying that "mplayer ain't a great choice for Linux newbies", which is definitely true), keep in mind that he's quite likely to have some bias, as A'rpi does when talking about Barr on the mplayer website. I take both with a big, big grain of salt.

Perhaps some collaboration between MPlayer and Xine should occur.

It does. Of course, it's full of people flaming each other for not giving sufficient credit, but the two projects have shared a *ton* of code in the past, and is the only reason either of them are as good as they are.

Re:Here we go again...

[dasmegabyte]

His argument:
"I am a user savvy enough to be running linux. I am bright enough to fix problems. And yet, it was not easy for me to install this application. Therefore, it will be even harder for somebody who is new to Linux."

Your argument:
"What an idiot! He should have read the acoryphal poorly laid out document! Things are easy if you do all the chores perscribed to you by developers with no talent for technical writing and different systems than you!"

My argument:
"RTFM is not a valid complaint. Windows software installs without a manual. It does not expect you to RENAME directories after installing things to get them to work. It does not expect you to KNOW what codecs you want to use and already have them downloaded. It allows somebody to do what they need to do before hacking the source code of the underlying software. Why can't linux software do this as well. Oh right. Because we're better than them."

Re:What about other software?

[Qzukk]

Won't this seem daunting to the end user ... having two different applications, with individual libraries, for doing the exact same thing.

So, you think we should all just go with one software project and kill the other? Which should we kill? Did you know that Xine did the GUI thing first? Mplayer has been the leader in figuring out how to play new formats (especially Quicktime codecs)

So if we had killed xine, would mplayer have developed a gui without competition? If we had killed mplayer, would we still be griping about not being able to watch Sorenson encoded movie trailers?

What future benefits will this competition bring?

To the users, I say: Try them both and stick to the one you like. This doesn't require genius or even much intelligence. If you can't get mplayer working, then xine. If you'd rather not deal with a GUI, then mplayer. (personally, I hate hunting through a list of files for the video I want, when I could just run mplayer -fs filename.avi and get full screen goodness straight from the start without having to move widgets out of the way or get them behind the video window.)

Re:What about other software?

[Dave_bsr]

Hey, it's the overly critical guy again. I really think you might just be a great, great troll. Either that or just really frustrating. Anyways...response:

For one, groups don't always just get along...I really don't think the xine guys and the mplayer guys would like to just drop everything and work together - both sides have thrown a couple of potshots too many - they don't like each other.

Secondly - competition increases output. It's one of those crazy things about life that two competing groups seem to get farther than only one alone.

Third - more people on a project does not (neccesarily) more code make - adding more developers means you have to merge maintainers, and "people in charge" - etc - it has to be very well organized to get O(n) increase in production... You can't just throw people from well-defined, properly working groups together...it doesn't work! Good people can be left out (and unused)...

Fifth - mplayer and xine do share some libs. I'm almost positive that xine is using some of mplayer's win32 code, but i'm not sure - but the logical thing is that they are both open, and why wouldn't one project "borrow" code from another, if it was great. Emulation is the sincerest form of praise - I think that's how that goes.

Mplayer is a great project, xine - last I checked - was...decent. I think seperately you get _more_ output - and thus having two seperate groups is better for you. *shrug*.