Why it sucks to be a poor, stupid Windows user

[ChristianHJW]

http://article.gmane.org/gmane.comp.video.mplayer.user/12833

Taken from mplayer-user ML. One of the main developers of mplayer is writing there :

D Richard Felker III wrote :
> What about Matroska?

I dunno, I stopped reading after the "inspired by XML" part... :))

Seriously, it looks like it suffers from serious bloat and nonsense
that make things unnecessary painful to implement. I have no idea how
efficient it is or whether it handles seeking, error recovery, etc.
well, but since it seems to be written by stupid DShow fans like MCF,
I imagine it's brain damaged in that area too.

[Ookami]

And? There are (and were) many TMF/MCF/Matroska sucks people out there...

The best solution/cure is to come out with a good container format that actually works and will convert all those non-believers...

But, even if it is very good there are many other factors to look at. For example, IMO, the best Win98 player (under XP I had some probs with it, altough I haven't tested it on my new PC) is Sasami2k, just perfect. But, when I look around even a average player like GDivx is more used and hyped by those average mags/users/websites... Why? No one knows...

Like always, if someone tells you you're stupid, the best reply is "Yes,yes!" and keep up your work. Hopefully, you proove that he/she is wrong, if not, you're too stupid to notice it :D .

IMO, instead of investing energy in such x vs. y wars, it can be used on developing and testing. The Xvid devs are a very good example... This wasn't directed at you or anyone specifically, just an observation by myself.

All the best,

Mijo.

[int 21h]

All of these revolutionary containers have suffered from such hype its unbearable.

[ChristianHJW]

Originally posted by int 21h All of these revolutionary containers have suffered from such hype its unbearable.

We chose a different approach than most other developers do. Instead of simply doing 'something' and 'giving' it to the world afterwards as a 'gift' we wanted to bind all camps in, normal Linux people as well as the Gstreamer guys, Windows users and Open BeOS folks also. All were invited to contribute and to comment what we have done, and to make suggestions where to improve. This needs a certain amount of advertising, to make people aware of what we are trying to achieve. If you call it 'hype' i am fine with that, although i am failing to see how it could be 'unbearable'. The number of posts about MCF/matroska to the mplayer lists shouldnt be higher than 5 or 6 in the last year, and that is to be seen in relationship to about 100 emails / week showing up there. Unbearable for them ??

Today we know it was the wrong thing to do after all. This doesnt mean we wouldnt do it again this way, because we still feel its the right and sensible way to start such a project normally. But the way how things worked out teached us a few important things about opensource development. In the end nobody is interested to help achieving a common goal ( Gstreamer people are really the exception from the rule here i guess ), but is just keen to get his name on the most used app/codec/proggie quickly, and to earn a lot of admiration for that. A quick hack that makes new things possible is always of a higher value than a solid solution, taking its time to design, document and finally develop/code.

mplayer people were invited to help us developing MCF/matroska, and in fact there was some valuable input from Arpi and Atmos in the beginning ( no Mr. Felker or Mr. Niedermeier posting there ), but it finally went into a flame war why we decided to use C++ instead of C for the main library, so the traffic on the list stopped. All to be found here : http://news.gmane.org/thread.php?group=gmane.comp.video.mcf.mplayer and here : http://sourceforge.net/mailarchive/forum.php?forum_id=9680 .

Now they tell us our specs are 'crap' and we are 'stupid DShow fans' doing 'brain damaged' things. Nice manners they have, the opensource developers. Highest possible respect for the work of other people, as it should be.

They are planning on making their own container format now. Michael Niedermeier ( heck this guy IS a genius, no doubt here ) had a first draft of a spec done overnight. It sooo easy with the background these people have about other containers, and if you dont have to ask anybody else, just do your own thing and then give it to the rest of the world. They wanted to call it 'EGG' for 'Extensible OGG' , but dropped the name again to avoid confusion .. lol

We certainly wont let ourselves being influenced by this behaviour. libmatroska is finished, uploaded to CVS and will be relased this week, and so is the documentation ( please, show me one single opensource project where the docs are released on day 1, together with the code ! ). Cyrius is working on matroskadub and kromyx on the DShow parser, so we should have something for the alpha test team by the end of this week, latest next week ( we said end of february ).

I just hope not all developers being interested in making use of a new container format will judge that quickly about matroska than the mplayer people obviously did .....

[gabest]

Hehe, don't be surprised of their attitude over mplayer, they are always like that... Anything heavier than simple c with the standard lib is just too much for them. Ok, nearly everything is possible in c too, but also in assembly or pure machine code, it just takes a "little" longer.

[Neo Neko]

The blessing and curse of opensource development. Nearly unlimited human resources and nearly unlimited oppinions as to how it should be done. :P

The sucessful ones are the ones with a vivid vision and force of will to get it done. You know what you want. Just do it.

If they were given the invite to contribute and refused to then they have no room to bitch.

If they were or are going to make a format based on the OGG container and call it extensible; especially in light of the fact that OGG itself is unfinished. What does that say about their though process? Not the brightest in the bunch. :P

Just keep on trucking Christian. Perhaps you guys could borrow a slogan from Microsoft. But unlike Microsoft you could adhear to it. "Make it first. Then make it better than anyone else." With Microsoft it is more like "Make it first and threatten anyone who does something similar with legal action. Or buy out those who did it first and or run them into the financial ground."

[BlackSun]

how cool, I'm brain damaged and a stupid directshow fan and on the top: I'm using windows :)

This kind of guys make me laugh loudly :sly:

[Beelzebubu]

Christian, don't forget two things:
1) mplayer developers are -censor-
2) you've got to proove yourself.

1 is obvious, won't go into that. 2 is the important thing here. I don't care whether a container format is bloated, basically. Have you ever looked at MPEG? From a programmer's point of view, it's so full of features and at the same point so hard to get right that it just plain sucks. But people use it! AVI, on the other hand, is so @$^! easy that nobody can do it wrong, but severly limited in its extendibility.

Matroska isn't aiming to be a look-a-like of either of them, which is good, because we don't need that. MPEG and AVI will stay the standard container formats for the next X years, that's pretty certain. After that, Matroska or OGM will take over parts of their market. Maybe even all of it. What each of them needs to proove is that A) users like it, use it, can click buttons and get a movie out of it that will play somewhere else and B) that the format is technically OK and achives the goals it needs for the users to be valuable to them, and better than its predecessors. In some respects, this requires "bloatedness", as these people call it.

How? Well, in the form of being able to embed any number of information streams (subtitle, audio, video, metadata, you-name-it), being able to be used in embedded devices AND on desktop and server systems (streaming!), being easy to convert to and being fairly easy to understand. libmatroska actually isn't of that much importance. I suspect that 50% of the people will write their own reader or will "fork" your code into their project. Why? Because you want as little dependencies as possible, clearly. Using a BSD-style license would help a lot here. You'll dislike it, just as me, but that's just how it works. Others will want to write their own reader, so you need good documentation on how it works and on how to use libmatroska AND on how to write your own matroska-reader. Same goes for OGM, obviously. The biggest requirement is of course extendibility, which is where AVI totally blew up (well, it was extendible, but these were all ugly hacks and required recoding of demuxers/readers).

The key poitn is to proove them that all that efforts is worth it, or to do (parts of) it yourself, like the OGM people are/were doing. Yes, write a matroska reader, get it into the Nimo codec pack or even in standard windows (well, you gotta set goals ;) ). Make a linux matroska reader and get it into gstreamer and xine (and maybe even mplayer). Proove us that it's worthy, and we'll use it. This requires good code, good PR, good design, etc.

Just proove us that it's good enough, that's up to you. :).

[TactX]

Originally posted by Beelzebubu
1) mplayer developers are assholes

Come on.
You can say that one mplayer developer is an a**ho**. Maybe you can even say that some mplayer developers are a*sh**es. But if you say that _all_ mplayer developers are **sho***, it's quite clear who the real ******* is.
Sorry, this has to be said.
And btw... If it needs as**ol*s to write the imho best player for Linux, can't you live with it?

I can...

[Ookami]

Ok, people, asshole, here, asshole, there, please keep it,er, nice and clean.

No insulting, please. We/you can discuss this without such insults, especially if the other side is not here to tell their side of the story.

Thanks.

Good night,

Ookie.

[ChristianHJW]

Originally posted by Beelzebubu Just proove us that it's good enough, that's up to you. :)

Thanks BBB. I promise we will work hard to achieve this. So, may we expect support of our format in gstreamer one day ;) ?

I was posting a few job proposals on sourceforge lately, and although a few people want to help with porting libmatroska to C ( funny enough .. lol ) there was nobody replying to my proposal about making a matroska plugin for gstreamer yet :( .... maybe you could help us :D ??

[Beelzebubu]

Originally posted by TactX
And btw... If it needs -censor- to write the imho best player for Linux, can't you live with it?

Well, no.

Opensource is about more than just making some code available...

[TactX]

Originally posted by Beelzebubu
Well, no.

Opensource is about more than just making some code available...

OS is only about making code available. If you release crap as OS it's crap, but it's still OS, isn't it?

Well,
If you don't like the mplayer developers, O.K.. But if you say s.th. like that....

"mplayer is mostly a piece of crap that has to be mentioned because it exists, but besides that, there's no point in looking at it"

... you're obviously biased. There are thousands of users that are very happy with mplayer... more happy than they are with xine (your obvious favourite)...

And xine does not have the fastest encoder, right?

[Beelzebubu]

Originally posted by TactX
OS is only about making code available. If you release crap as OS it's crap, but it's still OS, isn't it?

Technically. But OS (what it _stands_ for) is about cooperation, building a community, working together. Making some code available is something anyone could do, but working together in creating software in a team, with user<->developer communication, that's where the advantages of OS as we know it versus closed-source become really obvious.

Well,
If you don't like the mplayer developers, O.K.. But if you say s.th. like that....

"mplayer is mostly a piece of crap that has to be mentioned because it exists, but besides that, there's no point in looking at it"

... you're obviously biased. There are thousands of users that are very happy with mplayer... more happy than they are with xine (your obvious favourite)...

I'm biased, definately. But I have other reasons to say that mplayer isn't worth looking at. For example, it's external programming interface isn't good at all. For mplayer itself, this doesn't matter, since the only goal is to provide an internal lib on which they can build a player (mplayer) and an encoder (mencoder). It's API is badly documented (or mostly not documented at all), which actually goes for most of the code too. This makes cooperation and code reviewing really hard. One of the things it shares with other projects is the forking of other projects (xine does that too, for example, as do most other media players, even GStreamer's plugin package includes one other project in the tarball (ffmpeg)), but it doesn't keep these external versions up-to-date and doesn't send patches back upstream automatically. Now of course, the source is available so you could just do diff -u, but again - this isn't what opensource is intending to be, it's just the legal rules that have to be there. The whole idea is to go much further and let things go automatically. There's basically no community building around mplayer as an effect of all this. It has a large number of users, and it can do FIFO-data-exchange with a number of other packages (e.g. mjpegtools), but it's still all commandline. There's no runtime library backend that's good enough to choose when you're building a multimedia application. Now look at xine or GStreamer, they have several external projects. Xine has totem, qxine, enix, etc.. GStreamer has Juk, Rhythmbox, Aleader, etc.. mplayer has... nothing. My conclusion: from a programmer's point of view, mplayer likely isn't what you want.

As a player, it currently stands out in performance only, though. It doesn't look good imho (though that's a personal thing), and it's currently just too hard to start up for anyone but a computer expert. I'm not talking about average-joe-that-installs-linux or RMS' Aunt Tilly, I'm talking about my parents and the other 99% of the computer users in the world who don't even want to know what commandline is. Xine and GStreamer's Media Player are fully graphical in both configuration and in file selection. For mplayer, this doesn't work *at all*. Yes, I know that it has a GUI, but that GUI doesn't work at all. Last time I tried it, the GUI wasn't even able to be clicked on while a movie was playing. Well, that's useful! Not good enough for my parents and/or me - and that's the audience you'll have to target in the future if you want Linux to become big. My second conclusion: from a user's point of view, mplayer is very interesting if you're more than the average windows user that knows CLI. Besides that, mplayer likely isn't what you want.

[rhetherington]

Originally posted by Ookami
Ok, people, asshole, here, asshole, there, please keep it,er, nice and clean.


Yeah it's never pleasant when people don't wipe properly. :D

Anyway i think these arguments are extremely silly. One mplayer developer slags off matroska and Christian goes in flaming and then proceeds to wander around the 'net making a massive fuss about one persons opinion.

The flaming just aint gonna help. It's not going to change the developers opinion and by attacking one of them the rest will close ranks against you and what you represent.

You then turn it into a pride thing where the developers are determined to beat your offering and show you how it should be done.

Not that i'm against a better container or anything, just that i don't think it should be motivated by pride.

Moritz has said he'll still code the Matroska support for mplayer, so there's no worries about it not playing them.

Whether i'll use Matroska or not i don't know. I don't care what container i use as long as it's the best tool for the Job. If Matroska is the best tool then i'll use it, if the mplayer container is better then i'll use that. Since neither of them are currently available i'll just keep on using Avi and Ogg for the time being.

BTW, AFAIK the mplayer container isn't anything to do with Ogg, .egg was just a suggestion someone made for the extension. I suggested .ova in keeping with the egg theme (and it satisfied my Anime interests too ^_^), but i think .nut is just as good, it's a hardy container for the sweet goodness inside. *g*

[robUx4]

Here is my little contribution on what has been said.

First, I use MPlayer on Mac OSX and it's the best player you can find on that platform (along with VLC/Videolan). And as a Mac "port" it is as easy to install as clicking on the program once it's decompressed...

I also consider that for an OS project to succeed it requires more than a working code. Working hard, pushing hard is another factor. But the main important IMO is : listen to what the users say/want. And AFAIK neither Xiph or MPlayer ever spent much time doing that. They only listen to their small user circle but not get very different opinions/needs (until recently where Emmett and Monty appeared here (and still don't lsiten)). And that's the main reason why matroska/MCF took so much time. Hopefully that will be the key to our success :D

[ChristianHJW]

Originally posted by rhetherington Anyway i think these arguments are extremely silly. One mplayer developer slags off matroska and Christian goes in flaming and then proceeds to wander around the 'net making a massive fuss about one persons opinion.
I clearly overreacted on the list. The last week was extremely hard and enervating for me, so i made a big mistake by replying, especially with respect to the wording i was using. Its just, you feel so helpless in such a situation. Although they over and over repeated they 'know very well how matroska and MCF are like' ( in fact, both are completely different now, although having the same roots ) it became clearer and clearer they didnt have the faintest idea about it. Felkers email was showing the real reason why they were against is, its mainly based on prejudices because its coming from the bad, bad Windows world.

The flaming just aint gonna help. It's not going to change the developers opinion and by attacking one of them the rest will close ranks against you and what you represent. You then turn it into a pride thing where the developers are determined to beat your offering and show you how it should be done. Not that i'm against a better container or anything, just that i don't think it should be motivated by pride.

Well, they didnt even adress to us trying to find out if they were right about their assumptions when they decided to make their own container instead. You are helpless in such a situation, as they dont even give you the chance to come up with arguments that could change their minds.

About the motivation to make matroska, well, pride is a factor in that that shouldnt be neglected ! Why do people invest their free time in such a project, without any payment or such ? First its the joy to create something, 2nd its the pride you feel if its working and people like it. Nothing is more demotivating than other developers stomping on your work with their feet, without any logical reason to do so.

Moritz has said he'll still code the Matroska support for mplayer, so there's no worries about it not playing them.

We have 3 new dev team members now ( all joined the team because of my advertising on sourceforge ) and 2 of them will port the complete library to C now, so Moritz shouldnt have any problems supporting matroska for mplayer. I am glad he wants to help us, he is a good guy !

Also i have high hopes on the gstreamer guys, they are much more open for new, unusual solutions it seems, plus they dont judge that quickly about the work of other people.

Whether i'll use Matroska or not i don't know. I don't care what container i use as long as it's the best tool for the Job. If Matroska is the best tool then i'll use it, if the mplayer container is better then i'll use that. Since neither of them are currently available i'll just keep on using Avi and Ogg for the time being.

We will work hard and harder to make our users happy. I still fail to see why a container format had to be 'simple' . A coder may like to have a 'simple' and 'clean' code, resulting in an 'elegant' solution, assuring him the respect of other developers. The users are not interested in such things, not at all. They dont see the code, they dont understand the structure, and they arent interested in that either.

What they are interested in is what the container will allow them to do. In the end, and i am convinced here, the bad bad 'complexity' of matroska will make it attractive for the potential user. We will see.

Our next steps now will be to make sure the basic functionalities will work on Windows, so the container can be used. The future will bring if people like it or not.

BTW, AFAIK the mplayer container isn't anything to do with Ogg, .egg was just a suggestion someone made for the extension. I suggested .ova in keeping with the egg theme (and it satisfied my Anime interests too ^_^), but i think .nut is just as good, it's a hardy container for the sweet goodness inside. *g*

Well, yes, this new container will not be based on Ogg at all, in fact it became clear from reading through all the numerous posts that they dont like Ogg as well ....

[midiguy]

people are WAY too concerned with technicalities and crap.. as long as it works like it is suppose to, and it's fast enough, and it's not impossible to use or to figure out, and is EFFECTIVE in what it is SUPPOSE to do, then people will be happy with it. Honestly, do you think the end user will really give a **** or decide not to use the container if you guys decided to use C++ instead of C for the main library? no. they won't care at all. do whatever works best for you, and all your hard work will pay off when you feel the satisfaction of allowing people to use such a nice container format.

[robUx4]

In the other hand if we want the format to be successfull we have to either code every app by ourself, or get support from other developpers, that we'll have to convince.

[midiguy]

Originally posted by robUx4
In the other hand if we want the format to be successfull we have to either code every app by ourself, or get support from other developpers, that we'll have to convince.
I have to disagree with that. Sure, developers need to be able to support the format once it is out there, but if it is an effective format, it will pick up by itself, and will get the support from developers. If it is good, they will be convinced, and will support the format. It is much harder to rally support for a format that is not yet compelte and is still under heavy development. if developers of other softwares (ie: mplayer devs) have views about a project, they can go ahead and give their opinion, that is what OS is all about, but to start calling people "stupid dshow fans" and to start flame wars and the etc. is just insulting to all the hard work that people ahve put into the project. I call that UNconstructive critisism. rather than flaming back, I would just ignore people like that, not get discouraged, and continue work on the project.

[robUx4]

Don't worry, we continue to work and are as motivated as ever (if not more since we're close to real world applications) !

But history shows that better technical solutions don't get used by the majority of people (look at Mac or BeOS). It requires more than good technical specs and working code to be successfull. If there is a solution it's work, user-friendlyness (?), price and PR. We're working on all fronts :)

[ChristianHJW]

quick update for all mplayer fans :

Moritz Bunkus is planning to work on the Linux counterpart of Suiryc's matroska muxer tool this weekend, and he will use the C++ lib for that.

At the same time we have a dedicated team of 3 people now that will do the porting of the complete lib to C, so mplayer support shouldnt be an issue then anymore .... hopefully ...

[kempodragon]

@ ChristianHJW:
Just wanted to pass a along a few words of encouragement. I've seen your posts in other forums and I must say your project sounds intriquing. About that silly "bloat" argument, my definition of bloat code is stuff that has been just tossed together with no thought of optimization or clarity. ANY language, even assembly (my favorite), can produce bloated code. What matters is how you much time you put into streamlining the code.

Your goal of replacing the avi format is laudable, but in order to do this your container must be able to be used for capturing with Huffyuv, editable with VirtualDub or something similar, can be supported by media players, and finally, support error correction for use in Mode2 cd's, which is the avi's main weakness. I'm always willing to try something new, but it has to be able to stand on its own. I mainly video capture, and am only able to keep a few source files on hand at any time. This means I need a proven format for editing. Meet all of the above requirements, and you will have truly sounded the death knell of the avi format. Editability is still the avi format greatest strength and that is what your format must match. Look at MPEG2, for playback it is great, but doing any editing is a royal pain in the you know where. :devil: I look forward to giving your project a fair shake when it is ready.

[robUx4]

We have made matroska with all that in mind, and we can produce files now with a modified version of VirtualDub. So yes, you should like Matroska when it's finally there :)

[ppera2]

OK, OK... we will test it gladly..

Just don't disappoint us :D

[ChristianHJW]

Originally posted by ppera2 OK, OK... we will test it gladly..Just don't disappoint us :D
Without a working DShow parser our hands are bound with respect to releasing something, and also our test crew is bored :( . There is not much fun in creating files you can only play in matroskadub's preview window ....

As a result, we are searching for more alpha testers now, and these people should have both Windows and Linux installed, and be willingly to completely screw thier mplayer installation with our crazy C++ playback patch :D :D ... lol

If anybody is interested please drop me a PM ....

[Mosu]

Hehe :) My patch won't screw your complete installation. At least not intentionally ;)

At the moment I would say that I still need a week or so before a more wide-spread testing would be useful. E.g. I haven't done testing with files created with Cyrius' Matroskadub... and there are still a lot of issues with my own tools.

But I'd appreciate testers just as Christian has said. Especially if they're running something other than Linux/x86, e.g. BSD.

[robUx4]

I can try testing it with mplayerosx :D