The XMMS Future in an interview with Dev
Hexdancer wrote to us with the latest interview at
theLinux MusicStation with Dev Mazumdar, co-founder (with Hannu, the guy who wrote the original
OSS kernel sound drivers) of 4Front Technologies the commercial Linux/Unix sound driver guys. He's talking about the future of XMMS, and trying to make sound w/XMMS a killer app for Mac/Windows people, as well as the problems with working with some of the high end sound cards.
How is a clone of Windows software supposed to draw people from Windows to Linux? They could just run the original software and stick with Windows.
Perhaps there's something about XMMS I'm missing? Is there anything it provides that doesn't exist in an equivalent Windows tool?
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
Winamp is uncrippled shareware that costs $10 to register. Legally and ethically, it costs $10. In practice, most people use without paying.
As for performance, I don't see a problem with Winamp. It consistantly uses less than 5% CPU on my pII 266, and uses very little RAM. It has a fully customizable interface, and you can modify virtually everything, even including the mp3 decoding (if you want to write your own mp3 decoder plugin to replace Nitrane).
It's never crashed on me.
10 PRINT CHR$(205.5+RND(1)); : GOTO 10
However, if the Alsa guys were to make an AlsaLib, that would be different...
However, this really bugged me.
Well, to tell the truth, if ALSA does its job, OSS is obsolete. If it is what everyone wants, supporting new cards and new features (I like multiple-open of the sound card!!) OSS will go the way of the dinosaur.However, it isn't that way right now, and while ALSA is working on obsoleting OSS, the majority of programs were developed in the time when OSS was the only solution. How else can you support ALSA but by adding its support? (If you need to bolt it on, your program is broken to begin with.)
As for the whole proprietariness.. I'm more mad at sound card vendors than 4front. Obviously if a sound card vendor is brain damaged enough not to release specs, support has to come somehow.. Even so, as ``pragmatic'' as the 4front guys are, they are selling something which should be Free. By using binary-only drivers, you're tying yourself to one or two kernel versions, and if 4front went out of business (probably not going to happen, but within the realm of possibilities) you're up the proverbial polluted tributary without a means of locomotion.
Personally, I hope alsa's header files stay GPL'd. The ALSA development team has worked hard to create an entirely Free set of sound card drivers, and I'm one who thinks that people should come to us on our terms, not any other way. Our way is the GPL, so do what Trident did and release GPL'd drivers, or do what .. creative used to do .. and release detailed specs on the card so we can release GPL'd drivers. Anything else is undesirable and difficult to work with.
What I really hate is the way there's bugs in the free drivers which don't exist in the commercial ones. For example, the free CS4236 driver has a problem where (even if the module is already loaded) the speaker pops loudly every time a sound begins playing. The commercial driver works perfectly.
I guess my point is: it's slighly annoying to have features withheld for commercial gain, but withholding fixes to sell more product seems sleezy.
--
The GPL on the alsa.h file is not a problem for applications or kernel sound driver modules. Linus specifically makes an exception to the GPL for drivers that are loaded as modules (which must call into the GPL-ed kernel APIs) and in the same document where he makes that exception, he states that applicaitons that run on top of the Linux kernel are not derived works of the Linux kernel and need not be GPL-ed.
Thanks
Bruce
Bruce Perens.
Is the ALSA API really GPL'd. While that seems very nice, surely it will inhibit acceptance of the API and the amount of support it will receive; they're not exactly in a position to dictate terms, are they? Or has Linus blessed them? It just seems like an X-style license would be more approriate. Any ALSA people reading want to explain the decision (I have no doubt they put some thought into this)? Also, any reason alsa-project.org is down?
I was kindof interviewing a member of the devlopment team informally on IRC out of curiosity, he mentioned quicktime (yes, even 4.0) and mpeg video, but not indeo and stuff... for mpeg they are using the mpegtv SDK, according to this guy.. I have no idea about the others, but of course they *have* to be binary-only... I think it would be cool to work with xanim and try to make the video codec archicecture binary compatible with xanim's... then things would take off much faster and have a better support base.. Of course the feasibility of this might be low... but then again, xanim is opensource so the loader code is right there for exanimination, and since this is a completely new feature, I don't see how it could be infeasible..
XML is like violence. If it doesn't solve the problem, use more.
Well, I can see it, as long as we forget that XMMS is for X :) I personally think development towards windows would be a waste of resources, they already have these tools... I think the community would best be served by focusing on linux and trying to make a good multimedia player.. with the audio flexibility of what was x11amp, the video flexibilty of xanim+mtv, and the interface features of mtv (full-screen playback is cool), then there would be far fewer complaints about multimedia lacking in linux..
XML is like violence. If it doesn't solve the problem, use more.
Despite all the supposed good intention of buying x11amp, I suspect a bit more sinister motives... They already shut down the cvs server to the public (public access only through tarballs). And the statement about the mpeg support already being done obviously indicates that development isn't released even to the cvs available to the public. I see this as closing development off bit by bit... My extremely paranoid theory is that they saw OSS being overtaken by ALSA in the future, and will be closing it up and making it a commercial product in the future.. Well, even if they did it, if they made xmms a good product with true multimedia support that would interface with ALSA, I would be tempeted to buy it, provided that they conduct business sanely. I see a lot of comments of comments talking about how evil OSS is because they don't make all their source code available and free of charge. I think it is stupid that people REQUIRE that a person give away this/her work in order to be accepted. GPL is nice, but after someone earns a reputation, they ought to be able to sell a product without being put down.
XML is like violence. If it doesn't solve the problem, use more.
4Front made the following claims wrt ALSA:
1. ALSA copied the OSS API and added a couple of things to it.
2. ALSA is for experimenters, not people who need commercial support
3. ALSA's use of the GPL will taint every product that uses it.
4. ALSA is spreading FUD about OSS, calling OSS "obsolete".
5. OSS hasn't challenged the claims in 4 publically, because polite people don't dis others hard work.
6. ALSA has managed to come up with a decent MIDI interface, which OSS will take and credit the ALSA team for.
Item 5 confuses me. I mean, except for the MIDI stuff, all the 4Front rep did was dis ALSA. He acted as if the unreleased ALSA as is now in development was a reasonable thing to compare to a released, commercial product -- ALSA doesn't have software mixing, but OSS does (yet software mixing is a major things ALSA will have that OSS/lite lacks); ALSA is for adventurous developers only, without any commercial support (true, but it's a system in development, not released yet), etc.
Who is spreading FUD?