Messages in this thread | | | From | "Khimenko Victor" <> | Date | Sat, 2 Jan 1999 18:44:29 +0300 (MSK) | Subject | Re: [linux-audio-dev] ISA Plug & Play support in kernel |
| |
In <199901020813.DAA24629@saturn.cs.uml.edu> Albert D. Cahalan (acahalan@cs.uml.edu) wrote:
> Richard Stallman writes:
>> 1. For Ghostscript, everything added to the non-free version >> will appear in the free version a year later.
> This can not happen. For many cards, NDAs would make a source > code release illegal. The sound system must be closed-source and > must be commercial to financially support the sole developer.
NO ! The sound system MUST BE open-source and MUST NOT include binary-only drivers without sources. Commercial or not commercial -- not such big deal. Ghostscript way (free version is 1 year behind) is acceptable even if not perfect.
> It is easy for you to say we shouldn't support such hardware.
This is what we SHOULD DO ! REALLY ! NOT support such hardware... If we could not develop driver without signing NDA, that is. Doing otherwise will kill Linux eventually.
> Owners of such hardware are likely to go back to Windows though, > which is even less free than Linux + OSS. When they throw out > Linux, they generally throw out all free software and advise > friends and family to avoid it too. ("It can't do sound.")
This is possible harm but this is by far MUCH LESS harm then possible corruption of Linux. Open linux-kernel archives and take a look on discussion about UDI -- there are was a lot of explanation WHY binary-only drivers are BY FAR worse then absence of any drivers at all. You think about Linux from marketing viewpoint ("how to sell a lot of linux to normal humans"). You are from wrong world. Marketing problems are exactly type of probems which are not considered in Linux development at all. What we need is open-source drivers for sound cards (other drivers could be consideren as TEMPORARILY SOLUTION ONLY and SHOULD NOT be included in major distributions). There are two ways: oscilloscope and such to produce drivers without needed info and bombing of vendors with documentation requests. Situation with SCSI drivers shows that both ways have good chances for success. And this all works MUCH better when there are no drivers at all then when there are exists binary-only drivers (commercial or not -- no matter yet again) since more peopleas are involved when there are no drivers at all :-))
> ObKernel: Lack of PnP in the kernel is going to hurt! Normal humans > don't figure out the userspace tools and insert kernel modules > with the right parameters. Besides using an off-the-wall syntax, > the whole idea of a config file destroys PnP totally. We've invented > software jumpers. People hate jumpers with good reason, software > or not. Systems that can't determine their own configuration are > not somehow better; they are worse. I seem to recall we had code > for this too.
Linux (kernel) is NOT ABOUT "hormal humans". Linux (kernel) is all about joy and doing "right things". If someone could try to sell Linux to "normal humans" he is free to do this. Till he will insist on augmeting kernel with some unclear parts to make linux "more sellable to normal humans". At this point he shuld be rejected without talks.
What about P-n-P: this is DEFINITELY not kernel task. This is belond to userspace. If current version of tools is not clear -- change tools. Do not pullute kernel with such crap as P-n-P ...
- To unsubscribe from this list: send the line "unsubscribe linux-kernel" in the body of a message to majordomo@vger.rutgers.edu Please read the FAQ at http://www.tux.org/lkml/
| |