> >> I'd love to see more vendors providing specs, and doing so more
> >> openly, and preferably without NDAs. Ragging on vendors who do
> >> permit access to docs under NDA to people of their choosing, for
> >> not providing them to the world, is more likely to dry up access
> >> to specs for _EVERYONE_, and make binary only drivers the only
> >> way of getting modern hardware to work.
> >> In other words, I believe that whining about these certain
> >> realities, is equivalent to shooting one's self in the foot.
> >Mike you're quite the downer at the moment, been a rough weekend? <g>
> Neither really. ;o) I'm just expressing my opinion on how
> things are, and what we can realistically expect now, and in the
> near future, at least from my perspective. I might not be 100%
> correct, but it's how I see things from my current viewpoint
> >I couldn't care two hoots about whether or not ATI sits on the hardware
> >documentation or starts distributing it to univertsities as teching
> >What I'm saying is that if they'd decide "gee this document can
> >be released without problem, along with this pile over here and
> >this lot over here can probably be released for use only for
> >writing OSS drivers" then they should go ahead and do it.
> Absolutely. I think they'd (any vendor, not just ATI) do that if
> they really wanted to do that. I think the fact that some
> vendors do not do that is indicative that they don't want to do
> that however, or they would. ;o)
I am also implying that they should push those docs into the OSS community
rather than make the OSS pull them from ATI.
> >It would make life a lot simpler for all concerned. Why should
> >people have to fight to get documentation when ATI is in reality
> >quite happy to give out certain docs, but because they have
> >ceated an awkward process.
> I don't see it as a fight at all. Aside from the very few
> vendors who have publically released documentation (such as 3Dfx
> Voodoo 3, some older Intel docs, etc.), ATI is one of the vendors
> who provides docs to more people under NDA than any of the other
> vendors, with the exception of the cards mentioned above and some
> other older things here and there. In other words, if the
> alternative to a vendors docs under NDA, is no docs at all from
> the vendor, I don't think we should complain.
Hey ATI is this nicest vendor by far wrt releasing docs / supporting OSS
drivers, which is one of the reasons why you should continue to try and
get a bit more out of them, I'm not talking to the extent that would
damage the comapany here, the end result is OSS drivers for their cards
which is hardly a bad thing.
> >At the end of the day an NDA isn't much protection, eventually
> >the doc will fall into the hands of someone they don't want it
> >to, whether someone has to steal it off someone who has signed a
> >NDA, find it in the trash, bribe the night staff.
> Well, if people do not honour the NDAs that vendors give, it is a
> no brainer what will happen. If someone leaks documentation and
> breaks the NDA and it gets back to the vendor, the vendor is most
> likely just going to do one of:
> 1) Not provide documentation to people anymore period
As I said, all or nothing.
> 2) Make the NDA more restrictive and provide documentation to
> less people
People who violate NDA's don't really care how strict it is.
Sort of like:
If gun ownership is illegal then only crimanls will have guns.
> 3) Provide watermarked docs under NDA to individuals. If docs
> leak, they can then sue the person who leaked them, as it is
> obvious due to the watermarking
Watermarking can be defeated. You can literally retype the entire document
in another language with spelling errors. Well at least that's teh
first idea that pops into my head<g>
> 4) Force developers to work right in the vendor's headquarters in
> a monitored room with access to docs that don't leave the
> premises (such as some of the Serverworks IDE work, etc.)
Is that indentured labour. <g>
> >It pretty much is an all or nothing approach.
> >If they're prepared to release docs to members of TG, why don't they
> >release them to TG directly?
> I really don't understand your point here. You are suggesting
> that ATI release docs to TG, and then let TG decide who gets them
> and who does not get them. ATI is capable of deciding who they
> want having their docs, and if they wanted TG to be the people to
> decide that, they would ask TG to do that. The fact that that
> has not taken place means that they are capable of deciding this
> on their own, and that that is not an option that they consider
Assuming they've thought of it, and that they think TG is willing to do
it, then they'd ask them. Heck if I was at TG I would probably have put it
to ATI to relase docs to TG under whatever agreement they want, why should
the docs not all be on hand, it just strikes me as useful.
> I don't see the point of it anyway.
> >What I was doing was putting forward a suggestion that TG may be
> >able to get docs out of ATI easier (without screwing over ATI in
> >the preocess).
> And my suggestion, is that if ATI wanted docs to go into the
> hands of random open source developers through TG, that they
> would themselves just give docs to those random open source
> developers, which is the way it is now. Developers get the docs
> from ATI, or they do not get the docs from ATI. I completely
> fail to see how/why/what TG has anything to do with this
It's about making a library / repositary of docs that is available in one
place, there are similarities between cards so even if you're working on
the r300 but can't get docs for it referring to r200 will probably help a
little bit, is it worth having to apply for, probably not.
I'm just asking why there is no repositary for all public docs, with a
restricted acccess to NDA docs, if that repositary won;t happen at ATI
then it can happen at XF86 or TG.
When I join a library I expect to be able to check out all books (bar the
reference books), similarly if I sign an ATI NDA I should be able to
access all the documents under NDA, and if I don't I should still be able
to access all the public info docs.
> >It was just a suggestion, maybe after I learn C I'll care, and
> >argue my points with a lot more conviction.
> If you do not know C, the documentation would be useless to you
> in any case. It always seems to be the people who don't even
> know how to write helloworld.c that are the ones who complain
> about a vendor like ATI not providing them with hardware
> documentation that they couldn't do anything but make paper
> airplanes out of anyway.
lol, that's why I'm suggesting it on the list for comment, and not
bothering ATI politely.
I'm sure there are plenty of people who have enough paper to make
airplanes, and who actually want the docs for the info contained therein.
Basically ATI makes HW, they write the docs for it, the OSS comunity needs
/ wants those docs, if there i ssomething in those docs that ATI doesnt
want public then they should "sanitize" the document and just leave a note
in there saying "access to this info under NDA only" or "snowball's chance
Some is better than nothing. Some freely available docs may well be enough
to allow someone to fix whatever is bothering them.
btw one of the mail servers is down on my side of the pond so I'm not
getting the list email today...