ATI Releasing Specs for TV Tuner 52
Jonathan C. Masters
writes "As you may be aware, a long battle with ATI has
been going on to persaude them to release developer
information about their TV Tuner products to the Linux
community. Well, it's happened! Sorry folkes, there is no
Web page yet (I was notified by a mail list). Checkout
this page for general info (CT:Seems to be down)
and this one for more ."
Someday we'll live in a world where specs aren't propriatary.
Specifications (Score:1)
This raises an interesting issue. If a consortium of for-profits comes together to design a specification, then there is normally a fee charged to each member of the consortium. This covers staffing, secretarial (they all need to eat) and equipment costs.
Linux is now a "For Profit". Whilst the OS core and libraries are free, there are lots of companies getting in on the act of making money out of it. So there is an obvious problem. What is the difference between Red Hat and IBM, Sun, HP, SCO, Compaq etc ?
At what level of "wealth" does a OSS community really have to start pulling it's financial weight for development work?
Whilst some people are liable to work for free and provide input on a part time basis (a muddying of the pond -- life is never b&w), I expect that OSS (and Linux specifically) is going to have to start paying $$$ into the pot just like everyone else.
Which gets into another interesting point. Which OSS company (e.g. Caldera, RedHat,
I do ask myself sometimes if it is really fair for some companies to fork out $200,000 and Linux $0 (even when they are making lots of $$$ out of it).
Are we likely to see more stuff under patent?
Let's see... Who uses ATI chipsets? (Score:1)
On the flip side, nVidia is helping developers (after a tussle over XFree86), and nVidia is wooing manufacturers. For example, Dell now uses nVidia boards on their high-end, eating a bit of ATI's glory. Matrox has been quite developer-friendly lately, too. The high-end today is tomorrow's low-end, so they see troubles.
Seems like a critical mass is approaching. People are starting to feel competative pressure to open up. This is damn good, and how it should be (more or less). And a good bit of the pressure seems to have come from the community, not only the `leaders.' People have explained why they have chosen other solutions, providing a lot of the pressure.
Jason
WinTV (Score:1)
Let's use the Slashdot Effect for good... (Score:1)
ttyl
Farrell
Parallel scanners (Score:1)
Anyhow, though... I think I've seen some parallel->SCSI kernel-level emulation support. I'd expect it to be possible to use that with the standard Microtek SANE drivers (assuming Microtek, like most other parallel scanner manufacturers, just takes a SCSI scanner and throws on some SCSI->parallel conversion hw for their parallel line).
WinTV (Score:1)
IMHO bt848 cards have the best Linux support of any card right now. Well worth the extra PCI slot needed.
Unfortunately, the latest drivers won't compile under 2.0.x series kernels, and the author has removed old drivers from the webpage claiming that these drivers should compile.
Specifications (Score:1)
The problem is, even if some wealthy developer (or company) does shell out $200,000 for the spec, they won't be allowed to release the source code for the driver. That's the problem. How can it be a 'standard' if a part of the community is legally barred from ever supporting it?
For example, Red Hat would have to break it's practice of releasing it's software in GPL form and risk it's market in order to comply with those terms. As if it doesn't get enough flak as it is.
I argue that DVD et. al. is nothing but a widely licensed proprietary system. It is not a standard in any way.
A true DVD standard form the free software point of view would be a hardware api that is shared by all DVD hardware and is openly published.
So what about the other specs? (Score:1)
I assume that the remaining specs will be along shortly, together with the 3D acceleration API..? The ATI FAQ [atitech.ca] says not, but I'd definitely like to see it. I mean, someone offers to write them drivers and therefore sell lots more of their cards, and they say no...? Boneheads.
--
W.A.S.T.E.
But most *laptops* come with winModems. (Score:1)
Also, some laptops, *cough* *cough* *compaq* *cough*, have CDROM drives with nonstandard software interfaces that Linux will not fully recognize.
It WORKED!! (Score:1)
I'm not watching TV under my linux box...
Wow.. now if RealMagic Releases their DVD Decoder specs I would never have to switch back into windows... =)
You don't know how happy I am..
ChiefArcher
It allows competitors to make compatible hardware (Score:1)
--
Timur Tabi
Remove "nospam_" from email address
Commence the celebration (Score:1)
I almost bought one of their TV cards a few months
ago, until I learned they were withholding specs.
I might actually reconsider my decision in light of
their reversal...
If you're whinin', you're not fixin' (Score:1)
Problem is, dearies, that
But really, as any OSS programmer should say, "Want it? Write it!" Most folks who ask aren't willing to code, hence the attitude from the folks who DO code.
In other words, if you're whining, you're obviously not part of the solution.
Specifications (Score:1)
If noone from the Linux community joins, then Linux isn't represented in the design of a new spec.
It's not about fairness. But whether the Linux community wants to have a voice or not.
Specifications (Score:1)
If specs for DVD & TV tuners were all patented, then I imagine we'd see drivers almost overnight. Instead, they are treated as Trade Secrets, kind of like the formula for Coca Cola. These restrict sharing much more than patents do.
ATI (Score:1)
I plan to learn to code... (Score:1)
I agree, I hate that! (Score:1)
ATI has seen the light! (Score:1)
ATI TV Spec for Linux (Score:1)
About bleed'n time! (Score:1)
and to think they got rude with me when I asked them about TV Support in Linux.
Too little, too late (Score:1)
It allows competitors to make compatible hardware (Score:1)
If you're whinin', you're not fixin' (Score:1)
> "Want it? Write it!" Most folks who ask aren't
> willing to code, hence the attitude from the
> folks who DO code.
I won't challenge the arguments for not coming up with drivers for WinModems/ParPort Scanners because they seem good, but this "Programmers only need apply" chauvanistic attitude would be funny if it's naive mindset wasn't so nauseating.
I think that most people would be willing to code, but a lot of people can't and the learning curve is precariously high. It's not a simple manner of just learning basic C syntax, there's learning some high-level C programming techniques for things like memory management, many poorly documented libraries and APIs, the OS interface, and finally in this example, unusual or poorly documented hardware interfaces.
Given a person with a full-time job, the most basic of personal responsibilities and a rudimentary social life, they might have 8 hours a week to really focus on the above tasks. Just becoming a halfway decent C programmer might take them a year, and the rest of it could take another year. This doesn't include the catch-up time required to keep up with the shifting ground underneath their feet.
You could probably compress that timeline to about six months if you're a complete loser with no job, family, sex life, dog, friends or anything else to distract you. If you are, you're probably also a sociopath, and I don't want you living in my neighborhood even if you CAN write WinModem drivers.
So I ask you, is it really fair to claim that someone who wants something isn't willing to do it? I complain about my car not doing what I want, but I don't hear Toyota griping that people who aren't willing to make theit own motors aren't part of the solution either.
Like it or not, Linux/OpenSource/what-have-you can't expect that the cost of participation to necessarily be the skills needed to create it in the first place. If it is, Linux proponents are seriously overestimating its appeal.
Developer's Paranoia (Score:1)
> of their stuff to get released and getting burned by it?
Isn't this something along the lines of what happened when the IBM PC got cloned? IBM didn't want to waste R&D dollars on the PC, so they used non-proprietary off the shelf components. This left just the job of reverse engineering the BIOS in order to create a clone.
I'm not sure what would've happened if IBM had used proprietary stuff in the PC, but it would have probably not been cloned as easily. I'm not sure what the computing landscape would look like if PCs hadn't descended into commodity pricing.
DEC, HP, IBM and Apple computers all running mutually incompatible operating systems? Complete systems with only 8MB RAM, limited graphics and 100 MB hard disks still going for $3k due to expensive designs and lack of economies of scale?
WinModems Suck! (Score:1)
It WORKED!! (Score:1)
Interesting spec sheets.. (Score:1)
Does this mean that it could very well be possible for my ATi TV Tuner to function as a regular FM Stereo, in conjunction with my sound card? Or did ATi decide not to install the requisite crystals that would be needed to support this feature..?
Thanks..
Now let's get the winModem specs. (Score:1)
What a relief! (Score:1)
Developer's Paranoia (Score:1)
It's not as if IBM actually got burned by using off the shelf components...
They are an industry supplier of PC components, if I am not mistaken, including LCD screens, monitors, hard drives, etc.
They also happen to have a very nice line of notebooks, and a very very nice PowerPC line as well. So both parties, the consumer and IBM have gained from this, as not only has their pie grown, their slice of the pie is actually fairly significant, considering how big they are.
AS
Too little, too late - not really (Score:1)
And, for an initial release, it seems to run quite well...
Now, we just need to get Creative Labs to write software for their DVD kits or open up the specs...
PHB thinking (Score:1)
Nah, it's just that THEY SUCK!
YESSSS! (Score:1)
Life is good.
Next Big Thing to wish for: support for my Sony DVD CDROM under Linux. (hey, ATI came around!)
:)
Internal release? (Score:1)
This is an internal release to ati-team. It is not secret or closed-source, you can give it to whoever you want to, just don't publicly announce it just yet until we can nail down any remaining issues.
It allows competitors to make compatible hardware (Score:1)
finally I can use it! (Score:1)
Specifications and influence (Score:1)
XATITV-GATOS (Score:1)
The URL is...
http://www.core.binghamton.edu/~ins omnia/gatos/ [binghamton.edu]
YES!! (Score:1)
About time. (Score:2)
GATOS (Score:2)
And for the record: ATI has been very cooperative, they believed that releasing their specs would give competitors an edge (which is understandable, wrong or not - I'm not going to debate that issue here).
As soon as we (Mostly Øyvind Aabling [slashdot.org]) managed to reverse engineer most of the workings of the cards, they realized it was no longer a secret and have been very helpful so far.
Developer's Paranoia (Score:2)
Why is it no one ever thinks that by being liberal with their information they will in return get more hardware level support? Is there any history somewhere of company XYZ allowing info of their stuff to get released and getting burned by it? Yeah, it is sorta relying on social reciprocity, which is a social and not a legal contract...
It's not like being the Most Widely supported, understood, acknowledged, and programmed for is a bad thing, though with mass public release I could see a company losing some of its freedom, being tied to its customers when releasing new/future products. But isn't that actually called customer loyalty and having an installed user base?
One thing I can think of, if 3dfx released info on their Glide API, then other companies could ostensbily offer support for it, either as drivers or in hardware, and then 3dfx would seriously have to compete in an additional arena, since Glide capability would no longer be a sign of uniqueness. Heck, free coders could release better wrappers for Glide for alternative devices, like ATI's cards, which have *huge* installed bases, and hurt 3dfx immensely.
Or is that analogy and concept different than a hardware company releasing spec and info?
AS
About time, indeed. (Score:2)
Time was, in the days before hardware became complicated, when it was dead easy to write to, say, Motorola and get them to send you the specs for the 68B54 chip which was the key device in every Acorn Econet system. As a result, there were lots of different fileservers available for the Econet, and it was possible to plug in most computers you'd care to name (down to and including the ZX Spectrum, so I'm told!) to an Econet system. Did this harm Motorola in any way? Of course not - they sold many more 68B54 chips than they would have done otherwise.
If this recent announcement amounts to anything, this is a welcome change of tune - it's about time the hardware manufacturers took this line again. Producers of soundcards, 3D cards, tuner cards and PCI modems - listen up!
"If it's a bad idea, trash it. If it's a good idea, steal it and release the source code."
Why they withhold specs (Score:3)
Actually, you can get a fair bit of insight into a graphics chip's inner workings by looking at the register specifications, and if I understand correctly this is indeed what they're worried about. I speak as a graphics driver writer
The main disadvantage to releasing your specs that I can see is that your competition gets to see exactly how ugly some of your chip architecture is. I can't mention names due to NDA, but at least one of the chips we have to program for makes me want to crawl under my desk and whimper - and I'm an excellent programmer with a very good grasp of graphics hardware.
Fortunately, their competition's chips are much nicer.