Trident Micro Changes Policy Toward XFree86

Alex writes: "According to Egbert on the Xpert Xfree86 mailing list, Trident Microsystems, who makes video chipsets for low end PC's and notebooks, has changed its policy towards open source developers. Get the details here." If you want to email Trident Micro Public Relations, please be polite! Flaming will only hurt the chances that Trident will reverse this decision.

Comments

No need to write to Trident

[fobbman]

As soon as vendors announce that they will be CyberBladeXP or later Trident chipsets simply send an email to the vendors sales department notifying them that you will not be buying their laptop because the video subsystem does not work with your chosen security-based operating system.

Trident won't respond to a few users, but they will respond to vendors who are fielding complaints.

CyberBlade XP, eh?

[EnderWiggnz]

hrm... "XP"

i wonder if MS is in the backrooms twisting some arms...

Why bother complaining?

[Rimbo]

Trident hasn't been a player in graphics technology for years.

Probably as a result of poor business decisions like this, too.

If they make something worth buying, I'll worry about it. Until then...

Why do companies do this type of thing?

[schatt]

This is a serious question:
Why do companies do this sort of thing with their products? It would seem to me that having the interface to a particular chip would not be particularly helpful to designing a competitor, ("Well, if I tell it to draw a blue square, it draws a blue square! I know how to copy that!") so what good does this do?
I've always been under the (possibly mistaken) impression that it made more sense to distribute specifications to everyone, so that others could use your hardware. If you have to write the drivers yourself for every operating system that you are going to allow to use your hardware then that would add quickly up to a rather large expense, wouldn't it?
Are drivers really that much of a proprietary, critical secret for hardware companies? Does having the source code for your drivers help anyone else create drivers for their products? What benefit is there in preventing others from having the drivers?
Sorry if these questions seem silly or unimportant, but I've never understood the other side of the secrecy of our drivers argument.

Re:Why do companies do this type of thing?

[mandolin]

If you expose an API, and particularly, unrestricted driver source for a chip, it makes it considerably easier for other companies to reverse-engineer and "clone" a knockoff.

Example: look at all the tulip (network card) chipset clones out there. Last I checked they don't win on being better than the tulip, most of them are actually crappier; they win on being cheaper parts that (basically) work with already-written software.

And apparently it's not that hard. Quoting a coward from an early soundblaster article (only the most reliable sources here!), "weitek reverse engineered one of Sun's graphics chipsets because they got hold of a single .h with the register specs".

However it seems to me this strategy is only of any use when you're a big player with a popular chipset, trying to keep the little players down. My perception is that Trident is neither, so why they are doing this is beyond me.

Re:Code monkeys

[LordNimon]

It's not worth the effort. Video hardware is immensely complicated. A Linux video driver programmer could add support for 10 other chips in the time it would to reverse-engineer the Windows drivers.

Re:Code monkeys

[ZxCv]

It's not so much the chip won't be supported at all-- the standard Trident driver should still work. What they're referring to more is the proprietary acceleration features built into the chip. Those types of added features and benefits (which are actually probably required for any decent output from the card) are what will be missing.

Trident is still making chipsets?

[A Commentor]

Man, the last trident video card I saw was PCI... With all the major brands going under, I wouldn't have expected Trident to last.

What do they expect to get from keeping their specs private?? It's not like they are the leading chipset maker, and other companies are attempting to steal some secrets.

at this point can we help or hurt?

[Frizzled]

(from the last post off xfree86.org on this issue):

Significant amounts of well-reasoned arguments in emails from end users might possibly have an impact. Of course, as soon as this hits /., they will get 10 times as much flamage as reasoned argument, and be even more convinced that Open Source is not where they care to invest their efforts.

sad, but true. there are a lot of good arguments to be made for keeping this information open to the public. but when an issue like this gets pushed into everyone's view it tends to generate comments that might push the company further away from open source instead of closer to it.

granted, the "cat's out of the bag" at this point, let's just hope trident sees the light and reverses their decision (before 3d acceleration took off, all i used was a trident, 2MBs of video RAM ... woot!)

_f

Slashdot and flames

[VP]

Looks like some of the replies on the mailing list are also worried about the response Trident may get from Slashdot....

Re:Boo Hoo

[jchristopher]

ATI's laptop support is HORRIBLE. The chips themselves may be fine, but their driver support is abominable. What's a good card without drivers?

They have basically stopped driver updates on the Mobility series, even though that chipset is used in many CURRENTLY shipping products.

They won't fix dual display under Windows 2000/XP, even though every other manufacturer has figured it out.

If you are in the market for a laptop, I would highly recommend getting something with the new Nvidia chipset.

Please enlighten me

[novastyli]

I have been wondering the same. I really would like to hear an opinion from someone who has been in iteraction with these hardware vendors.

My hypotheses are:

• the internal communication is not going well in such a company and the people who decide these things are completely clueless.
  
• people making drivers in such a company are trying to make their work look more important.
  
  
  

putting on the pressure

[darkonc]

If you follow the thread, they mention that the most effective avenue might be to go after OEMs and discourage them from using Trident chipsets. An effective way of doing that may be to go one level further back for leverage.

Push the purchasers for your company and/or school to notify suppliers that you won't be accepting Trident chipsets because of this decision. Inform them that you need to be able to use your machines interchangably, and if Trident chipsets are not being supported by Linux, you won't be able to use them in your Linux boxes...

It's easier if you know that, in a crunch, you're not going to have interchangability problems with a machine because of Trident's unwillingness to support Linux. This leverages a possible 5-20% linux market share into a 100% purchase decision, on firm financial/operations grounds.

Something to note is that, even where Linux may not be a high percentage of a groups machine count, Linux boxes are often in a high profile or critical area. Being unable to deploy a machine into such a location could be a real impact to the company. If nothing else, it's just an unwelcome annoyance.

An OEM faced with a choice between losing a medium-large customer or switching to a 'widely supported chipset', is more likely to walk away from Trident. that sort of pressure is something that is likely to be 'heard' by the company.

confusion over trident

[johnjones]

there has always been a bit of confusion over trident drivers

e.g. the CYBER9385 this had at one stage 3 drivers distributed in a major release this is because they named chips the same

Trident supplies low cost chips because they are small (as in die size) and thus makes them less power hungry which is essential in laptops

the problem is that lately they have done into the onboard chipset market with Mother Board manufactures garbing them as a cheap way to stick video on board

then trident accelerated parts of their chips for these vendors

they have always been tight but allowed NDA people to help out writing drivers

the people you should complain to are the MB manufacturers who properly paid for the work to be done

so this begs the question who uses trident that you know ?

me I know SIS do so

write to
China
Ms. Ellie Yin
Tel:886+2+29161619 ext.346
E-mail: ellie@sis.com.tw.

Europe, Taiwan, Japan, Korea:
Ms. Jessie Lee
Tel:886+2+29161619 ext.341
E-mail: jessie@sis.com.tw

America(Canada,U.S., and Latin America), Oceania:
Miss Michele Huang
Tel:886+2+29161619 ext.345
E-mail: michele@sis.com.tw.

for your appropriate dealer

regards

john jones

Here's one helpful arguement

[rkent]

there are a lot of good arguments to be made for keeping this information open to the public.

Very true. And since the contact address given was public_relations@tridentmicro.com, I chose a PR-related arguement: first of all, all open source users (growing in number!) will have to avoid this new chipset, since it won't be supported. But more importantly, we'll remember Trident's decision and be less likely to support them in the future. Here's the letter I sent to that address; feel free use it as an inspiration for a note (not flame) of your own:

Hi -

I recently learned that Trident has decided not to provide chipset
documentation for the CyberBladeXP chipset to open source developers.
The effect of this decision is that Trident customers who choose to use
open source operating systems such as Linux or BSD with their computers
will not enjoy the full functionality of their CyberBladeXP video
systems. In fact, the systems may not work at all.

Besides being rude and alienating to your own customers, this news of
these non-functioning systems will spread by word of mouth, and people
will avoid Trident chipsets intentionally. At first, it will only be
certain chipsets that they try to avoid. But, as I'm sure you know, once
a company's name has been associated with a poor product, it becomes
difficult to trust that company for other products, as well.

In short, I'm not sure if I would even have bought a CyberBladeXP chip
from any vendor. But now that I know it won't work on my system, I will
be sure to avoid it. And now that I know Trident is upsupportive of my
software, I will probably have to avoid your products altogether in the
future.

Please reconsider your decision about the chipset documentation.

Sincerely,
[my name was here, put in yours]

Cut and paste if you care

[cnkeller]

As others have stated, I don't use Trident chipsets, but anytime that people take this attitude it bothers me, so here you go to those lazy people.

public_relations@tridentmicro.com

To whom it may concern,

I'd just like to express my disappointment at your recent decision of no documentation to open source projects.

As a proponent of open source, specifically linux, I've always purchased hardware that was linux compatible. It is likely that forthcoming hardware from Trident Microsystems will not be be supported by the open source community, due to a lack of vendor documentation.

If this is the case, I, and many others I'm sure, will be forced to choose hardware that does not make use of the Trident chipsets.

I'd like to urge you to reconsider your decision.

Thank you for your time.

Good question.

[jcr]

I just sent them this:

Dear Sirs:

How can it possibly benefit your company to require an NDA for the documentation on how to use your graphics chipsets?

Several years ago, when I was consulting for Hewlett-Packard, I was informed that HP considered the data on instruction latencies in the PA/RISC architecture to be highly proprietary, and access to it was strictly controlled by their in-house compiler development group. Indeed, even other parts of H-P were unable to get it.

The upshot was that GCC was unable to include a decent optimizer for PA/RISC, and consequently designers chose PPC, SPARC, MIPS, and the products of other, more cooperative vendors for their designs. Heard anything about PA/RISC lately?

Frankly, I'm as astounded by your stupidity as I was at theirs.

-jcr

"These kids today don't know the simple joy of saving four bytes of page-0 memory on a 6502" - unknown

Its just Open Source vs. IP confidentiality

[kbonin]

Competitive advantages w/ hardware often turn on clever use of data flow within chip designs, and it is often possible to obtain patents on such designs. At the same time, aspects of these designs are often exposed in driver API's.

It is a common beleif (I've had lawyers give conflicting advice in this area) that protecting API's under NDA's helps defend against a competitor figuring out what you're doing from "public domain" information and thereby having a legal basis to circumvent a patent.

The technical and legal merit of this position are certainly arguable.

You could also add that "opening" an API requires spending some effort (and $) spent on creating publicly readable documentation (although I've had to work with documentation from many non-English companies that apparently hired elementary school students to translate...)

Closed source drivers for such cards seem like a great solution for this problem, but many people in the open source community have 'religious' problems with this.

Re:Its just Open Source vs. IP confidentiality

[nyet]

When you patent something, you make its design public knowledge. No one can use that knowledge without your permission. Thus refusing to document an API because it's patented is inherently contradictory.

Very true. But in this PARTICULAR situation, Trident is no doubt in the process of obtaining patent(s). Which means that from their perspective they need to play it safe until the patent has been awarded. I'm not a fan of patents, but this is the way the game is (and always has been) played.

In many cases, API hoarding is done by a CTO or a product manager or two who thinks their technology is so wonderful and original that 1) nobody has done it before 2) nobody is smart enough to do it on their own and 3) exposing the API will allow somebody to "steal" their brilliant idea.

Very rarely does this type of CTO/manager have any academic/scientific background; normally they are MBA types who think every passing clever idea they have is a potential make-money-fast scheme.

Most REAL engineers/scientists realize that most everything has already been done, and most "innovations" are built upon millions of other (much older) ideas.

Unfortunately, most of the population does not belong in this category, and thinks that Salad Shooters(TM) need patent protection.

No real conspiracy, just business

[alsta]

Part of this information is retrieved from a person I know and part of it is fiction. The relevance to the XFree thing is the last two paragraphs.

So here is the thing.

Trident has filed a patent for a technology that they want to use in their new line of chips. We can all agree on how evil patents are, but they are allowed to file patents. In order to use a patent pending status, you can't disclose source code or specs. That being freely available invites others to implement the technology. So to be awarded a patent one must take reasonable steps to prevent others from copying the design. If that isn't being done, the patent can be considered void in a patent lawsuit. After that, the technology is considered Public Domain.

If Trident wishes that this patent goes through, they need to take this action. Otherwise somebody can simply retrieve a spec from them and/or work out the source code and reimplement in another chip.

I could care less if Microsoft subsidised this or not. Trident is allowed to publish specs if they want to and withold them if they so desire. This is their right.

All I know is that I will not be purchasing any Trident products in the future. This is my right. Until the day that legislators tell me that I MUST buy a Trident product I am not going to complain about this. There are plenty of other chip makers out there and they make good stuff. In my opinion, Trident chips have always been flaky and low budget.

Alex

Re:Boo Hoo

[Nailer]

Really? At Linux.conf.au I asked a member of the XFree86 team who'd just finished a taslk what chipset he'd recommend for mobile users. His response was the ATI Rage Mobility (16Mb model). Open Source drivers, the all important XVideo support, and good performance (better than the same card under Windows) was the absis for his decision.

Re:Who cares?

[Raven667]

While I wouldn't buy Trident for performance but they are traditionally very reliable. The people I know have built OEM whitebox computers in the past know which vendors are reliable and which aren't, they prefer Trident based boards because they are cheap and people don't have to return them. Other, whizzier, video chipsets tend to have more wierd issues with particular software or they overheat and die, but the Trident based boards didn't have these problems.