Holes they might be, but the performance will suffer if you did it much of any other way. There's a reason why they did it the way they did- fastest way to accomplish the task without packet inspection and state slowing down connectivity. Just skim the packet traffic, if you spot one of the forbidden items in content, you issue disconnects to both endpoints in the middle and kill the link. Now you've got to do a real firewall and proxy filter- which means they're going to spend a hell of a lot of cash doing that, which they really, really don't want to do and honestly can't afford cash-wise.
They largely don't exist any more because of the actions- their win is a pyhrric victory, as was Digital Research's.
Unless you can afford the legal battle, you're almost always going to lose against someone like Microsoft- they can abjectly outspend you in that arena without even hurting their bottom line. Most people and organizations can't do that or even come a third of the way there.
That's why I keep telling people that the "there's someone to sue" line is nothing but folly- even in the case of proprietary software, there's nobody really to sue. So, if this is the case, why are you spending all that money? A false sense of security? Someone to blame instead of yourself when you screw things up?
"I think the real reason they're all going proprietary (and not providing SDKs) is because the service providers don't want there to be an easy way for anybody but them to make applications for the phones."
I don't think that's the case, I KNOW that's the case. And, better yet, it's not just that they don't want people making apps, they don't want people copying the apps from one to the next- most of the Mobile Phone vendors gig you each and every time you get a new app for a new phone. That's $2-15 each and every time, or a recurring service fee of the same amount per month if you have the apps- with a standard framework that everyone has access to, it'd be more like Windows or Linux is now, instead of the way the phones are now and the way they want it. The big problem is that they can't seem to come to an agreement on how they want to screw all of us- some have done Brew, some have done a modified version of the official Java, etc. Therefore, we have all these different consortiums of Linux on the Smartphone companies.
Typically, I go and do the research and go looking for brick and mortars like Fry's to buy the stuff.
Yes. Fry's. And, if it's not Fry's, it's out of a local wholesaler- I've got a DBA and all anyhow.
Why? Because the places I buy from have a return policy that while it's obnoxious at times to use, it's still there.
With the online stuff, you may/may not get a return policy, and it's even MORE obnoxious to use it (when it's available) if something doesn't work QUITE as intended, they changed the damn chipsets on you and it doesn't work with Linux like the earlier models did (WiFi cards are a GOOD example of THAT one...), etc.
I don't do the asking of anyone except another geek like myself- they wouldn't know and largely don't give a damn anyhow. Now, I question his geekiness, for other reasons- I mean, Best Buy, really... They honestly haven't had a decent price in that place for computer gear in years and I've seen what they think of their customers.
In reality, we do... Otherwise, we wouldn't have all the nice vector fonts that Linux/*BSD users take for granted... Some of them are not quite as good as the commercial ones, some of them are as good. However, this wouldn't have helped the printing house of the original article as they've got to interact with each and every customer, etc. who happen to HAVE licenses, etc. but then they need to print for the customer in question, etc.
With ActiveX/IE, you only provide Windows support. (Now, hear me out here before you dismiss that statement as Windows-Only isn't being a problem...)
With that, your application will only be runnable there. With XUL, and some work, your app will run on the following:
MacOS X (x86 AND PPC) Linux
and of course...
Windows
There's a bunch of others, but they're subsets of the above in most cases. If you think Windows is the thing, you'd be rather mistaken since approximately 10-20% of the global end-user workstation OS install base is one of the other two. To be sure, you don't need to do XUL- you can do the application in Qt, GTK+, Fltk, and a few others and get the same results with less effort unless you need some HTML rendering support- then it requires more effort to integrate a browser in your code.
Only where it's applicable. If Microsoft has no Copyrights in the countries in question either because of an oversight or because they've been stripped of them, then the Berne Convention doesn't apply and the WTO won't directly intervene. If you didn't know, even in the US, they may strip a rights holder of their rights if they're guilty of using the same to violate the Anti-Trust Acts. It just doesn't happen all that often.
Those deals are being made by the legitimate seller to get you to buy the service- the razor versus the razor blades deal. In the case of the dude in the subway, there's NO good reason to be selling a phone to a random stranger for $50 under most circumstances- because typically, a brand new Sidekick is attached to 2 year service contracts for that special price, subject to penalties or the like if you close out your account prematurely- except under extenuating circumstances such as going somewhere long term that doesn't have their services offered. In his case, it's VERY LIKELY to be a stolen item. It falls back to the "known or should have known" liability issue- you're actionable if you don't use good sense in these matters and your argument wouldn't hold water in a court of law.
Actually, it's not a problem of the format, per se, but one of the tools producing the results in question and a mis-understanding of what is actually being done by the people using the tools. In the case of PDF, ODF, MS Office, etc. the type of redactions done happen to only work if your app 100% honors the concept of a redaction- but it doesn't mean you can't have redactions. Nor does it mean you can't have portable imaged documents in electronic format that can't be usefully redacted.
Take TIFF for example. If you redact a TIFF document, you can still have the redacted content present in the document in a locked by encryption bitmap annotation that can be dropped on top of the rendering if your app supports the annotation type. If it can't or you don't have the key to unlock the bitmap annotations, you get just the redacted document as a rendering- period. The same can be done with PDF and other document formats, but unless someone thinks about this and produces the desired formatting and engine to work with it, you're going to keep getting results like this.
Excuse me, any electronic format, unless it is a bitmap format, will have this problem unless all the viewers 100% honor the redaction as it's intended. In the case of a bitmap format, you can burn a black or white rectangle into the original image and then add an annotation a la TIFF's annotations that contains the original portion of the image that was redacted in an encrypted format so that it's difficult to expose the redaction- IF you need to have the redaction exposed. If not, you hand across the redacted image as-is without annotations.
This has NOTHING to do with PDF or ODF at all- trying to make this a connection to these is bogus to say the least. In this case, I believe that the people doing it used the MS Office redaction capabilities and then exported the redacted content to PDF, which the export carried the same sort of redactions across to the other format. What happened is because someone didn't understand the tools they were using, not because of PDF or ODF.
Oh, and buying a $400-500 retail phone from a "psst...wanna buy a cell phone" type vendor in the Subway constitutes purchasing stolen goods if it is in fact not the person's to sell- it's a known or should have known" issue which means "Mom" is guilty of a misdemeanor or worse (depending on how NY views that sort of thing) herself- and she should clam up and count herself lucky that the NYPD doesn't have an interest in pursuing her alleged trafficing in stolen goods (The purchase or sale is that- both sides are typically actionable on that little equation, but an honest mistake on the part of the purchaser is a defense to prosecution. She's got nothing of the sort to lean back on...).
Odds on, the "I bought it" is more of a pathetic ruse to cover up for the fact that they pocketed it themselves.
Too many of the interchanges, obviously not contested by them, and the conflicting stories from them lead one to believe that Luigi or mom pocketed the Sidekick, not thinking that they'd get caught up with.
In the case of contracting out the work to the OS developers, it would very likely be a situation of them fixing the things because they're personally accountable.
In the case of a company like Oracle, it would very likely be a situation of them trying to fix it, not because they're accountable but because it's a potential customer loss situation. If your problem is the only one and you're not looking to be costing them customers or money to NOT fix it- it plain flat won't get fixed. If you end up getting financially damaged by the whole affair, unless the company can be shown to be criminally negligent, you won't be able to sue for anything because it's been disclaimed that the software isn't really guaranteed to serve any purpose other than to part dollars from your wallet. (Just read their licensing on the stuff sometime...)
A business that uses your thinking to justify all of this is not looking at what they agree to when they buy their software and are foolishly thinking they're "okay" when they buy Oracle's SQL DBMS or Microsoft's Office suite or their OS products- that there'll be someone to fix it for them if it breaks and if not, they can sue.
Try suing them. It'll be funny. It'll be hillarious. In most cases, the legal disclaimers they pile on your usage limit what you can and can't do in regards to a system failing and doing financial damages or physical ones. Unless you can prove gross negligence on their part with a flaw (and even then...it's hard to make it stick in the first place. Otherwise we'd be seeing class-actions being fielded against Microsoft- and been seeing them some time now...) you're not going to get very far with a lawsuit.
The "we can sue someone if we buy it" line has worn quite thin. Businesses may THINK that they can sue someone over a failure in a system component, but in business, it's "you're on your own"- if you get fubared over something that someone else did and you didn't do your due dilligence it's your own damn fault things went south unless the other party can be proven to have been criminally negligent. It's always been that way.
You didn't NEED to get the IDL on all the platforms.
1) The IDL's only useful for making the interface skeletons. If you needed that, you had a pretty twisted ORB. 2) I think you're referring to the IOR which is a reference to the remote object so the ORB can find it. 3) I will believe the IOR "problem" being obnoxious- it was evil in CORBA v.2 and prior implementations, for version 3, however, you could use iioploc:// type URI's for the same thing, and if your ORB properly implemented things, if you are on the same network subnet you didn't even need that as there's a magic IOR to feed into the load up to bring in the local NamingService which simply makes everything totally plug and play. TAO offers BOTH.
In my case, I've experienced the opposite thing. No IDL issues. No need to plug in IOR hex number sets. Only the rare, occasional iioploc:// invocation to the NamingService because we weren't on the same subnet and had to go through a port-forward on a firewall. It's all in the ORB implementation- and most of the ORBs just stink on ice.
When I used TAO, it just all snapped together. Minimal noise. No goofy setup, etc. If it's on the same network segment, it just finds the NamingService object on the segment (If you've properly started it...) and just works. I know, I've designed massively distributed trasaction processing systems with it that had to handle situations like vending a toll gate with a vehicle going through it to speeds up to 55 mph.
Now, to be sure, this is more the ACE/TAO people's doing, but there's NOTHING in CORBA that requires all the noise that you typically see with other ORB implementations. And CORBA's neither slow nor cumbersome- it's all in the ORB implementations. Of which, many just plain flat suck.
No, this is the average working week for the average wage whore (otherwise known as a consultant or contractor over here...). Considering that I'm NOT salaried right at the moment and I'm raking it in per hour at a high rate of money per hour... More bang for their buck in a shorter time-if I can get it done in the extra 10 per week worked and not burn out, etc. it ends up actually being cheaper for them, they get to market faster, etc.
No, I know whom this is- and he's about as talented as I. He's just not as motivated as I- and it's very indicative of India in general; I know, I've had to work with TI India before in the past as a contractor for TI- and it was the same story.
Gonna tell you the same thing... This server's always been in Germany. The site's very likely to be bogus, but it's NOT for the reasons people keep proposing.
The odds it's somebody at the university that has access that decided to be funny are good. But because it's where it is has no bearing on it. Keep in mind, also, that what graphics are there on the main front page are professional quality work- this is not to say that the site's legit because of that, but someone put a lot of work into making a logo that doesn't exist just for a prank. Again, not to say this isn't a prank or hoax it's very likely to be one- but again, none of the things you or anyone else have come up with actually make it so.
Uh, it's always been in Germany for this DNS entry... That's the rub. It's likely to be bogus, yes- but for NONE of the reasons you folks keep proposing.
...SCO owns the domain (Do a 'whois' next time before proclaiming "fakeness"...) and the site IS on one of their servers. Not to mention that the site's main page happens to be referring to the same thing, coming Early Q1 of 2006.
Just like SCO... Promise everything, have nothing in hand.
If you've got one developer dedicated to the work, what do you have him working on?
1) Adding OpenGL 2.0 features (including things like GLSL)- which are sought after but not all 100% done... 2) Fixing an obscure performance problem with the middle of the line Laptop GPUs...
If I were the manager, it'd be #1 that'd get done first- period. How many people do they have there at ATI doing the Linux stuff? A couple at least- things are shaping up quite nicely as the drivers have stabilized and while they're not QUITE performant on Linux as they are under Windows. They still DO have issues lurking in the woodwork- things that DO take time to sort out. If you've got only a couple (i.e. 2-4 or so) what all do you do to ensure conformance on the BROAD line of GPUs that they have to support under Linux?
Basically, it's NOT as simple as you'd think- and it's not anywhere as easy as many make it out to be. It's not simple doing code that doesn't stall the pipelines on a GPU or something like an iWARP card (Current contract work for moi- trying to drive a certain vendor's iWARP card to full wireline speeds (10Gbps, here I come!)...)- not everyone can code for it. ATI and NVidia's right in that it's not a simple thing to code for a GPU's drivers. What they're not right in is that they're the sole source for this sort of capability in programming on their chipsets- after all, where did they get their developers in the first place?
More eyes always works better than less of them. More minds working on a problem will invariably produce results not always envisioned by the original designers. John Carmack, for example, came up with a clever (but rather evil) hack that allowed the RagePRO UtahGLX support not need DMA allocated memory, special allocations of RAM, etc.- it just simply started up, allocated it's driver buffers and then using a feature of the/proc filesystem, figured out the bus memory addresses for the memory just allocated so you could hand it to the chip. No kernel space driver needed- only root privs which the X server had at the time. Had it been closed technical info, it probably would have never happened as a supported chipset.
I steadfastly hold to the principle that an SDK interface is an API on the silicon itself and you shouldn't be hiding things or worrying about "tipping your hat" on things patentable. APIs are supposed be available in some fasion. I would have problems with requiring that someone register for the info- but I'd have less of a problem with that than the current state of affairs. Interested parties would be able to at least do the things I've implied in this conversation.
Slashdot Mirror
Holes they might be, but the performance will suffer if you did it much of any other way. There's
a reason why they did it the way they did- fastest way to accomplish the task without packet inspection
and state slowing down connectivity. Just skim the packet traffic, if you spot one of the forbidden
items in content, you issue disconnects to both endpoints in the middle and kill the link. Now you've
got to do a real firewall and proxy filter- which means they're going to spend a hell of a lot of cash
doing that, which they really, really don't want to do and honestly can't afford cash-wise.
They largely don't exist any more because of the actions- their win is a pyhrric victory,
as was Digital Research's.
Unless you can afford the legal battle, you're almost always going to lose against someone
like Microsoft- they can abjectly outspend you in that arena without even hurting their
bottom line. Most people and organizations can't do that or even come a third of the
way there.
That's why I keep telling people that the "there's someone to sue" line is nothing but
folly- even in the case of proprietary software, there's nobody really to sue. So, if
this is the case, why are you spending all that money? A false sense of security?
Someone to blame instead of yourself when you screw things up?
I don't think that's the case, I KNOW that's the case. And, better yet, it's not just that they don't want people making apps, they don't want people copying the apps from one to the next- most of the Mobile Phone vendors gig you each and every time you get a new app for a new phone. That's $2-15 each and every time, or a recurring service fee of the same amount per month if you have the apps- with a standard framework that everyone has access to, it'd be more like Windows or Linux is now, instead of the way the phones are now and the way they want it. The big problem is that they can't seem to come to an agreement on how they want to screw all of us- some have done Brew, some have done a modified version of the official Java, etc. Therefore, we have all these different consortiums of Linux on the Smartphone companies.
Typically, I go and do the research and go looking for brick and mortars like Fry's to buy the stuff.
Yes. Fry's. And, if it's not Fry's, it's out of a local wholesaler- I've got a DBA and all anyhow.
Why? Because the places I buy from have a return policy that while it's obnoxious at times to use,
it's still there.
With the online stuff, you may/may not get a return policy, and it's even MORE obnoxious to use it
(when it's available) if something doesn't work QUITE as intended, they changed the damn chipsets
on you and it doesn't work with Linux like the earlier models did (WiFi cards are a GOOD example
of THAT one...), etc.
I don't do the asking of anyone except another geek like myself- they wouldn't know and largely
don't give a damn anyhow. Now, I question his geekiness, for other reasons- I mean, Best Buy,
really... They honestly haven't had a decent price in that place for computer gear in years
and I've seen what they think of their customers.
In reality, we do... Otherwise, we wouldn't have all the nice vector fonts that Linux/*BSD users take for granted... Some of them are not quite as good as the commercial ones, some of them are as good. However, this wouldn't have helped the printing house of the original article as they've got to interact with each and every customer, etc. who happen to HAVE licenses, etc. but then they need to print for the customer in question, etc.
Cross. Platform. Support.
With ActiveX/IE, you only provide Windows support. (Now, hear me out here before you dismiss that statement as Windows-Only isn't being a problem...)
With that, your application will only be runnable there. With XUL, and some work, your app will run on the following:
MacOS X (x86 AND PPC)
Linux
and of course...
Windows
There's a bunch of others, but they're subsets of the above in most cases. If you think Windows is the thing, you'd
be rather mistaken since approximately 10-20% of the global end-user workstation OS install base is one of the other
two. To be sure, you don't need to do XUL- you can do the application in Qt, GTK+, Fltk, and a few others and get
the same results with less effort unless you need some HTML rendering support- then it requires more effort to integrate
a browser in your code.
Only where it's applicable. If Microsoft has no Copyrights in the countries in question either because of an oversight or because they've been stripped of them, then the Berne Convention doesn't apply and the WTO won't directly intervene. If you didn't know, even in the US, they may strip a rights holder of their rights if they're guilty of using the same to violate the Anti-Trust Acts. It just doesn't happen all that often.
McNealy: Check...
Messman: Check...
Gates: Nope...He's still Chairman of the BoD. You don't get to count that one- at least not yet.
Jobs...well...Now that he's solidly enmeshed with Disney on their BoD, etc. I don't see him going that way yet or for a while to come.
It is that simple.
Those deals are being made by the legitimate seller to get you to buy the service- the razor versus the razor blades deal.
In the case of the dude in the subway, there's NO good reason to be selling a phone to a random stranger for $50 under
most circumstances- because typically, a brand new Sidekick is attached to 2 year service contracts for that special
price, subject to penalties or the like if you close out your account prematurely- except under extenuating circumstances
such as going somewhere long term that doesn't have their services offered. In his case, it's VERY LIKELY to be a stolen
item. It falls back to the "known or should have known" liability issue- you're actionable if you don't use good sense in
these matters and your argument wouldn't hold water in a court of law.
Actually, it's not a problem of the format, per se, but one of the tools producing the
results in question and a mis-understanding of what is actually being done by the
people using the tools. In the case of PDF, ODF, MS Office, etc. the type of redactions
done happen to only work if your app 100% honors the concept of a redaction- but it
doesn't mean you can't have redactions. Nor does it mean you can't have portable
imaged documents in electronic format that can't be usefully redacted.
Take TIFF for example. If you redact a TIFF document, you can still have the redacted
content present in the document in a locked by encryption bitmap annotation that can
be dropped on top of the rendering if your app supports the annotation type. If it
can't or you don't have the key to unlock the bitmap annotations, you get just the
redacted document as a rendering- period. The same can be done with PDF and other
document formats, but unless someone thinks about this and produces the desired formatting
and engine to work with it, you're going to keep getting results like this.
Excuse me, any electronic format, unless it is a bitmap format, will have this problem unless
all the viewers 100% honor the redaction as it's intended. In the case of a bitmap format,
you can burn a black or white rectangle into the original image and then add an annotation
a la TIFF's annotations that contains the original portion of the image that was redacted
in an encrypted format so that it's difficult to expose the redaction- IF you need to have
the redaction exposed. If not, you hand across the redacted image as-is without annotations.
This has NOTHING to do with PDF or ODF at all- trying to make this a connection to these
is bogus to say the least. In this case, I believe that the people doing it used the MS Office
redaction capabilities and then exported the redacted content to PDF, which the export
carried the same sort of redactions across to the other format. What happened is because
someone didn't understand the tools they were using, not because of PDF or ODF.
Oh, and buying a $400-500 retail phone from a "psst...wanna buy a cell phone" type vendor in the Subway
constitutes purchasing stolen goods if it is in fact not the person's to sell- it's a known or should
have known" issue which means "Mom" is guilty of a misdemeanor or worse (depending on how NY views that
sort of thing) herself- and she should clam up and count herself lucky that the NYPD doesn't have an
interest in pursuing her alleged trafficing in stolen goods (The purchase or sale is that- both sides
are typically actionable on that little equation, but an honest mistake on the part of the purchaser
is a defense to prosecution. She's got nothing of the sort to lean back on...).
Odds on, the "I bought it" is more of a pathetic ruse to cover up for the
fact that they pocketed it themselves.
Too many of the interchanges, obviously not contested by them, and the
conflicting stories from them lead one to believe that Luigi or mom
pocketed the Sidekick, not thinking that they'd get caught up with.
Here's the reality of it.
In the case of contracting out the work to the OS developers, it would very likely
be a situation of them fixing the things because they're personally accountable.
In the case of a company like Oracle, it would very likely be a situation of them
trying to fix it, not because they're accountable but because it's a
potential customer loss situation. If your problem is the only one and you're
not looking to be costing them customers or money to NOT fix it- it plain flat
won't get fixed. If you end up getting financially damaged by the whole affair,
unless the company can be shown to be criminally negligent, you won't
be able to sue for anything because it's been disclaimed that the software isn't
really guaranteed to serve any purpose other than to part dollars from your
wallet. (Just read their licensing on the stuff sometime...)
A business that uses your thinking to justify all of this is not looking at
what they agree to when they buy their software and are foolishly thinking
they're "okay" when they buy Oracle's SQL DBMS or Microsoft's Office suite
or their OS products- that there'll be someone to fix it for them if it
breaks and if not, they can sue.
Try suing them. It'll be funny. It'll be hillarious. In most cases, the legal disclaimers
they pile on your usage limit what you can and can't do in regards to a system failing and
doing financial damages or physical ones. Unless you can prove gross negligence on their
part with a flaw (and even then...it's hard to make it stick in the first place. Otherwise
we'd be seeing class-actions being fielded against Microsoft- and been seeing them some time
now...) you're not going to get very far with a lawsuit.
The "we can sue someone if we buy it" line has worn quite thin. Businesses may THINK that
they can sue someone over a failure in a system component, but in business, it's "you're on
your own"- if you get fubared over something that someone else did and you didn't do your
due dilligence it's your own damn fault things went south unless the other party can be proven
to have been criminally negligent. It's always been that way.
I'd be careful slinging that sort of comment about.
GNOME's no more bloated or slow than KDE and I've observed conditions where GNOME was more responsive than KDE on the low end.
GNOME uses CORBA. KDE uses DCOP. Both inject quite a bit of size and overhead and doing fanboy stuff doesn't make you look good.
Oh, I forgot... This is Slashdot... Forgive me, I'll move on- it's all about opinions here, not facts.
You didn't NEED to get the IDL on all the platforms.
1) The IDL's only useful for making the interface skeletons. If you needed that, you had a pretty twisted ORB.
2) I think you're referring to the IOR which is a reference to the remote object so the ORB can find it.
3) I will believe the IOR "problem" being obnoxious- it was evil in CORBA v.2 and prior implementations, for version 3, however, you could use iioploc:// type URI's for the same thing, and if your ORB properly implemented things, if you are on the same network subnet you didn't even need that as there's a magic IOR to feed into the load up to bring in the local NamingService which simply makes everything totally plug and play. TAO offers BOTH.
In my case, I've experienced the opposite thing. No IDL issues. No need to plug in IOR hex number sets. Only the rare, occasional iioploc:// invocation to the NamingService because we weren't on the same subnet and had to go through a port-forward on a firewall. It's all in the ORB implementation- and most of the ORBs just stink on ice.
When I used TAO, it just all snapped together. Minimal noise. No goofy setup, etc.
If it's on the same network segment, it just finds the NamingService object on the
segment (If you've properly started it...) and just works. I know, I've designed
massively distributed trasaction processing systems with it that had to handle situations
like vending a toll gate with a vehicle going through it to speeds up to 55 mph.
Now, to be sure, this is more the ACE/TAO people's doing, but there's NOTHING in CORBA
that requires all the noise that you typically see with other ORB implementations. And
CORBA's neither slow nor cumbersome- it's all in the ORB implementations. Of which, many
just plain flat suck.
No, this is the average working week for the average wage whore (otherwise known as a
consultant or contractor over here...). Considering that I'm NOT salaried right at the
moment and I'm raking it in per hour at a high rate of money per hour... More bang for
their buck in a shorter time-if I can get it done in the extra 10 per week worked and not burn
out, etc. it ends up actually being cheaper for them, they get to market faster, etc.
No, I know whom this is- and he's about as talented as I. He's just not as motivated
as I- and it's very indicative of India in general; I know, I've had to work with
TI India before in the past as a contractor for TI- and it was the same story.
Gonna tell you the same thing... This server's always been in Germany. The site's
very likely to be bogus, but it's NOT for the reasons people keep proposing.
The odds it's somebody at the university that has access that decided to be funny are good.
But because it's where it is has no bearing on it. Keep in mind, also, that what graphics
are there on the main front page are professional quality work- this is not to say that the
site's legit because of that, but someone put a lot of work into making a logo that doesn't
exist just for a prank. Again, not to say this isn't a prank or hoax it's very likely to be
one- but again, none of the things you or anyone else have come up with actually make it so.
Uh, it's always been in Germany for this DNS entry... That's the rub. It's likely to be
bogus, yes- but for NONE of the reasons you folks keep proposing.
...SCO owns the domain (Do a 'whois' next time before proclaiming "fakeness"...) and
the site IS on one of their servers. Not to mention that the site's main page happens
to be referring to the same thing, coming Early Q1 of 2006.
Just like SCO... Promise everything, have nothing in hand.
If you've got one developer dedicated to the work, what do you have him working on?
/proc filesystem, figured out the bus memory
1) Adding OpenGL 2.0 features (including things like GLSL)- which are sought after but not all 100% done...
2) Fixing an obscure performance problem with the middle of the line Laptop GPUs...
If I were the manager, it'd be #1 that'd get done first- period. How many people do they have there
at ATI doing the Linux stuff? A couple at least- things are shaping up quite nicely as the
drivers have stabilized and while they're not QUITE performant on Linux as they are under Windows.
They still DO have issues lurking in the woodwork- things that DO take time to sort out. If you've
got only a couple (i.e. 2-4 or so) what all do you do to ensure conformance on the BROAD line of
GPUs that they have to support under Linux?
Basically, it's NOT as simple as you'd think- and it's not anywhere as easy as many make it out
to be. It's not simple doing code that doesn't stall the pipelines on a GPU or something
like an iWARP card (Current contract work for moi- trying to drive a certain vendor's iWARP card
to full wireline speeds (10Gbps, here I come!)...)- not everyone can code for it. ATI and
NVidia's right in that it's not a simple thing to code for a GPU's drivers. What they're not
right in is that they're the sole source for this sort of capability in programming on their
chipsets- after all, where did they get their developers in the first place?
More eyes always works better than less of them. More minds working on a problem will invariably
produce results not always envisioned by the original designers. John Carmack, for example,
came up with a clever (but rather evil) hack that allowed the RagePRO UtahGLX support not need
DMA allocated memory, special allocations of RAM, etc.- it just simply started up, allocated
it's driver buffers and then using a feature of the
addresses for the memory just allocated so you could hand it to the chip. No kernel space driver
needed- only root privs which the X server had at the time. Had it been closed technical info,
it probably would have never happened as a supported chipset.
I steadfastly hold to the principle that an SDK interface is an API on the silicon itself and
you shouldn't be hiding things or worrying about "tipping your hat" on things patentable. APIs
are supposed be available in some fasion. I would have problems with requiring
that someone register for the info- but I'd have less of a problem with that than the current
state of affairs. Interested parties would be able to at least do the things I've implied
in this conversation.
You try getting even half of it sometime- it's not been fun for me and I've got things as important
or moreso to produce.