ANNOUNCE: Riva 128/glQuake/D3D driver
Rob Rodgers
David Springer <spri...@matrix.eden.com> wrote:
>This is a D3D wrapper for running glQuake in Win95.
>It's only working with Riva 128 and glQuake 0.93.
Oh, hey, that's great. Now all of the cards I care about have support
for OpenGL. Whew, no more driver worries.
----
http://www.wam.umd.edu/~rsrodger
"I did not get my Spaghetti-O's, I got spaghetti! I want the
press to know this." - Thomas Grasso, executed (OK) 3/20/95
sumatose
Rob Rodgers wrote in message <345a47f4....@news.wam.umd.edu>...
>David Springer <spri...@matrix.eden.com> wrote:
>>This is a D3D wrapper for running glQuake in Win95.
>>It's only working with Riva 128 and glQuake 0.93.
>
>Oh, hey, that's great. Now all of the cards I care about have support
>for OpenGL. Whew, no more driver worries.
So David Springer solved the OpenGL driver availability problem? Now
_that's_ ironic! 8-)
David Matiskella
Rob Rodgers wrote:
> David Springer <spri...@matrix.eden.com> wrote:
> >This is a D3D wrapper for running glQuake in Win95.
> >It's only working with Riva 128 and glQuake 0.93.
>
> Oh, hey, that's great. Now all of the cards I care about have support
> for OpenGL. Whew, no more driver worries.
>
Don't confuse QuakeGL and OpenGL support.
----
> http://www.wam.umd.edu/~rsrodger
>
> "I did not get my Spaghetti-O's, I got spaghetti! I want the
> press to know this." - Thomas Grasso, executed (OK) 3/20/95
--
David Matiskella
Kendall Bennett
In article <63dct7$s84$1...@boris.eden.com>, spri...@matrix.eden.com
says...
> This is a D3D wrapper for running glQuake in Win95.
>
> It's only working with Riva 128 and glQuake 0.93.
I guess you havn't tested it on any other graphics hardware, since I just
tried to run it on my ATI 3D Rage Pro and all I got was a black screen.
Oh well. Perhaps now that your software doesnt work on my system, I
should start flaming you just like your flamed SciTech and our MGL
graphics library. But I wont.
--
+----------------------------------------------------------------------+
| SciTech Software - Building Truly Plug'n'Play Software! |
+----------------------------------------------------------------------+
| Kendall Bennett | To reply via email, remove nospam from |
| Director of Engineering | the reply to email address. Do NOT send |
| SciTech Software, Inc. | unsolicited commercial email! |
| 505 Wall Street | ftp : ftp.scitechsoft.com |
| Chico, CA 95928, USA | www : http://www.scitechsoft.com |
+----------------------------------------------------------------------+
Paul Hsieh
In article <345a47f4....@news.wam.umd.edu>, rsro...@wam.umd.edu says...
> David Springer <spri...@matrix.eden.com> wrote:
> >This is a D3D wrapper for running glQuake in Win95.
> >It's only working with Riva 128 and glQuake 0.93.
>
> for OpenGL. Whew, no more driver worries.
Well, as the Springer himself likes to say:
"Vapourware"
Paul Hsieh
q...@chromatic.com
Mathew J. Binkley
David Springer (spri...@matrix.eden.com) wrote:
: This is a D3D wrapper for running glQuake in Win95.
: It's only working with Riva 128 and glQuake 0.93.
: However, it's pretty awesome on the Riva. On my
: Dell Dimension PII/333 it's getting 30 fps at 800x600.
It also works on the Verite 1000 chipset. It still isn't VQuake, but
I'm getting 10 fps at 640x480 on my P100 with Intergraph Reactor.
You 'da Man!
Mat
---
Mathew J. Binkley - Astronomy TA | "That which we are, we are: one equal
Wake Forest Univ. - Olin 304 | temper of heroic hearts, made weak by
Phone: (910) 759-4957 | time and fate, but strong in will to
WWW: http://www.wfu.edu/~binklmj5 | strive, to seek, to find, and not to
-=* WATCH BABYLON 5 *=- | yield." - Ulysses
Sean Timarco Baggaley
In article <MPG.ec43dd3b...@news.scitechsoft.com>, Kendall
Bennett <Kend...@nospam.scitechsoft.com> writes
>> This is a D3D wrapper for running glQuake in Win95.
>> It's only working with Riva 128 and glQuake 0.93.
>I guess you havn't tested it on any other graphics hardware, since I just
>tried to run it on my ATI 3D Rage Pro and all I got was a black screen.
>Oh well. Perhaps now that your software doesnt work on my system, I
>should start flaming you just like your flamed SciTech and our MGL
>graphics library. But I wont.
Firstly, David Springer wrote this wrapper for one particular card: the
Riva 128.
Secondly, he wrote it ON HIS OWN. IT'S FREE. UNSUPPORTED. Written out of
the goodness of 'is own 'eart.
And most importantly...
Exactly which part of the phrase "It's only working with Riva 128 and
~~~~~~~~~~~~~~~~~~~~~~~~~~
glQuake 0.93" did you not understand? You even quoted it in your reply,
FFS!
--
Sean Timarco Baggaley
Owen Evans
> And most importantly...
> Exactly which part of the phrase "It's only working with Riva 128 and
> ~~~~~~~~~~~~~~~~~~~~~~~~~~
> glQuake 0.93" did you not understand? You even quoted it in your reply,
> FFS!
too. Runs decently on v2100, and horribly on V1000 (what I have) and it
also runs on GLQuake 0.92.
-OrulZ
>
> --
> Sean Timarco Baggaley
David Springer
Kendall Bennett <Kend...@nospam.scitechsoft.com> wrote:
: In article <63dct7$s84$1...@boris.eden.com>, spri...@matrix.eden.com
: says...
: > This is a D3D wrapper for running glQuake in Win95.
: >
: > It's only working with Riva 128 and glQuake 0.93.
: I guess you havn't tested it on any other graphics hardware, since I just
: tried to run it on my ATI 3D Rage Pro and all I got was a black screen.
: Oh well. Perhaps now that your software doesnt work on my system, I
: should start flaming you just like your flamed SciTech and our MGL
: graphics library. But I wont.
Kendall, I did this specifically to support the STB Velocity 128 AGP
sold with Dell Dimensions. Also, it clearly states in a message box
every time you use it what it's been tested with and that it is unsupported.
Now if I claimed that I'd support it for anything more and it didn't
work as claimed I'd expect flamage and gracefully accept it. Hell,
I expect flamage anyhow, that's just the nature of usenet.
David Springer
--
*************** IGAMES INTERNET GAME LOBBY ****************
* *
* NOW SUPPORTING MICROSOFT DirectPlay 3 LOBBY STANDARD ! *
* *
* A real-time game lobby for the internet with many *
* exciting games and thousands of players. Game *
* developers, players, and ISP's can try it out at: *
* *
****************** http://www.igames.com ******************
Javier Arevalo
Kend...@nospam.scitechsoft.com (Kendall Bennett) wrote:
>> It's only working with Riva 128 and glQuake 0.93.
>I guess you havn't tested it on any other graphics hardware, since I just
>tried to run it on my ATI 3D Rage Pro and all I got was a black screen.
Man I donæ„’ think David needs any defense, but to flame someone
because their explicitly unsupported product doesnæ„’ work in places it
says it doesnæ„’ work sounds not too intelligent.
>Oh well. Perhaps now that your software doesnt work on my system, I
>should start flaming you just like your flamed SciTech and our MGL
>graphics library. But I wont.
If you really believe what you are saying I think you should flame or
not say it at all. Wait, wasnæ„’ flaming what you were doing? Itæ„€ like
saying "I would tell you that you are stupid but for other reasons I
wonæ„’". Do you think that person has actually been called stupid? What
would be your reaction?
I think too many people believe that flaming is a universal right
regardless of the actual intelligence shown doing it. When you are an
unknown individual itæ„€ probably ok (you are the only loser), but when
you are representing your company it hurts not just you but your
company. BTW I donæ„’ see David representing Dell (he _is_ representing
Igames), and you are certainly taken as speaking on behalf of Scitech.
You run the risk of damaging Scitechæ„€ public image by doing this. As
a developer who has evaluated Scitechæ„€ products in the past I offer
you my view as honestly as I can: Scitech is not gaining any points by
its founder/CEO/main guy getting into stupid flame wars or posting
messages where most of the content is pure emotional, unthought and
often incorrect information. It tells me you value more your
self-image than Scitechæ„€ stability, and therefore I have to conclude
that Scitech products run a serious risk of going unsupported in the
near future (this is the damage you愉e doing to your own company).
Why are you posting here? To defend Scitech or to defend yourself? If
the former, as a developer I悲 say you better quit. If the later, I悲
say as a friend that you quit as well. Or at least take some more time
before writing.
Does David愀 behaviour piss you off? It pisses me off as well but I惴
not his wife and I donæ„’ have to share his strong attitude at life
every day. So I just extract whatever wisdom is in his messages (a lot
IMHO), and discard the hard words.
Greets
Javier Arevalo
http://www.jet.es/jare
change nospam to jare in the address to send email
Peter Söderman
>In article <63dct7$s84$1...@boris.eden.com>, spri...@matrix.eden.com
>says...
>
>> This is a D3D wrapper for running glQuake in Win95.
>>
>> It's only working with Riva 128 and glQuake 0.93.
>
>I guess you havn't tested it on any other graphics hardware, since I just
>tried to run it on my ATI 3D Rage Pro and all I got was a black screen.
>
He did say that it would only work on the Riva, but it work almost
okay on my V1000.
>Oh well. Perhaps now that your software doesnt work on my system, I
>should start flaming you just like your flamed SciTech and our MGL
>graphics library. But I wont.
>
>--
>
>+----------------------------------------------------------------------+
>| SciTech Software - Building Truly Plug'n'Play Software! |
>+----------------------------------------------------------------------+
>| Kendall Bennett | To reply via email, remove nospam from |
>| Director of Engineering | the reply to email address. Do NOT send |
>| SciTech Software, Inc. | unsolicited commercial email! |
>| 505 Wall Street | ftp : ftp.scitechsoft.com |
>| Chico, CA 95928, USA | www : http://www.scitechsoft.com |
>+----------------------------------------------------------------------+
PeterS
Michael I. Gold
Sean Timarco Baggaley <stbag...@solarflair.demon.co.uk> writes:
| Firstly, David Springer wrote this wrapper for one particular card: the
| Riva 128.
|
| Secondly, he wrote it ON HIS OWN. IT'S FREE. UNSUPPORTED. Written out of
| the goodness of 'is own 'eart.
Actually I believe he obtained the code from someone at Microsoft.
This is the wrapper demonstrated at Meltdown.
--
Michael I. Gold gold at berkelium dot com
Pardon the dust, this .signature is under construction
Daniel Phillips
nos...@jet.es says...
>
>Kend...@nospam.scitechsoft.com (Kendall Bennett) wrote:
>
>>> It's only working with Riva 128 and glQuake 0.93.
>
>>I guess you havn't tested it on any other graphics hardware, since I just
>>tried to run it on my ATI 3D Rage Pro and all I got was a black screen.
>
>because their explicitly unsupported product doesn´t work in places it
>says it doesn´t work sounds not too intelligent.
Right. Do you realize that you're the first to jump into this with a
long, contentless flame-like article? Please give the previous poster
the benefit of the doubt and call it a "bug report". We'll get a lot
further that way.
What about the 3D rage? What's the problem?
--
Daniel Phillips
phillips at dowco.com
eric.w....@lawrence.edu
Actually, it works kinda with the Mystique. I ran it on mine and it
does work but transparancy, (i know that mystique only does alpha
stipple but it is better then none) doesn't work, IE. the flashes on
picking up things, and the textures are handled weird. I ran it with
width=200 and got 11fps in my 6x86MX 200. I hope you refine this d3d
version of the opengl32.dll, it is really cool. I have wanted to
write one myself but am still learning how D3D works.
---Eric W. Bressler
>Doesn't work with Matrox Mystique, although the D3D Quake viewer that
>came out last month did.
>
>
Glenn Corpes
go...@berkelium.com.xyzzy (Michael I. Gold) wrote:
> Sean Timarco Baggaley <stbag...@solarflair.demon.co.uk> writes:
>
> | Firstly, David Springer wrote this wrapper for one particular card:
> the
> | Riva 128.
> |
> | Secondly, he wrote it ON HIS OWN. IT'S FREE. UNSUPPORTED. Written ou
> t of
> | the goodness of 'is own 'eart.
>
> Actually I believe he obtained the code from someone at Microsoft.
> This is the wrapper demonstrated at Meltdown.
A guy I work with who was at meltdown reckons that the lighting was
better.
-=< gco...@ea.com, Project Leader Bullfrog >=-
Jeff Atwood
I am using 0.93. Have you tried it with 0.93? I can't imagine what else
could be wrong with my system, I am incredibly fastidious about drivers and
software. I'm annoyed that I can't get this to work on 3dfx. Guess I'll try
to locate a copy of 0.92.
Jeff
sumatose wrote in message <63l0se$msq$1...@wagner.videotron.net>...
>
>Jon Graehl wrote in message <345EAA5C...@usc.edu>...
>>sumatose wrote:
>>> It's works on my 3DFX here. It's slow, but it looks the same. This
>>> is more
>>> a testimony of Riva's performance and driver, than anything else, I
>>> guess.
>>
>>How did you get it to work on 3dfx? I have the latest drivers for my
>>Righteous 3D ...
>
>I really haven't done anything special. I've gotten GLQuake 0.92, replaced
>the opengl32 in the quake directory with Springer's. I have DX 5.0
>installed. It poped a dialog saying 'Riva Driver.. not surported.. bla
bla'
>clicked OK, and then it was slower, a sure sign that it has worked..
>
>
Jeff Atwood
Hmm, I agree, I have a canopus Pure 3D with latest 3dfx reference
everything, and it doesn't work for me. I really don't think it works on
3dfx.
Jeff
Nathan Rubio
it runs fine in my Verite 1000 Chipset board. Laters....
~Nate
Kendall Bennett <Kend...@nospam.scitechsoft.com> wrote in article
<MPG.ec43dd3b...@news.scitechsoft.com>...
> In article <63dct7$s84$1...@boris.eden.com>, spri...@matrix.eden.com
> says...
>
> > This is a D3D wrapper for running glQuake in Win95.
> >
> > It's only working with Riva 128 and glQuake 0.93.
>
> I guess you havn't tested it on any other graphics hardware, since I just
> tried to run it on my ATI 3D Rage Pro and all I got was a black screen.
>
notoriousGIB
I can't get it to work on my system :( Are there any docs anywhere on the
net? I wish there were some included with it... Anyway, here's what went
wrong:
I have a STB ViRGE/VX (yes. the 3d decellorator :P)
I ran the upgrade, it wrote the DLL file.
I tried GLQW 2.1, the GLQuake that comes with glq3_28.zip, glq8_09.zip and
glq8_27.zip and all three gave the same error:
|------------------------------------ -|
| D3D Wrapper Error |
|--------------------------------------|
| |
| Unable to find 5551 texture |
| |
|--------------------------------------|
and
|--------------------------------------|
| Quake Error |
|--------------------------------------|
| |
| wglCreateConnect failed |
| |
|--------------------------------------|
(I obviously have too much time on my hands to make those illustrations..
:P)
They look good in my font, dunno about yours.. Anyway, you get the idea..
Any idea what cause this, can I fix it, and if so, how?
Thanks
-notoriousGIB
sumatose
Try to switch the current resolution to something like 15/16 bit color
before starting quake.
notoriousGIB wrote in message
<19971104.181919...@WVNVM.WVNET.EDU>...
>I can't get it to work on my system :( Are there any docs anywhere on the
>net? I wish there were some included with it... Anyway, here's what went
>wrong:
>
>I have a STB ViRGE/VX (yes. the 3d decellorator :P)
>I ran the upgrade, it wrote the DLL file.
>I tried GLQW 2.1, the GLQuake that comes with glq3_28.zip, glq8_09.zip and
>glq8_27.zip and all three gave the same error:
>
>| Unable to find 5551 texture |
>
>-notoriousGIB
>
>
>
Kendall Bennett
In article <63i6lt$7d9$3...@boris.eden.com>, spri...@matrix.eden.com
says...
> Kendall, I did this specifically to support the STB Velocity 128 AGP
> sold with Dell Dimensions. Also, it clearly states in a message box
> every time you use it what it's been tested with and that it is unsupported.
> Now if I claimed that I'd support it for anything more and it didn't
> work as claimed I'd expect flamage and gracefully accept it.
Well, since you are going about saying that this is 'your' driver, I
seriously hope that you have not made a big mistake with this. The driver
you released was not written by you; it is a slightly modified version of
the original driver shown at Meltdown by Microsoft. A disassembly of the
driver you released and the one distributed to IHV's by Microsoft, shows
pretty clearly that the two drivers are one and the same. The fact that
the 'GL_RENDERER' string returned by the driver says 'Microsoft' was the
first giveaway.
But the strangest thing of all? Microsoft does not appear to have any
knowledge that this driver has been released, and more specifically I
highly doubt that Microsoft would have given permission to David Springer
to release their driver freely on the internet. I also highly doubt that
Microsoft would have wanted this driver to be distributed anyway.
Did you really get permission to distribute the Microsoft D3D Quake
driver? Or did you think that no-one would be able to tell that the
drivers were one and the same and you would release it anyway? Perhaps a
clear explanation of exactly where you got permission to release this
driver would be in order (I'm sure Microsoft would love to know also).
Wilbur Streett
David Springer <spri...@matrix.eden.com> wrote:
>Kendall Bennett <Kend...@nospam.scitechsoft.com> wrote:
>
>[snip typical Bennett ramblings]
>
>: clear explanation of exactly where you got permission to release this
>: driver would be in order (I'm sure Microsoft would love to know also).
>
>I owe you no explanations Bennett. If you've got a problem with that,
>too bad.
Hmm..
David, you got caught with your hand in the cookie jar..
Fess up.
So how much did you cost?
Wilbur
David Springer
> Hmm..
> Fess up.
> Wilbur
Sorry to ruin your moment but it was all perfectly above board.
Wilbur Streett
David Springer <spri...@matrix.eden.com> wrote:
>
>Sorry to ruin your moment but it was all perfectly above board.
No, it wasn't.
Above board means in the public view.
The back room isn't above board.
Wilbur
David Springer
> No, it wasn't.
> Wilbur
Get a dictionary and use it before attempting to correct my grammar.
From American Heritage Dictionary:
aboveboard adj. Without deceit or trickery; straightforward in manner.
As I said, it was all above board.
Wilbur Streett
David Springer <spri...@matrix.eden.com> wrote:
>Get a dictionary and use it before attempting to correct my grammar.
>
>From American Heritage Dictionary:
>
> aboveboard adj. Without deceit or trickery; straightforward in manner.
>
>As I said, it was all above board.
If there is no deception involved, then why aren't you able to discuss the
terms of the agreement? Why were you the only one that was able to get
around the NDA?
I know what deception is. I also know what aboveboard means.
Making a deal that you can't discuss with Microsoft isn't "aboveboard".
Being able to make that deal because you work at Dell while all of the rest
of the world can't make the same deal isn't "aboveboard".
Want to try again?
Wilbur
David Springer
> Want to try again?
Sure, I'm patient with children and animals so why not you ?
First, you're now making deception and NDA synonymous. I must again
ask that you use a dictionary to discriminate between the two. You
DO know how to use a dictionary, right ?
Secondly, you yourself have refused to divulge details relating to
various things in your experience due to NDAs. Working under your
definitions of deception and NDA one must assume you are also
deceitful. And a hypocrite. But we already knew that didn't we ?
Thirdly, I have no reason to believe that any IHV has the same
access to the wrapper source that I do. Do you have some reason
to believe differently ? It is possible that I was the only one
that thought to ask - "Hey, the source is under NDA, but are
binary compilations of it NDA also ?" Or possibly no IHVs wanted
to support the beast if they let it out. Or possibly no IHV,
except perhaps NVidia had any use for it ?
Lastly, there was no "backroom" involved. There were exactly
four emails involved, with 3 people on the address list,
almost literally as short as:
1) hey, can I check out the wrapper ?
2) sure, here's the source but it's under NDA
3) hey, is a binary compilation of the source also NDA ?
4) nope, if you want to use it for something go ahead
Of course you'll now modify the definition of "back room" to be
a couple of people exchanging simple one line emails... :-)
Wilbur Streett
David Springer <spri...@matrix.eden.com> wrote:
>Wilbur Streett <WStr...@shell.monmouth.com> wrote:
>> David Springer <spri...@matrix.eden.com> wrote:
>
>> >Get a dictionary and use it before attempting to correct my grammar.
>> >
>> >From American Heritage Dictionary:
>> >
>> > aboveboard adj. Without deceit or trickery; straightforward in manner.
>> >
>> >As I said, it was all above board.
>
>> If there is no deception involved, then why aren't you able to discuss the
>> terms of the agreement? Why were you the only one that was able to get
>> around the NDA?
>
>> I know what deception is. I also know what aboveboard means.
>
>> Making a deal that you can't discuss with Microsoft isn't "aboveboard".
>> Being able to make that deal because you work at Dell while all of the rest
>> of the world can't make the same deal isn't "aboveboard".
>
>> Want to try again?
>
>Sure, I'm patient with children and animals so why not you ?
>First, you're now making deception and NDA synonymous.
No, you are claiming that I'm making them synonymous. I'm not. I'm
claiming that making a distribution deal with Microsoft on code that's only
available under NDA isn't quite aboveboard. My metaphore for aboveboard is
that the cards are played "under the board" rather than above them. There
are two players at the table that made a deal that wasn't shown above the
table, or above board.
>I must again ask that you use a dictionary to discriminate between the two. You
>DO know how to use a dictionary, right ?
Well, I'll dig up deception so that you can see that you don't know these
words as well as you think you do.
At http://www2.lib.uchicago.edu/cgi-bin/www2/WEBSTER.sh?WORD=deceive
You'll find:
ARTFL Project: Webster Dictionary, 1913
Searching for: "deceive"
Found 1 hit(s).
Deceive (Page: 375)
De*ceive" (?), v. t. [imp. & p. p. Deceived (?); p. pr. & vb. n.
Deceiving.] [OE. deceveir, F. décevoir, fr. L. decipere to catch, insnare,
deceive; de- + capere to take,
catch. See Capable, and cf. Deceit, Deception.]
1. To lead into error; to cause to believe what is false, or disbelieve
what is true; to impose upon; to mislead; to cheat; to disappoint; to
delude; to insnare.
Evil men and seducers shall wax worse and worse, deceiving, and being
deceived. 2 Tim. iii. 13.
Nimble jugglers that deceive the eye. Shak.
What can 'scape the eye Of God all-seeing, or deceive his heart?
Milton.
2. To beguile; to amuse, so as to divert the attention; to while away; to
take away as if by deception.
These occupations oftentimes deceived The listless hour. Wordsworth.
3. To deprive by fraud or stealth; to defraud. [Obs.]
Plant fruit trees in large borders, and set therein fine flowers, but
thin and sparingly, lest they deceive the trees. Bacon.
Syn. -- Deceive, Delude, Mislead. Deceive is a general word applicable to
any kind of misrepresentation affecting faith or life. To delude,
primarily, is to make sport of, by deceiving,
and is accomplished by playing upon one's imagination or credulity, as by
exciting false hopes, causing him to undertake or expect what is
impracticable, and making his failure ridiculous.
It implies some infirmity of judgment in the victim, and intention to
deceive in the deluder. But it is often used reflexively, indicating that a
person's own weakness has made him the sport
of others or of fortune; as, he deluded himself with a belief that luck
would always favor him. To mislead is to lead, guide, or direct in a wrong
way, either willfully or ignorantly.
I can easily see how what you are talking about fits into all 3 of the
definitions above. Disclosure is the antithesis of deception. Having
signed an NDA by definition means that deception is occuring. My brat (as
in military brat) upbringing means that I know all about truth and
deception, those are very strong standards in the military, my father made
it very clear to me how not disclousing was a lie. Sorry if you don't like
it, but that is the definition.
>Secondly, you yourself have refused to divulge details relating to
>various things in your experience due to NDAs. Working under your
>definitions of deception and NDA one must assume you are also
>deceitful.
Yes, I have been a party to deception in the specific case that I signed an
NDA on a significant bug in chips that are in millions of computers around
the world. I was working on a professional services basis, so by the
accepted ethics of that professional services agreement I was bound to not
disclouse an issue which would be detrimental to the company that
contracted my services. I have rethought that particular NDA even recently
and I think that I may need to contact the legal staff of the company that
I signed the NDA of to discuss this particular situation. The NDA had
language to the effect that if the NDA was to be in excess of a year that I
was supposed to be flown out to California, so I initially signed the NDA
for period of one year, then it was returned to me and I was asked to
modify it for life. The nature of the issue that I agreed to not disclose
is significant enough that I am sorry that I even found the problem. I
have had the NDA reviewed and I've been told that the NDA is invalid, since
by law an NDA must have a limited term. On the other hand, I think it
would be in the interest of the company that I signed the NDA with to
resolve this particular problem and hang the people out to dry who created
this mess inside the company.
On the other hand, in private email with you I gave you a couple of hints
on how to find the problem. (There may have been some public one's as
well). Every chip known to man has an errata sheet, except those chips
that have problems so large that they don't want any liability on
disclosure. Find the chip without an errate sheet and you'll be onto the
problem. I also spelled out how if this particular chip was in Dell
equipment that your company would have probably signed the an NDA about the
problem since the manufacturer required every vendor to sign this NDA
before they released the software patch that "sort of" fixed the problem.
Of course, every vendor that got this chip with this problem was dying for
an excuse to not have to do a recall, or take any liability, so sign they
did. I also offered to look over a component list of all Dell equipment
and tell you if the chip was or was not in the component list.
Of course, the problem exhibits itself in a Windows machine that locks up,
but we all know that Windows crashes all the time...
So, yes, I do feel bad about that one. But I didn't set up the rules of
society, nor am I releasing a product based on that particular NDA
agreement.
>And a hypocrite.
Any time that I am, I admit it. But I'm not in this case. I stated that
there are chips with bugs, that's not a new story. I'm not claiming that
any technology is better or worse because of the problems that I am aware
of, I have been around long enough to know that things break down, and shit
get's swept under the carpet.
>But we already knew that didn't we ?
No, we already knew that you were grasping for straws..
>Thirdly, I have no reason to believe that any IHV has the same
>access to the wrapper source that I do. Do you have some reason
>to believe differently ?
Not relative.
>It is possible that I was the only one that thought to ask -
>"Hey, the source is under NDA, but are binary compilations of it
>NDA also ?"
I believe that the NDA agreement probably included language the to the
effect of "derivative works" which includes binary compilations... Or are
you such the dupe of Microsoft that you don't realize your own part in the
deception?
>Or possibly no IHVs wanted to support the beast if they let it out.
>Or possibly no IHV, except perhaps NVidia had any use for it ?
Or possibly, not IHV was given the ability to release it, and had no
motivation since they had already been told by Microsoft that they had to
support DirectX and have been specifically told by Microsoft what the
policy is on OpenGL?
Like I said, not above board.
>Lastly, there was no "backroom" involved. There were exactly
>four emails involved, with 3 people on the address list,
>almost literally as short as:
>
>1) hey, can I check out the wrapper ?
>2) sure, here's the source but it's under NDA
>3) hey, is a binary compilation of the source also NDA ?
>4) nope, if you want to use it for something go ahead
I hope that you got it in writing, since I'd bet that the NDA includes the
language "derivative works" in it..
>Of course you'll now modify the definition of "back room" to be
>a couple of people exchanging simple one line emails... :-)
Back room means non-public. Private means non-public. No change in my
definition.
Wilbur
Daniel Phillips
WStr...@shell.monmouth.com says...
>David Springer <spri...@matrix.eden.com> wrote:
>>
>>Lastly, there was no "backroom" involved. There were exactly
>>four emails involved, with 3 people on the address list,
>>almost literally as short as:
>>
>>1) hey, can I check out the wrapper ?
>>2) sure, here's the source but it's under NDA
>>3) hey, is a binary compilation of the source also NDA ?
>>4) nope, if you want to use it for something go ahead
>>Of course you'll now modify the definition of "back room" to be
>>a couple of people exchanging simple one line emails... :-)
>
>Back room means non-public. Private means non-public. No change
>in my definition.
>
>Wilbur
Does anybody really believe David Springer got hold of this source
after 4 one-line emails? Has anyone *ever* seen David Springer
limit himself to one-line *anything*? Credibility==0. Just more
cynical deception.
q...@chromatic.com
> >>1) hey, can I check out the wrapper ?
> >>2) sure, here's the source but it's under NDA
> >>3) hey, is a binary compilation of the source also NDA ?
> >>4) nope, if you want to use it for something go ahead
> [...]
Unfortunately folks, this conversation is quite likely to have taken
place. A little while ago, a bunch of graphics IHVs (including mine)
were *given* the source to this abomination for the expressed purpose of
"doing what we like with it." I say abomination because ... well it
makes even the Riva look bad. This "2 week job" did indeed yield "2 week
quality" only nobody got to look close enough when Microsoft first
displayed it.
The intent, of course, is for GLQuake to work on D3D so that Microsoft
can do that maximum possible damage to the OpenGL movement (striking at
id). (Man, when does the $1 million a day start kicking in?) IHVs will
react to this as required, I should think, but it seems rather pointless
considering Quake 2 is basically days away from shipping and the wrapper
aint going to work on it.
--
Paul Hsieh
q...@chromatic.com
-------------------==== Posted via Deja News ====-----------------------
http://www.dejanews.com/ Search, Read, Post to Usenet
Wilbur Streett
q...@chromatic.com wrote:
>Unfortunately folks, this conversation is quite likely to have taken
>place. A little while ago, a bunch of graphics IHVs (including mine)
>were *given* the source to this abomination for the expressed purpose of
>"doing what we like with it." I say abomination because ... well it
I wouldn't doubt that Microsoft would look to change the definition of the
term OpenGL so that it might be thought to include "runs on D3D". Their
express purpose in life seems to be to own words and change their meanings.
For some reason they think that this is a viable business practice. They
are playing that game with the word "innovation" right now. Look up
innovation and you'll see that it's got a definition that includes,
"doesn't have to be new, just as long as the customer never saw it before."
I think that we know what they have done with "Windows" "Word", etc..
Wilbur
Glenn Corpes
WStr...@shell.monmouth.com (Wilbur Streett) wrote:
> q...@chromatic.com wrote:
>
> >Unfortunately folks, this conversation is quite likely to have taken
> >place. A little while ago, a bunch of graphics IHVs (including
mine)
> >were *given* the source to this abomination for the expressed purpose
> of
> >"doing what we like with it." I say abomination because ... well it
>
> I wouldn't doubt that Microsoft would look to change the definition of
> the
> term OpenGL so that it might be thought to include "runs on D3D". The
> ir
> express purpose in life seems to be to own words and change their mean
> ings.
I think it's completely obvious why MS wrote it, it was to 'prove' that
Quake could have been done with D3D and to shut up the people who
claimed it was twice as slow as openGL, it undoubtedly did this
although, after seeing it run I think that it sucked a bit, sure it
threw polygons around faster than some GL drivers and only marginally
slower than others (quite impressive when you think it's running through
two APIs) but entirely failed on the texture uploading front, what
happened to the lightmaps?. I also think that David genuinely wanted to
release it to get Quake running on the Riva. Put your paranoia on hold
for a second and think about it, if Microsoft wanted to use it for the
purposes you suggest they would have been better off leaving it under
NDA and not letting people see how unfinished it was.
Shawn Swift
q...@chromatic.com wrote:
>
>> >>1) hey, can I check out the wrapper ?
>> >>2) sure, here's the source but it's under NDA
>> >>3) hey, is a binary compilation of the source also NDA ?
>> >>4) nope, if you want to use it for something go ahead
>
>Unfortunately folks, this conversation is quite likely to have taken
>place. A little while ago, a bunch of graphics IHVs (including mine)
>were *given* the source to this abomination for the expressed purpose of
>"doing what we like with it." I say abomination because ... well it
>quality" only nobody got to look close enough when Microsoft first
>displayed it.
>
>The intent, of course, is for GLQuake to work on D3D so that Microsoft
>can do that maximum possible damage to the OpenGL movement (striking at
>id). (Man, when does the $1 million a day start kicking in?) IHVs will
>react to this as required, I should think, but it seems rather pointless
>considering Quake 2 is basically days away from shipping and the wrapper
>aint going to work on it.
So, does this wrapper perform as advertised? Running a normal game of
GLQuake against a normal game of GLQuake run throguh this wrapper on a
3DFX card, do yuo get similar framerates, or is that only if you turn
off all the lighting?
Wilbur Streett
gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
>I think it's completely obvious why MS wrote it, it was to 'prove' that
>Quake could have been done with D3D and to shut up the people who
>claimed it was twice as slow as openGL,
Glenn.. Perhaps you forgot that it required a total rewrite of D3D in
order to emulate the tiny subset of OpenGL that Quake uses?
>it undoubtedly did this
>although, after seeing it run I think that it sucked a bit, sure it
>threw polygons around faster than some GL drivers and only marginally
>slower than others (quite impressive when you think it's running through
>two APIs) but entirely failed on the texture uploading front, what
>happened to the lightmaps?.
Well, I really haven't bothered to check it, so I'll have to take your word
on that.
>I also think that David genuinely wanted to
>release it to get Quake running on the Riva. Put your paranoia on hold
>for a second and think about it, if Microsoft wanted to use it for the
>purposes you suggest they would have been better off leaving it under
>NDA and not letting people see how unfinished it was.
I'm not paranoid.. Microsoft doesn't change NDA's on a whim.. They are run
by a Lawyer's son..
Wilbur
Glenn Corpes
WStr...@shell.monmouth.com (Wilbur Streett) wrote:
> gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
>
> >I think it's completely obvious why MS wrote it, it was to 'prove' th
> at
> >Quake could have been done with D3D and to shut up the people who
> >claimed it was twice as slow as openGL,
>
> Glenn.. Perhaps you forgot that it required a total rewrite of D3D in
> order to emulate the tiny subset of OpenGL that Quake uses?
Are you trying to suggest that DrawPrimitive was done to enable the
writing of the Quake wrapper?, if so it nicely illustrates how screwed
up your mind is on this subject,
[...]
> >I also think that David genuinely wanted to
> >release it to get Quake running on the Riva. Put your paranoia on hol
> d
> >for a second and think about it, if Microsoft wanted to use it for th
> e
> >purposes you suggest they would have been better off leaving it under
> >NDA and not letting people see how unfinished it was.
>
> I'm not paranoid.. Microsoft doesn't change NDA's on a whim.. They ar
> e run
> by a Lawyer's son..
What does that say about my last comment?, it's just a statistical
likelihood in a country with 5% of the worlds population and 70% of it's
lawyers.
Daniel Phillips
gle...@cix.co.uk.notthisbityoumoron says...
>
>>I'm not paranoid.. Microsoft doesn't change NDA's on a whim.. They
>
>What does that say about my last comment?, it's just a statistical
>likelihood in a country with 5% of the worlds population and 70% of it's
>lawyers.
Glenn, I genuinely believe that you are naive - you're not just cynically
regurgitating the Microsoft spin. Gates built Microsoft mainly be being
extremely agressive legally. Being a lawyer's son has *everything* to do
with his success. Gates has been so successful using the U.S. legal
system to his advantage that he now thinks he's above the law - Balmer
summed it up: "to heck with Janet Reno". Gates now believes he's more
powerful legally than the Department of Justice.
Wake up, Glenn.
Wilbur Streett
gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
>WStr...@shell.monmouth.com (Wilbur Streett) wrote:
>
>> gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
>>
>> >I think it's completely obvious why MS wrote it, it was to 'prove' th
>> at
>> >Quake could have been done with D3D and to shut up the people who
>> >claimed it was twice as slow as openGL,
>>
>> Glenn.. Perhaps you forgot that it required a total rewrite of D3D in
>> order to emulate the tiny subset of OpenGL that Quake uses?
>
>Are you trying to suggest that DrawPrimitive was done to enable the
>writing of the Quake wrapper?, if so it nicely illustrates how screwed
>up your mind is on this subject,
Glenn, you really have to stop projecting how paranoid you think other
people are on to them.
I'm sure that Microsoft was looking for some response to the Carmark plan
comments.. or do you think that the wrapper happened "by accident?" I
haven't dug into the OpenGL subset enough to be able to make a reasoned
decision as to whether the design of the D3D5 API's design was skewed to
include the functionality necessary to do a Quake clone on MS Technology...
There are numerous examples of this sort of tactic in Microsoft's past, so
why would I expect anything different this time?
<snip>
>> I'm not paranoid.. Microsoft doesn't change NDA's on a whim.. They ar
>> e run
>> by a Lawyer's son..
>
>What does that say about my last comment?, it's just a statistical
>likelihood in a country with 5% of the worlds population and 70% of it's
>lawyers.
Glenn, you overestimate the number of son's of lawyers that there are as a
percentage of the population. You also underestimate Microsoft's legal
department's influence on the running of the company. Or do you think that
a bunch of idiots can hold the US Department of Justice at bay with
language games?
Wilbur
sumatose
Wilbur Streett wrote:
>
> gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
>
> >I think it's completely obvious why MS wrote it, it was to 'prove' that
> >Quake could have been done with D3D and to shut up the people who
> >claimed it was twice as slow as openGL,
>
> Glenn.. Perhaps you forgot that it required a total rewrite of D3D in
> order to emulate the tiny subset of OpenGL that Quake uses?
That's completely false. Try to argue with true facts, please.
Are you talking about DrawPrimitive? It's a function call to
draw triangles -are you saying that D3D didn't draw triangles
before? Com'on!
Everything was in Direct3D from version 1.0 to do what GLQuake
does, from primitives to filtering - you should know that.
Much uglier to program and read before DrawPrimitive, sure.
But DrawPrimitive does not add any features to Direct3D and the wrapper didn't
need any additionnal feature to do what Quake needed. In fact, Carmack has
said in his famous .plan that he doesn't need a wide array of features - and
Direct3D provides what the current accelerators are able to do, anyway.
Showing that GLQuake could have been made with Direct3D was precisly
the point of the driver - and running the D3D wrapper on a 3DFX shows
this clearly. Both APIs have the same features : you can read Carmack
say it in his old .plan and in the current BOOT magazine.
Wilbur Streett
sumatose <suma...@NOSPAM.usa.net> wrote:
>Wilbur Streett wrote:
>>
>> gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
>>
>> >I think it's completely obvious why MS wrote it, it was to 'prove' that
>> >Quake could have been done with D3D and to shut up the people who
>> >claimed it was twice as slow as openGL,
>>
>> Glenn.. Perhaps you forgot that it required a total rewrite of D3D in
>> order to emulate the tiny subset of OpenGL that Quake uses?
>That's completely false. Try to argue with true facts, please.
Well, then by your claims, the demonstration of a Quake wrapper at the DX5
release had nothing to do with any sort of reality at all..
You can't have it both ways..
Wilbur
Kent Sandvik
(Daniel Phillips) wrote:
> You put your finger on what's wrong with D3D. *Current* accelerators.
> Current accelerators become obsolete acclerators very quickly. Who
> wants to program a platform that's primarily designed for obsolete
> acceleraters?
It's true that HW accelerators today are old meat next year this time -
whether they support OpenGL or D3D, or both. But they will still work.
I.e. don't see the issue of them becoming obsolete. Maybe a post PCI bus
structure will make some of the cards obsolete, but then you just
purchase a new sub-200$ card at Fries, and that's it. I don't think
anyone in the computing industry any longer expects that their computers
are state of the art more than about three months max.
Cheers, Kent
--
Remove z from my email address above if you want to respond directly
(this is to avoid spam emails).
Kent Sandvik
David Springer wrote:
>
> I think you guys should check out the class action lawsuits filed
> by two major stock brokerage firms against SGI. It seems there's
> some executives at SGI who've been cooking the books to pump up
> the stock price to $30+ for a quarter, selling their holdings,
> and then the stock tumbles to $15- the next quarter when the
> real company performance re-emerges.
>
> How does this news stack up with your view of SGI as the white
> knights and Microsoft as the black knights ? Can you point to
> any lawsuits where senior executives at Microsoft are accused
> of screwing the shareholders ?
This has really nothing to do with r.g.p, but any stock that suddenly
takes a dive will most likely generate similar class action lawsuits.
Anyone following this in the silicon valley industry sees it happening
every month. The main reason MS has not encountered this before that
very seldom has their stock dropped dramatically -- it's indeed a sign
of a well managed company and good sales, but has nothing to do with
good cops and bad cops, that's just pure speculation.
Cheers, Kent
PS: And yes, I also own a big pile of SGI stock, and it's not fun owning
it just now.
Daniel Phillips
suma...@NOSPAM.usa.net says...
>In fact, Carmack has said in his famous .plan that he doesn't need a
>wide array of features -
******
>and Direct3D provides what the current accelerators are able to do,
>anyway.
******
You put your finger on what's wrong with D3D. *Current* accelerators.
Current accelerators become obsolete acclerators very quickly. Who
wants to program a platform that's primarily designed for obsolete
acceleraters?
Especially when there's a better platform available that was actually
designed with a consistent rendering model... provides the basics and
looks forward too, right from the word go. And is cross-platform too.
<drifting further towards rantzone>
Microsoft made a big mistake in spamming the game developers with D3D
instead of comitting fully to OpenGL - it hasn't done their
credibility a bit of good. Not too late to kiss and make up, though.
I think that's what we're really waiting for. It takes a lot of
maturity to admit you were wrong, and I don't think Billco has that
maturity.
David Springer
Wilbur Streett <WStr...@shell.monmouth.com> wrote:
> gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
> >I think it's completely obvious why MS wrote it, it was to 'prove' that
> >Quake could have been done with D3D and to shut up the people who
> >claimed it was twice as slow as openGL,
> Glenn.. Perhaps you forgot that it required a total rewrite of D3D in
> order to emulate the tiny subset of OpenGL that Quake uses?
That' just plain wrong Wilbur. I've got the code and AFAIK there's
only one DX6 extension in it and that's DirectPrimitive().
Moreover, ExecuteBuffer() could be substituted for DirectPrimitive()
but unless one added some code to cache streams of OpenGL calls
into larger EBs the EBs wouldn't be very efficient. It would also
be a LOT of work getting the precedural streaming OGL calls into
the OO execute buffers. Different paradigms. OpenGL is procedural
and EBs are object oriented. I can understand why the vast majority
of coders, with procedural 3D engines, were screaming bloody murder
about execute buffers.
EBs will, IMO, undoubtedly be the highest performance method for
hardware whose native language includes EB commands. Think of a
higher level graphics language, with pointers, jumps, branches,
calls, and returns along with 2D and 3D primitives for triangles,
vectors, polys, textures and so forth. Now think of bug chunks
of that code that make up bigger objects and parts of scenes.
Now imagine that code sitting in system memory with a graphics
chip that can haul it in and execute it directly over a 1.2 gigabyte
per second AGP bus.
That's what EBs do and that's how consumer PCs are being built.
OpenGL doesn't stand a chance at competing in that archtectural
space. You simply can't subjugate the x86 CPU to streaming a
zillion separate OpenGL function calls for every minor scene
change. You instead change a few key parameters in a big EB,
tell the graphics chip to execute it (ExecuteBuffer()), and get
the x86 doing other useful things in the meantime, like processing
the geometry for the next frame.
If consumer PC hardware were being architected like SGI workstations
it would be a different story, but they ain't. They're being built
for low cost with geometry predominantly handled by the x86 (so
Intel can sell ever more powerful processors) and graphics boards
with limited on-board memory (so you can intelligently re-allocate
system memory according to need when you aren't doing intensive 3D),
graphics boards that are mostly just 2D and 3D rasterization engines.
No amount of crying or handwaving is going to change this route. It's
a done deal and a road already chosen by the forces that control
what PC hardware looks like year over year. It's BEEN a done deal
for quite a while now... much longer than this silly OpenGL vs.
D3D war has been raging. Maybe you'll believe me now that Dell is
selling exactly this architecture and it won us PC Computing Product
of the Year at Comdex, MVP for power consumer desktop, and if not for
the immature Riva 128 drivers a "Kick Ass" rating in Boot Magazine ?
What more need I say or point out to convince you ? You're an
extremly bright guy Wilbur when you don't let Wintel paranoia take
over your higher brain functions. Stand back for a moment and consider
this with the names Intel and Microsoft removed and see if it
doesn't make a whole bunch of good sense.
David Springer
Dimension Product Group
Dell Computer Corporation
> >it undoubtedly did this
> >although, after seeing it run I think that it sucked a bit, sure it
> >threw polygons around faster than some GL drivers and only marginally
> >slower than others (quite impressive when you think it's running through
> >two APIs) but entirely failed on the texture uploading front, what
> >happened to the lightmaps?.
> Well, I really haven't bothered to check it, so I'll have to take your word
> on that.
> >I also think that David genuinely wanted to
> >release it to get Quake running on the Riva. Put your paranoia on hold
> >for a second and think about it, if Microsoft wanted to use it for the
> >purposes you suggest they would have been better off leaving it under
> >NDA and not letting people see how unfinished it was.
> I'm not paranoid.. Microsoft doesn't change NDA's on a whim.. They are run
> by a Lawyer's son..
> Wilbur
David Springer
Microsoft undoubtedly deserves a spanking, which I think they've
received, for not seeing the short term need for DirectPrimitive()
and having it in there from the start. EBs are not a good answer
for fast painless ports of procedural software 3D engines. Nor
are they particulary advantageous for graphics hardware that
doesn't natively support direct execution of the EB instruction
set. They're still pretty cool for OO code though but that also
requires a major paradigm shift on the part of the game coders.
So, spank them for being short sighted, nobody is perfect, but
be honest and acknowledge the long term benefit of the architecture.
And don't be silly by suggesting that DX5 was changed just to facilitate
a D3D Quake driver. DirectPrimitive() was added to correct a
glaring oversight for 1997 and, to a lesser extent, 1998 3D games.
David Springer
Glenn Corpes <gle...@cix.co.uk.notthisbityoumoron> wrote:
> WStr...@shell.monmouth.com (Wilbur Streett) wrote:
> > gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
> >
> > >I think it's completely obvious why MS wrote it, it was to 'prove' th
> > at
> > >Quake could have been done with D3D and to shut up the people who
> > >claimed it was twice as slow as openGL,
> >
> > Glenn.. Perhaps you forgot that it required a total rewrite of D3D in
> > order to emulate the tiny subset of OpenGL that Quake uses?
> Are you trying to suggest that DrawPrimitive was done to enable the
> writing of the Quake wrapper?, if so it nicely illustrates how screwed
> up your mind is on this subject,
> [...]
> > >I also think that David genuinely wanted to
> > >release it to get Quake running on the Riva. Put your paranoia on hol
> > d
> > >for a second and think about it, if Microsoft wanted to use it for th
> > e
> > >purposes you suggest they would have been better off leaving it under
> > >NDA and not letting people see how unfinished it was.
> >
> > I'm not paranoid.. Microsoft doesn't change NDA's on a whim.. They ar
> > e run
> > by a Lawyer's son..
> What does that say about my last comment?, it's just a statistical
> likelihood in a country with 5% of the worlds population and 70% of it's
> lawyers.
> -=< gco...@ea.com, Project Leader Bullfrog >=-
--
David Springer
I think you guys should check out the class action lawsuits filed
by two major stock brokerage firms against SGI. It seems there's
some executives at SGI who've been cooking the books to pump up
the stock price to $30+ for a quarter, selling their holdings,
and then the stock tumbles to $15- the next quarter when the
real company performance re-emerges.
How does this news stack up with your view of SGI as the white
knights and Microsoft as the black knights ? Can you point to
any lawsuits where senior executives at Microsoft are accused
of screwing the shareholders ?
You can get URLs to the lawsuit news by going to www.yahoo.com,
go to stock quotes, pull up a quote for SGI, and then check the
URLs to recent news about the company.
I won't argue about Microsoft sometimes mercilessly putting the
screws onto OTHER companies, that's business and it's in the
best interest of the millions of people, Gates included, that
own Microsoft stock. Putting the screws to the very people that
believed in you and invested in you is, however, truly abhorrent
in comparison.
David Springer
Wilbur Streett <WStr...@shell.monmouth.com> wrote:
> gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
> >WStr...@shell.monmouth.com (Wilbur Streett) wrote:
> >
> >> gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
> >>
> >> >I think it's completely obvious why MS wrote it, it was to 'prove' th
> >> at
> >> >Quake could have been done with D3D and to shut up the people who
> >> >claimed it was twice as slow as openGL,
> >>
> >> Glenn.. Perhaps you forgot that it required a total rewrite of D3D in
> >> order to emulate the tiny subset of OpenGL that Quake uses?
> >
> >Are you trying to suggest that DrawPrimitive was done to enable the
> >writing of the Quake wrapper?, if so it nicely illustrates how screwed
> >up your mind is on this subject,
> Glenn, you really have to stop projecting how paranoid you think other
> people are on to them.
> I'm sure that Microsoft was looking for some response to the Carmark plan
> comments.. or do you think that the wrapper happened "by accident?" I
> haven't dug into the OpenGL subset enough to be able to make a reasoned
> decision as to whether the design of the D3D5 API's design was skewed to
> include the functionality necessary to do a Quake clone on MS Technology...
> There are numerous examples of this sort of tactic in Microsoft's past, so
> why would I expect anything different this time?
> <snip>
> >> I'm not paranoid.. Microsoft doesn't change NDA's on a whim.. They ar
> >> e run
> >> by a Lawyer's son..
> >
> >What does that say about my last comment?, it's just a statistical
> >likelihood in a country with 5% of the worlds population and 70% of it's
> >lawyers.
> Glenn, you overestimate the number of son's of lawyers that there are as a
> percentage of the population. You also underestimate Microsoft's legal
> department's influence on the running of the company. Or do you think that
> a bunch of idiots can hold the US Department of Justice at bay with
> language games?
> Wilbur
--
Daniel Phillips
zsan...@best.com says...
>
>(Daniel Phillips) wrote:
>> You put your finger on what's wrong with D3D. *Current* accelerators.
>> Current accelerators become obsolete acclerators very quickly. Who
>> wants to program a platform that's primarily designed for obsolete
>> acceleraters?
>
>whether they support OpenGL or D3D, or both. But they will still work.
>I.e. don't see the issue of them becoming obsolete.
The problem is that your Direct3D code gets obsolete along with the card.
I'd rather have something that automatically gets better as cards get
better.
David Springer
Daniel Phillips <xy...@plugh.com> wrote:
> The problem is that your Direct3D code gets obsolete along with the card.
> I'd rather have something that automatically gets better as cards get
> better.
I don't think so. The fastest consumer hardware will support EB instruction
set as native code. It'll be a long time before that paradigm is replaced.
Therefore, as long as the EB instruction set is backward compatible with
newer hardware, just like Pentium II's can still run 15 year old 8088
code, there's no problem. New cards will run the old code faster and
faster while the code itself can be given facelifts when and if appropriate
to take advantage of an expanded instruction set.
Pretty cool, eh ?
David Springer
go...@berkelium.com.xyzzy
David Springer <spri...@matrix.eden.com> writes:
>> I think you guys should check out the class action lawsuits filed
>> by two major stock brokerage firms against SGI. It seems there's
>> some executives at SGI who've been cooking the books to pump up
>> the stock price to $30+ for a quarter, selling their holdings,
>> and then the stock tumbles to $15- the next quarter when the
>> real company performance re-emerges.
Yawn. That's a nice story, but there are always two sides to every
story. Do you have any idea how many of these frivolous lawsuits are
filed every year? Virtually every time a company's stock drops more
than a few points, some contingency lawyer is there with a suit.
In fact the people who were talking up the quarter were not the same
people who sold during that period. I suppose you'll argue there were
kickbacks involved, though.
>> How does this news stack up with your view of SGI as the white
>> knights and Microsoft as the black knights ? Can you point to
>> any lawsuits where senior executives at Microsoft are accused
>> of screwing the shareholders ?
No, Microsoft just screws their customers and partners instead. Give
me a break!
>> I won't argue about Microsoft sometimes mercilessly putting the
>> screws onto OTHER companies, that's business and it's in the
>> best interest of the millions of people, Gates included, that
>> own Microsoft stock. Putting the screws to the very people that
>> believed in you and invested in you is, however, truly abhorrent
>> in comparison.
And how abhorrent is screwing the very people that believed in you and
invested in your products?
Check out SGI's customer satisfaction ratings over the last decade.
They consistently recieve high marks for product quality and support.
(Yes, there are always a few who are unhappy, but the majority seem
quite please, according to Datapro Information Services.
In contrast, there are millions of mindless zombies cussing their PC's
and assuming "that's just the way things are" and they don't even
realize how much better it could be, if their OS vendor cared about
anything but its own profits.
Sigh. If its not one thing its another.
Daniel Phillips
spri...@matrix.eden.com says...
>
>I think you guys should check out the class action lawsuits filed
>by two major stock brokerage firms against SGI. It seems there's
>some executives at SGI who've been cooking the books to pump up
>the stock price to $30+ for a quarter, selling their holdings,
>and then the stock tumbles to $15- the next quarter when the
>real company performance re-emerges.
From CNET <the part you ignored asshole=38>:
Silicon Graphics has managed to stave off
class-action suits before. Earlier this year, a federal
judge in California dismissed for a second time a
shareholder class-action case brought against SGI.
In that case, the judge cited Congress's adoption of
a more stringent standard for class-action suits.
That new law required plaintiffs to demonstrate that
a defendant deliberately acted fraudulently, moving
the burden of proof beyond a plaintiff merely
alleging false or misleading statements.
>How does this news stack up with your view of SGI as the white
>knights and Microsoft as the black knights ?
Easy, it makes you the Black Kights chamber pot. A very useful
function - a great source of slop to toss at the White Knight! :-)
James Shaw
>The problem is that your Direct3D code gets obsolete along with the card.
Hows that? My original D3D code still works with DX5, and it will
benefit from any hardware acceleration available.
Jim
David Springer
go...@berkelium.com.xyzzy wrote:
> David Springer <spri...@matrix.eden.com> writes:
> >> I think you guys should check out the class action lawsuits filed
> >> by two major stock brokerage firms against SGI. It seems there's
> >> some executives at SGI who've been cooking the books to pump up
> >> the stock price to $30+ for a quarter, selling their holdings,
> >> and then the stock tumbles to $15- the next quarter when the
> >> real company performance re-emerges.
> Yawn. That's a nice story, but there are always two sides to every
> story. Do you have any idea how many of these frivolous lawsuits are
> filed every year? Virtually every time a company's stock drops more
Hmmmm... no. I have no idea how many of these frivolous lawsuits
are filed each year. I only added SGI to stocks that I watch a few
months ago. Is it a regular thing for SGI ? I've never seen one
against Sun, IBM, Microsoft, Dell, Compaq, Intel, or Gateway which are
companies I watch along with SGI.
Do *you* have any idea how many of these suits were filed ? Is your
idea maybe -zero- if we discount these two against SGI ? Did you
bother to read the details or just instantly dismiss it ?
> than a few points, some contingency lawyer is there with a suit.
> In fact the people who were talking up the quarter were not the same
> people who sold during that period. I suppose you'll argue there were
> kickbacks involved, though.
Nope, I'll just direct people to the URL and let them follow the
story if they want. I don't make the news, I just read it.
Funny thing Michael, all the stocks I mentioned dropped considerably
more than a few points in the last month and I didn't see any other
lawsuits filed by angry stockholders. Again, I'd really appreciate
you pointing me towards these examples. You DO have data to back
up your claims of course.
> >> How does this news stack up with your view of SGI as the white
> >> knights and Microsoft as the black knights ? Can you point to
> >> any lawsuits where senior executives at Microsoft are accused
> >> of screwing the shareholders ?
> No, Microsoft just screws their customers and partners instead. Give
> me a break!
Uh huh. Microsoft's customers pay about $20 for Windows 95 and everything
that comes bundled with it. Care to itemize a comparable list of
goodies sold by any other major software vendor and compare value ?
I'm willing to entertain the idea that they've screwed a partner
before, given that they've got thousands of partners, but (your words)
there's two sides to all of those stories too. Any specific examples
you'd like to mention ?
> >> I won't argue about Microsoft sometimes mercilessly putting the
> >> screws onto OTHER companies, that's business and it's in the
> >> best interest of the millions of people, Gates included, that
> >> own Microsoft stock. Putting the screws to the very people that
> >> believed in you and invested in you is, however, truly abhorrent
> >> in comparison.
> And how abhorrent is screwing the very people that believed in you and
> invested in your products?
That's pretty abhorrent all right. Am I supposed to guess what people
you're alluding to under what circumstances ?
> Check out SGI's customer satisfaction ratings over the last decade.
> They consistently recieve high marks for product quality and support.
> (Yes, there are always a few who are unhappy, but the majority seem
> quite please, according to Datapro Information Services.
I didn't accuse SGI of screwing customers yet. That's next. They're the
last people you screw on the way down. Shareholders first, employees
second, customers last. Ok, ok, you might do the employees first by
laying off a boatload of them to convince the shareholders you're
really concerned about getting costs in line. That'll keep the
stock price up a while longer. But in this case SGI's stock cratered
BEFORE the layoff of what was it, 10% of the workforce, and the layoff
didn't bump it back up. That's REAL pessimism when a big blood letting
like that doesn't help.
> In contrast, there are millions of mindless zombies cussing their PC's
> and assuming "that's just the way things are" and they don't even
> realize how much better it could be, if their OS vendor cared about
> anything but its own profits.
I suggest in this case you check Dell's customer satisfaction rating and
keep in mind we ship nothing but Win95 and NT. Good hardware, good
drivers, good applications - no crashy, no complaint. It's really
that simple. You made an interesting career move given this FUD
you're still spouting.
> Sigh. If its not one thing its another.
Oh yes... Michael Gold knows how to service the needs of 100,000,000
PC users better than -everyone- else. Michael, you're a good
programmer but you've still got a way to go before you actually
arrive at the place you imagine yourself to be. You really don't
know how to do it better than everyone else in the world. Let's try to
stay grounded in reality here, ok ?
Everyone likes to think he knows the answers to all the worlds
problems but when push comes to shove most of us admit we really
have a hard time with just our own respecitive little corners of
it. If you could really do a better job than Bill Gates I'd expect
you'd have risen just a teensy bit farther than an individual
contributor by now, am I right ? If I knew how to do a better
job than Bill Gates I'd probably be more than an individual
contributor now too. Can't we agree that neither of us is a
Bill Gates ? My ego can stand the blow. How about yours ?
David Springer
Daniel Phillips <xy...@plugh.com> wrote:
> From CNET <the part you ignored asshole=38>:
Now -that- was funny. You're still a slimy bigot but perhaps I misjudged
you as a humorless fuck. :-)
> Silicon Graphics has managed to stave off
> class-action suits before. Earlier this year, a federal
Ah, I didn't know SGI is sued by its stockholders on a regular basis.
You taught me something today. SGI is slimier than I thought.
> >How does this news stack up with your view of SGI as the white
> >knights and Microsoft as the black knights ?
> Easy, it makes you the Black Kights chamber pot. A very useful
> function - a great source of slop to toss at the White Knight! :-)
I'll give you a 5 for creativity in the two allusions to shit and
offer a free upgrade to an 8 if you can show a solid connection
back to your usage of asshole earlier. All your talk about buttholes
and doody though is is giving me insights into your psyche that
I'll eventually use against you when I'm in a meaner mood. Fair
warning...
Daniel Phillips
j...@curved-logic.com says...
>>The problem is that your Direct3D code gets obsolete along with the card.
>Hows that? My original D3D code still works with DX5, and it will
>benefit from any hardware acceleration available.
Does your original D3D code support DrawPrimitive? ;-) Or is it still
using the old broken DX3 stuff? Get the idea?
Daniel Phillips
spri...@matrix.eden.com says...
>Now -that- was funny. You're still a slimy bigot but perhaps I misjudged
>you as a humorless fuck. :-)
That wasn't funny. Somehow you have come to believe that you can make
libels on Usenet with impunity. I'm forwarding this one to Dell, Mr.
Springer, along with a couple of clips that show you're still posting
over Dell's name. Enough is enough.
This might make you hopping mad, but consider: it's better than getting
a letter from a lawyer.
I'll give you a chance to apologize by email.
Russ Williams
Daniel Phillips wrote in message <668cqi$oc2$5...@usenet.kornet.nm.kr>...
>zsan...@best.com says...
>>
>>(Daniel Phillips) wrote:
>>> You put your finger on what's wrong with D3D. *Current* accelerators.
>>> Current accelerators become obsolete acclerators very quickly. Who
>>> wants to program a platform that's primarily designed for obsolete
>>> acceleraters?
>>
>>It's true that HW accelerators today are old meat next year this time -
>>whether they support OpenGL or D3D, or both. But they will still work.
>>I.e. don't see the issue of them becoming obsolete.
>
>better.
Yes, but OpenGL code wouldn't, would it? If you code for advanced,
unsupported feature X, then either X isn't implemented or is done in
software. You have to use only what the hardware can do, ergo
OpenGL offers no more future-proofing than D3D. The only *possible*
advantage you could derive from OpenGL is adding extra rendering
functions on a menu option. Who would go to the effort of writing extra
code and producing extra graphics in order to speculative future
hardware?
---
Russ
Russ Williams
Daniel Phillips wrote in message <668qt3$aar$2...@usenet.kornet.nm.kr>...
>spri...@matrix.eden.com says...
>>Now -that- was funny. You're still a slimy bigot but perhaps I misjudged
>>you as a humorless fuck. :-)
>
>That wasn't funny. Somehow you have come to believe that you can make
>libels on Usenet with impunity.
You can. If I call you a slimy fucking bigot and/or a humourless fuck,
who's libel laws get used? Jurisdiction, Daniel - in which country
did this alleged libel occur?
>I'm forwarding this one to Dell, Mr. Springer,
Back to your old tricks again?
>along with a couple of clips that show you're still posting over Dell's
>name. Enough is enough.
What the bits where he says "Dell PCs are good"? *That*'ll piss his
superiors off...
I think it's fairly safe to assume that David has only used a Dell .sig
for nice, safe content.
>This might make you hopping mad, but consider: it's better than getting
>a letter from a lawyer.
Saying what "The nasty man called my client bad names"? AIUI,
name-calling is merely the expression of a personal opinion. Given
that you can't prove any damages caused by such a slur, libel would
be kinda difficult to win...
---
Russ
James Shaw
<xy...@plugh.com> (Daniel Phillips) wrote:
>j...@curved-logic.com says...
>>>The problem is that your Direct3D code gets obsolete along with the card.
>>Hows that? My original D3D code still works with DX5, and it will
>>benefit from any hardware acceleration available.
>
>Does your original D3D code support DrawPrimitive? ;-)
Doesn't need to. Sure EB's were broken. DrawPrimitive is much easier
to code, but the functionality (and performance) is essentially the
same.
Jim
Wilbur Streett
David Springer <spri...@matrix.eden.com> wrote:
>How does this news stack up with your view of SGI as the white
>knights and Microsoft as the black knights ?
You must be confusing me with someone else.
>Can you point to any lawsuits where senior executives at Microsoft are accused
>of screwing the shareholders ?
Microsoft plays it's own set of games with the stock market. Check out the
$80 Million deal last year. I don't remember all the particulars, but I do
recall that Microsoft was releasing a limited stock tied to the price of
the common with a guarenteed 5% profit (or some other low number), the
proceeds of which were used to buy their own stock common stock. Made the
brokers very happy, they made their commissions. The stock price went up 5
or 6 points. Made Bill another couple Billion. Care to have an discussion
of the ethics of that move?
>I won't argue about Microsoft sometimes mercilessly putting the
>screws onto OTHER companies, that's business and it's in the
>best interest of the millions of people, Gates included, that
>own Microsoft stock. Putting the screws to the very people that
>believed in you and invested in you is, however, truly abhorrent
>in comparison.
Tell that to any company in any sort of competative market with Microsoft.
Wilbur
Wilbur Streett
David Springer <spri...@matrix.eden.com> wrote:
>Wilbur Streett <WStr...@shell.monmouth.com> wrote:
>> gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
>
>> >I think it's completely obvious why MS wrote it, it was to 'prove' that
>> >Quake could have been done with D3D and to shut up the people who
>> >claimed it was twice as slow as openGL,
>
>> Glenn.. Perhaps you forgot that it required a total rewrite of D3D in
>> order to emulate the tiny subset of OpenGL that Quake uses?
>
>That' just plain wrong Wilbur. I've got the code and AFAIK there's
>only one DX6 extension in it and that's DirectPrimitive().
OK, so the paradign shifted from EB to DrawPrimitive and then they were
able to do a Quake wrapper.. and your point is? (And aren't you under NDA
on the contents of the code? Don't want you to get into trouble defending
Microsoft.. ;-P )
>Moreover, ExecuteBuffer() could be substituted for DirectPrimitive()
>but unless one added some code to cache streams of OpenGL calls
>into larger EBs the EBs wouldn't be very efficient. It would also
>be a LOT of work getting the precedural streaming OGL calls into
>the OO execute buffers.
>Different paradigms. OpenGL is procedural
>and EBs are object oriented. I can understand why the vast majority
>of coders, with procedural 3D engines, were screaming bloody murder
>about execute buffers.
I don't disagree with your analysis, but you are mixing metaphores calling
Execute Buffers "OO". EB's were created with their own brand of
pipelining, and based on legacy design from the software rendering engine.
EB's aren't OO any more than packetizing data for communciations is.
>EBs will, IMO, undoubtedly be the highest performance method for
>hardware whose native language includes EB commands. Think of a
>higher level graphics language, with pointers, jumps, branches,
>calls, and returns along with 2D and 3D primitives for triangles,
>vectors, polys, textures and so forth. Now think of bug chunks
>of that code that make up bigger objects and parts of scenes.
>Now imagine that code sitting in system memory with a graphics
>chip that can haul it in and execute it directly over a 1.2 gigabyte
>per second AGP bus.
I think that you should investigate the OpenGL design, (Not that I'm an
OpenGL guru. Yet..) there are mechanisms for all sorts of pipelining.
Check out Triangle Strips, Vertex Arrays, Display Lists, etc.
Batching up graphics commands and data is a no brainer. Intel's now
defunct 3DR even had that.
>That's what EBs do and that's how consumer PCs are being built.
>OpenGL doesn't stand a chance at competing in that archtectural
>space. You simply can't subjugate the x86 CPU to streaming a
>zillion separate OpenGL function calls for every minor scene
>change. You instead change a few key parameters in a big EB,
>tell the graphics chip to execute it (ExecuteBuffer()), and get
>the x86 doing other useful things in the meantime, like processing
>the geometry for the next frame.
See above..
>If consumer PC hardware were being architected like SGI workstations
>it would be a different story, but they ain't. They're being built
>for low cost with geometry predominantly handled by the x86 (so
>Intel can sell ever more powerful processors) and graphics boards
>with limited on-board memory (so you can intelligently re-allocate
>system memory according to need when you aren't doing intensive 3D),
>graphics boards that are mostly just 2D and 3D rasterization engines.
The architecture of OpenGL was build on top of a variety of workstations
with a variety of architectures and has a lot of flexibility. And in case
you never heard, yesterdays's high end architecture is tomorrows mainstream
commodity technology..
>No amount of crying or handwaving is going to change this route. It's
>a done deal and a road already chosen by the forces that control
>what PC hardware looks like year over year. It's BEEN a done deal
>for quite a while now... much longer than this silly OpenGL vs.
>D3D war has been raging. Maybe you'll believe me now that Dell is
>selling exactly this architecture and it won us PC Computing Product
>of the Year at Comdex, MVP for power consumer desktop, and if not for
>the immature Riva 128 drivers a "Kick Ass" rating in Boot Magazine ?
I'm sure that Dell sells a lot of architectures, and that they'll be
supporting more and more of OpenGL in the future on their mainstream
platform.
>What more need I say or point out to convince you ? You're an
>extremly bright guy Wilbur when you don't let Wintel paranoia take
>over your higher brain functions. Stand back for a moment and consider
>this with the names Intel and Microsoft removed and see if it
>doesn't make a whole bunch of good sense.
David, I understand what you are saying, and I understand logistics and
market forces. I also understand that the reality is that I'll probably
have to work my way through all of the issues on both OpenGL and Direct3D
in order to be comfortable delivering my product in all possible target
markets.
I'm basically doing what everyone in the field is doing, writing my own
software and passing whatever parts are appropriate to whatever 3D pipeline
does exist. OpenGL is a more stable platform, so I don't have to screw
around wasting time with revisions, updates, signing up for beta lists,
while I work out the architecture of the product. I've taken Michael
Gold's RASONLY code and modified it so that I can track performance and
have checked out the entire machines performance, including the tasking
model and other issues. At end end of the product cycle I'll write an
interface to whatever version of Direct3D is guarenteed to be stable and
then release for both Direct3D and OpenGL.. but only if I'm convinced that
OpenGL isn't mainstream enough at that time. Right now it appears that
OpenGL is gaining momentum, and I may not have to do a Direct3D interface.
Intel is a part of the OpenGL ARB, (as is Microsoft), so I'd assume that
OpenGL is going to be on the WinTel platform for quite some time to come.
But then if a summer intern can port OpenGL Quake to Direct3D in a couple
weeks, then why shouldn't I use OpenGL for the time being, I'll just hire a
college kid to do the port. Or wait for you to release a better wrapper.
Wilbur
PS. I'm talking to the legal department of the company that I signed the
NDA for.
Glenn Corpes
xy...@plugh.com (Daniel Phillips) wrote:
> gle...@cix.co.uk.notthisbityoumoron says...
> >
> >>I'm not paranoid.. Microsoft doesn't change NDA's on a whim.. They
> >>are run by a Lawyer's son..
> >
> >What does that say about my last comment?, it's just a statistical
> >likelihood in a country with 5% of the worlds population and 70% of i
> t's
> >lawyers.
>
> Glenn, I genuinely believe that you are naive - you're not just cynica
> lly
> regurgitating the Microsoft spin. Gates built Microsoft mainly be bei
> ng
> extremely agressive legally. Being a lawyer's son has *everything* to
> do
> with his success. Gates has been so successful using the U.S. legal
> system to his advantage that he now thinks he's above the law - Balmer
> summed it up: "to heck with Janet Reno". Gates now believes he's more
> powerful legally than the Department of Justice.
>
> Wake up, Glenn.
I have no doubt that Microsoft are extremely aggressive legally but that
isn't the point, I suggested that Microsoft would have been better off
(in terms of PR) keeping their wrapper under NDA and letting people talk
about what they (thought they) saw at meltdown. Releasing it, as they
did through David Springer just let everyone see how unfinished and
basically crap it really was. Wilbur suggested they released it to
somehow make openGL look stupid, surely a company as scheming as you and
Wilbur say would have realised that the wrapper wasn't up to the job and
would have said, sorry Dave, no. I think David was motivated as he said,
entirely by wanting to get Quake running on their new Dimensions, how
much of an arsehole is he going to look to his boss if Quake players
don't by Dell Dimensions because he specced the wrong accelerator? (the
wrong accelerator for Quake that is, the Riva 128 is the right
accelerator for everything else).
Daniel Phillips
gle...@cix.co.uk.notthisbityoumoron says...
>I have no doubt that Microsoft are extremely aggressive legally but that
>isn't the point, I suggested that Microsoft would have been better off
>(in terms of PR) keeping their wrapper under NDA and letting people talk
>about what they (thought they) saw at meltdown.
Your right, there. That would have been a _much_ more effective sleazy
trick. Maybe they erred in playing the Springer card - thinking it was
an Ace, but later find out it was just a joker...
>Releasing it, as they
>did through David Springer just let everyone see how unfinished and
>basically crap it really was. Wilbur suggested they released it to
>somehow make openGL look stupid, surely a company as scheming as you and
>Wilbur say would have realised that the wrapper wasn't up to the job and
>would have said, sorry Dave, no. I think David was motivated as he said,
>entirely by wanting to get Quake running on their new Dimensions,
If that was his only motivation then why would he have distributed
it to everybody? Doesn't wash.
>how much of an arsehole is he going to look to his boss if Quake players
>don't by Dell Dimensions because he specced the wrong accelerator? (the
>wrong accelerator for Quake that is, the Riva 128 is the right
>accelerator for everything else).
It's hard to say. Maybe he actually talked himself into believing his
own FUD about Riva OpenGL drivers.
Jon Leech
As predicted, David Springer ignored my calling him on this 2 weeks ago
(see article <64vjq5$3t...@fido.asd.sgi.com> in r.g.p), and now continues to
demonstrate his lack of knowledge of 3D graphics hardware:
In article <6686vv$6b8$1...@boris.eden.com>,
David Springer <spri...@matrix.eden.com> wrote:
[...]
>EBs will, IMO, undoubtedly be the highest performance method for
>hardware whose native language includes EB commands. Think of a
>higher level graphics language, with pointers, jumps, branches,
>calls, and returns along with 2D and 3D primitives for triangles,
>vectors, polys, textures and so forth. Now think of bug chunks
>of that code that make up bigger objects and parts of scenes.
>Now imagine that code sitting in system memory with a graphics
>chip that can haul it in and execute it directly over a 1.2 gigabyte
>per second AGP bus.
>
>That's what EBs do and that's how consumer PCs are being built. [...]
No, they are not. Identify a single "consumer PC" that handles EBs in
the graphics card as you describe. Hint: Nvidia's Riva 128 does not, despite
your belief. How many times does this have to be pointed out before you
acknowledge it?
>OpenGL doesn't stand a chance at competing in that archtectural
>space. You simply can't subjugate the x86 CPU to streaming a
>zillion separate OpenGL function calls for every minor scene
>change.
Function call overhead is unlikely to be noticeable. If it is, there are
alternatives in OpenGL including display lists and vertex arrays of several
sorts.
>If consumer PC hardware were being architected like SGI workstations
>it would be a different story, but they ain't. They're being built
>for low cost with geometry predominantly handled by the x86 (so
>Intel can sell ever more powerful processors) and graphics boards
>with limited on-board memory (so you can intelligently re-allocate
>system memory according to need when you aren't doing intensive 3D),
>graphics boards that are mostly just 2D and 3D rasterization engines.
You do not understand how SGI workstations are architected; for a few
examples, see Mark Kilgard's Eurographics paper at
http://reality.sgi.com/mjk_asd/twoimps/twoimps.html
Those SGI machines which do have geometry acceleration - not all -
generally have very limited amounts of scratchpad memory in the geometry
engines. Display lists are typically stored in host memory and pushed or
pulled into the hardware when the list is executed.
Getting back to the main point, execute buffers CANNOT BE HANDLED by a
graphics accelerator that does not implement geometry acceleration, because
geometry processing is an intrinsic part of execution:
"Execute buffers are processed first by the transformation
module. This module runs through the vertex list, generating
transformed vertices by using the state information set up for
the transformation module. ... Then the vertices are processed
by the lighting module, which adds color to them according to
the lighting instructions in the execute buffer. Finally, the
rasterization module parses the instruction stream, rendering
primitives by using the generated vertex information."
- "Direct3D Immediate Mode Overview" from DX5 documentation
"When the IDirect3DDevice::Execute function is invoked by the
Direct3D application, Direct3D begins processing the execute buffer.
In the case of rasterization-only hardware, Direct3D will parse the
execution stream and call the two HAL callbacks RenderState and
RenderPrimitive when required."
- "Direct3D Hardware Abstraction Layer" from DX5 documentation
At such time as a future version of D3D allows hardware acceleration of
geometry processing, it will become technically feasible to do this. That
does not mean it will neccessarily be a good idea; it means it will be
possible. Today's "consumer 3D" cards do not and *cannot* process execute
buffers directly.
Give it up. You are *wrong*. Hardware that relies on the host CPU for
geometry processing, including all of today's "consumer 3D" cards, will
never be able to handle execute buffers. Microsoft's own documentation
clearly describes why. Repeating yourself over and over will not magically
change reality.
Jon Leech
Silicon Graphics
Chris Green
<xy...@plugh.com> (Daniel Phillips) writes:
>
> suma...@NOSPAM.usa.net says...
> >In fact, Carmack has said in his famous .plan that he doesn't need a
> >wide array of features -
>
> ******
> >and Direct3D provides what the current accelerators are able to do,
> >anyway.
> ******
This is false. Direct3d does not offer access to many useful
features of the 3dfx chipset, and I'm sure the same is true with some
other chipsets. This chipset is actually relatively old and very
popular, but you still can't use many features of it in d3d.
I've done lots of coding of 3dfx at the HW level, glide level,
and d3d level.
--
Chris Green
Wilbur Streett
gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
>> Wake up, Glenn.
>
>I have no doubt that Microsoft are extremely aggressive legally but that
>isn't the point, I suggested that Microsoft would have been better off
>(in terms of PR) keeping their wrapper under NDA and letting people talk
>did through David Springer just let everyone see how unfinished and
>basically crap it really was. Wilbur suggested they released it to
>somehow make openGL look stupid, surely a company as scheming as you and
>Wilbur say would have realised that the wrapper wasn't up to the job and
>would have said, sorry Dave, no. I think David was motivated as he said,
>much of an arsehole is he going to look to his boss if Quake players
>don't by Dell Dimensions because he specced the wrong accelerator? (the
>wrong accelerator for Quake that is, the Riva 128 is the right
>accelerator for everything else).
David admitted that he was releasing the driver in order to make sure that
the Dell Dimensions passed the "Run Quake" test, and would be reported as
such in the trades.
I think that you are mistaken when you think that Microsoft cares a whit
about how clean or elegant their technology is. They don't play to us,
they play to "Joe Consumer", as in the average guy that doesn't understand
what an API is, other than "It's some sort of technology thing, right?"
I doubt that the Quake wrapper is any more an abortion than what I have to
assume is the case based on the sort of behaviour that I see in run of the
mill Microsoft products. Microsoft is a marketing company, not
particularly interested in elegant technology, but much more interested in
being able to sell it.
So, in the Market oriented game, Microsoft can now claim that DirectX runs
Quake. The little bits about it not running the recently release version,
or that it doesnt' support the OpenGL drivers as blessed by the card
vendors, or that David Springer is the guy that actually got the drivers
working and supportable for the Dell Dimensions doesn't matter to Joe
Consumer. All that he knows is that Quake is "KEWL" and that he can run it
on his Dell Dimension.
Wilbur
Matt Gradwohl
Although this has nothing to do with RGP, I have another rant:
Company 1. Publicly held. Doing well.
Company 2. Publicly (or privately) held. Startup with promise. Great people,
great technology.
Company 1 calls Company 2 and says "Hey, you guys rule, can we buy you?"
Company 2 says "Yeah, that'd be cool. We wouldn't have to worry about lame
junk like copy machines, office space, etc. anymore. We just want to work on
our cool crap."
Company 1 says "Cool, here's a big pile of money. C'mon out, we've got
everything you asked for, we'll help you move, find a place, etc."
Scenario 1: Govt steps in and says "even though this is a free country, and
you both WANT to do this, we ain't gonna let you"
Scenario 2: Company 1 moves in with Company 2. Public cries "Monopoly, they
are buying everyone." Like it's any of their business what Company 2 decides.
-Matt
http://riot.lith.com/
http://www.geocities.com/SiliconValley/Pines/5444/
Matt Gradwohl
OK could the subject "Is Anything wrong with D3D"
Be any more than FlameBait?
Kent Sandvik wrote in message <3487A3...@best.com>...
>(Daniel Phillips) wrote:
>> You put your finger on what's wrong with D3D. *Current* accelerators.
>> Current accelerators become obsolete acclerators very quickly. Who
>> wants to program a platform that's primarily designed for obsolete
>> acceleraters?
>
>It's true that HW accelerators today are old meat next year this time -
>whether they support OpenGL or D3D, or both. But they will still work.
>I.e. don't see the issue of them becoming obsolete. Maybe a post PCI bus
>structure will make some of the cards obsolete, but then you just
>purchase a new sub-200$ card at Fries, and that's it. I don't think
>anyone in the computing industry any longer expects that their computers
>are state of the art more than about three months max.
>
>Cheers, Kent
Matt Gradwohl
There's no way, with any API, that your shipped compiled code can take
advantage of features and APIs that don't exist yet. When you write the code,
if you don't have the header (or in COM the tlb) then you don't know about
unknown APIs and can't use them. To be able to use them, your code would have
to read documentation for new features, no where it belongs, write the code,
compile, and test it. There's no technology that does that yet.
However, since MS has a huge install base, they guarantee that your DX3 app
will work in DX5. Actually, the COM spec guarantees that. IDirectDraw2 is
exactly the same whether it's exposed by DX3 or DX5.
So, with DirectX, the code you write using IDirectDraw2 is guaranteed to work
in DX5 (barring your hardware vendor having crappy drivers).
If you want to know how bad backwards compatibility can be, look at Glide 2.4x
and Glide 2.2x. Ouch.
Daniel Phillips wrote in message <668q8j$aar$1...@usenet.kornet.nm.kr>...
>j...@curved-logic.com says...
>>>The problem is that your Direct3D code gets obsolete along with the card.
>>Hows that? My original D3D code still works with DX5, and it will
>>benefit from any hardware acceleration available.
>
Sean Cavanaugh
Chris Green wrote:
> This is false. Direct3d does not offer access to many useful
> features of the 3dfx chipset, and I'm sure the same is true with some
> other chipsets. This chipset is actually relatively old and very
> popular, but you still can't use many features of it in d3d.
> I've done lots of coding of 3dfx at the HW level, glide level,
> and d3d level.
>
Could you please provide some specifics, I'm interested.
- Sean
Kent Sandvik
(Daniel Phillips) wrote:
> zsan...@best.com says...
> >It's true that HW accelerators today are old meat next year this time -
> >whether they support OpenGL or D3D, or both. But they will still work.
> >I.e. don't see the issue of them becoming obsolete.
> better.
A) This is pure speculation as D3D graphics acceleration cards have
barely been out on the market for less than a year.
B) Don't expect HW to stay on the same level for a long time, three
months is a very long time in this business today.
C) If someone really wants to play a game that is based on let's say
Apple II, then they keep an Apple II around.
D) I still don't see the logic why the D3D code gets obsolete along with
a card? Doesn't DX3 code still work with DX5, someone with more
knowledge, let me know? --Kent
Kent Sandvik
Matt Gradwohl wrote:
>
> There's no way, with any API, that your shipped compiled code can take
> advantage of features and APIs that don't exist yet. When you write the code,
> if you don't have the header (or in COM the tlb) then you don't know about
> unknown APIs and can't use them. To be able to use them, your code would have
> to read documentation for new features, no where it belongs, write the code,
> compile, and test it. There's no technology that does that yet.
>
> However, since MS has a huge install base, they guarantee that your DX3 app
> will work in DX5. Actually, the COM spec guarantees that. IDirectDraw2 is
> exactly the same whether it's exposed by DX3 or DX5.
>
> So, with DirectX, the code you write using IDirectDraw2 is guaranteed to work
> in DX5 (barring your hardware vendor having crappy drivers).
Yes. In addition, the HAL/HEL layer will provide compability in forms
that if a feature is missing in the 'old' driver, then this is done
inside HEL, and the future Pentium systems will be wicked fast, and the
future HW Acceleration cards even more wicked fast :-). --Kent
Kent Sandvik
David Springer wrote:
> Hmmmm... no. I have no idea how many of these frivolous lawsuits
> are filed each year. I only added SGI to stocks that I watch a few
> months ago. Is it a regular thing for SGI ? I've never seen one
> against Sun, IBM, Microsoft, Dell, Compaq, Intel, or Gateway which are
> companies I watch along with SGI.
Believe me, it's a curse, there are law firms specializing in opening
such law suits based on price drops with shares. Out-of-court
settlements are common, and thus these firms make a good living out of
it. For a while there was an attempt here in California to stop this
practice by instituting laws against unjustified such lawsuits. --Kent
Daniel Phillips
mat...@msn.com.--- says...
>However, since MS has a huge install base, they guarantee that your DX3 app
>will work in DX5.
Then why do they warn you not to use Execute Buffers??
Daniel Phillips
zsan...@best.com says...
>(Daniel Phillips) wrote:
>> zsan...@best.com says...
>> >It's true that HW accelerators today are old meat next year this time -
>> >whether they support OpenGL or D3D, or both. But they will still work.
>> >I.e. don't see the issue of them becoming obsolete.
>
>> The problem is that your Direct3D code gets obsolete along with the card.
>> I'd rather have something that automatically gets better as cards get
>> better.
>
Simple. With OpenGL it's easy to write you code to use something really
extravagant, like alpha blending throughout, or stencilled shadows, then
just flip off a few state variables and have it actually function at a
reasonable speed on a consumer card. A player can just flip those fancy
features back on in a simple configuration setup a couple of years from
now when cards that can actually accelerate those features appear. So
your OpenGL game is non-obsolescent - you have a chance at establishing
a real long runner, like Myst.
With Direct3D, you're pretty well stuck with writing to the spec it
supports today - your game is going to start looking pretty crude,
pretty fast, as hardware progresses and your game can't automatically
take advantage of it.
It's a pretty simple concept - write to a spec that was properly
designed in the first place, not one that merely promises you the
world sometime in the future.
Glenn Corpes
WStr...@shell.monmouth.com (Wilbur Streett) wrote:
> sumatose <suma...@NOSPAM.usa.net> wrote:
>
> >Wilbur Streett wrote:
> >>
> >> gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
> >>
> >> >I think it's completely obvious why MS wrote it, it was to 'prove'
> that
> >> >Quake could have been done with D3D and to shut up the people who
> >> >claimed it was twice as slow as openGL,
> >>
> >> Glenn.. Perhaps you forgot that it required a total rewrite of D3D
> in
> >> order to emulate the tiny subset of OpenGL that Quake uses?
>
> >That's completely false. Try to argue with true facts, please.
>
> Well, then by your claims, the demonstration of a Quake wrapper at the
> DX5
> release had nothing to do with any sort of reality at all..
>
> You can't have it both ways..
DX5 was planned way before anyone thought of writing a wrapper, in fact
it may have been planned before GLQuake was even up and running,
DrawPrimitive() was indeed a reaction to Carmack's (and many others)
criticism of the extremely fucked up execute buffers but it wasn't
designed to make writing a wrapper for the 'GL' game which didn't even
exist at the time. Or are you suggesting that Microsoft 'cheated' by
making D3D non shit?, I guess you are, people like you lost their main
logical argument against it when EBs were taken out of the equation.
Glenn Corpes
xy...@plugh.com (Daniel Phillips) wrote:
> suma...@NOSPAM.usa.net says...
> >In fact, Carmack has said in his famous .plan that he doesn't need a
> >wide array of features -
>
> ******
> >and Direct3D provides what the current accelerators are able to do,
> >anyway.
> ******
>
> You put your finger on what's wrong with D3D. *Current* accelerators.
> Current accelerators become obsolete acclerators very quickly. Who
> wants to program a platform that's primarily designed for obsolete
> acceleraters?
>
> Especially when there's a better platform available that was actually
> designed with a consistent rendering model... provides the basics and
> looks forward too, right from the word go. And is cross-platform too.
>
> <drifting further towards rantzone>
> Microsoft made a big mistake in spamming the game developers with D3D
> instead of comitting fully to OpenGL - it hasn't done their
> credibility a bit of good. Not too late to kiss and make up, though.
> I think that's what we're really waiting for. It takes a lot of
> maturity to admit you were wrong, and I don't think Billco has that
> maturity.
openGL with all it's guaranteed functionality and software fallbacks was
not a viable option when MS designed an API to support hardware which
often had no z-buffer and/or perspective correction, what would be the
point of an API that _always_ fell back to software?. Sure you can sit
here in late 97 with cards like Riva and the new relaxed SGI/ARB
attitude to partial drivers and argue that openGL is starting to make
sense as a game API but you can't argue that D3D never needed to exist.
Daniel Phillips
gle...@cix.co.uk.notthisbityoumoron says...
>openGL with all it's guaranteed functionality and software fallbacks was
>not a viable option when MS designed an API to support hardware which
>often had no z-buffer and/or perspective correction, what would be the
>point of an API that _always_ fell back to software?.
You're relying on a false assumption. OpenGL *can* provide hardware
acceleration in this case that the hardware has no depth buffer - you
just have to disable(DEPTH_TEST). There isn't any
disable(PERSPECTIVE_CORRECTION), as far as I can tell, but you'd also
have a pretty tough time finding a 3D card than can't do perspective
correction, today.
>Sure you can sit here in late 97 with cards like Riva and the new
>relaxed SGI/ARB attitude to partial drivers and argue that openGL is
>starting to make sense as a game API but you can't argue that D3D
>never needed to exist.
Can too. :-) Microsoft would have been a lot wiser to commit to
OpenGL right in the beginning and make the _small_ changes that
would have been needed at that time to support acceleration on all
those early, broken cards, for the modes most useful to game
programmers. As it is, M$ is fighting a losing battle against
OpenGL for games, and in the process has already lost a whole pile
of credibility with game developers.
Glenn Corpes
xy...@plugh.com (Daniel Phillips) wrote:
> gle...@cix.co.uk.notthisbityoumoron says...
> >I have no doubt that Microsoft are extremely aggressive legally but t
> hat
> >isn't the point, I suggested that Microsoft would have been better of
> f
> >(in terms of PR) keeping their wrapper under NDA and letting people t
> alk
> >about what they (thought they) saw at meltdown.
>
> Your right, there. That would have been a _much_ more effective sleaz
> y
> trick. Maybe they erred in playing the Springer card - thinking it wa
> s
> an Ace, but later find out it was just a joker...
>
Or maybe you are just a fucking paranoid twat...
Glenn Corpes
WStr...@shell.monmouth.com (Wilbur Streett) wrote:
> gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
>
> >WStr...@shell.monmouth.com (Wilbur Streett) wrote:
> >
> >> gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
> >>
> >> >I think it's completely obvious why MS wrote it, it was to 'prove'
> th
> >> at
> >> >Quake could have been done with D3D and to shut up the people who
> >> >claimed it was twice as slow as openGL,
> >>
> >> Glenn.. Perhaps you forgot that it required a total rewrite of D3D
> in
> >> order to emulate the tiny subset of OpenGL that Quake uses?
> >
> >Are you trying to suggest that DrawPrimitive was done to enable the
> >writing of the Quake wrapper?, if so it nicely illustrates how screwe
> d
> >up your mind is on this subject,
>
> Glenn, you really have to stop projecting how paranoid you think other
> people are on to them.
>
> I'm sure that Microsoft was looking for some response to the Carmark p
> lan
> comments.. or do you think that the wrapper happened "by accident?"
Of course the wrapper didn't happen by accident, and it's true that
DrawPrimitive was implimented, at least partialy because of Carmack's
criticisms, Microsoft were hoping that he would look at it, admit it was
fixed and write D3DQuake, for some reason he refused, some say because
it still sucked, others because he knew that supporting D3D would kill
GL as a game API. The reason is irrelevant to the 'wrapper' incident,
Microsoft wrote it to make the point that DX5 was capable of running
Quake. IMO it was far too bug ridden to do this on any level other than
speed, but it should have at least silenced the morons who claimed that
GL was 'twice as fast as D3D' but no..., everyone decided they would
rather ignore the issue and argue about the wording of the reporting in
a case of spin doctoring that MS themselves wouldn't even have the nerve
to try and get away with...
Glenn Corpes
spri...@matrix.eden.com (David Springer) wrote:
> Daniel Phillips <xy...@plugh.com> wrote:
>
> > The problem is that your Direct3D code gets obsolete along with the
> card.
> > I'd rather have something that automatically gets better as cards ge
> t
> > better.
>
> I don't think so. The fastest consumer hardware will support EB instr
> uction
> set as native code. It'll be a long time before that paradigm is repl
> aced.
> Therefore, as long as the EB instruction set is backward compatible wi
> th
> newer hardware,
This is complete crap David, I have never spoken to an IHV with any
intention of hardware EB support, have you recently been brainwashed by
one of the halfwits from Rendermorphics who fucked D3D with EBs in the
first place?. If EBs are the future of D3D, why are MS adding all of the
new features in to DrawPrimitive.
Glenn Corpes
xy...@plugh.com (Daniel Phillips) wrote:
> j...@curved-logic.com says...
> >>The problem is that your Direct3D code gets obsolete along with the
> card.
> >Hows that? My original D3D code still works with DX5, and it will
> >benefit from any hardware acceleration available.
>
> Does your original D3D code support DrawPrimitive? ;-) Or is it stil
> l
> using the old broken DX3 stuff? Get the idea?
That's like asking if your old 1.0 GL code supports vertex arrays or
some other feature new to 1.1.
Glenn Corpes
xy...@plugh.com (Daniel Phillips) wrote:
> mat...@msn.com.--- says...
>
> >However, since MS has a huge install base, they guarantee that your D
> X3 app
> >will work in DX5.
>
> Then why do they warn you not to use Execute Buffers??
Because EBs are very hard to work with and, on newer drivers,
DrawPrimitive is significantly faster.
Glenn Corpes
xy...@plugh.com (Daniel Phillips) wrote:
> zsan...@best.com says...
> >(Daniel Phillips) wrote:
> >> zsan...@best.com says...
> >> >It's true that HW accelerators today are old meat next year this t
> ime -
> >> >whether they support OpenGL or D3D, or both. But they will still w
> ork.
> >> >I.e. don't see the issue of them becoming obsolete.
> >
> >> The problem is that your Direct3D code gets obsolete along with the
> card.
> >> I'd rather have something that automatically gets better as cards g
> et
> >> better.
> >
> >D) I still don't see the logic why the D3D code gets obsolete along w
> ith
> >a card? Doesn't DX3 code still work with DX5...
>
> Simple. With OpenGL it's easy to write you code to use something real
> ly
> extravagant, like alpha blending throughout, or stencilled shadows, th
> en
> just flip off a few state variables and have it actually function at a
> reasonable speed on a consumer card. A player can just flip those fan
> cy
> features back on in a simple configuration setup a couple of years fro
> m
> now when cards that can actually accelerate those features appear. So
> your OpenGL game is non-obsolescent - you have a chance at establishin
> g
> a real long runner, like Myst.
>
> With Direct3D, you're pretty well stuck with writing to the spec it
> supports today - your game is going to start looking pretty crude,
> pretty fast, as hardware progresses and your game can't automatically
> take advantage of it.
>
> It's a pretty simple concept - write to a spec that was properly
> designed in the first place, not one that merely promises you the
> world sometime in the future.
If you think you can predict what hardware will actually be able to do
in the future, look at PowerVR for example, it has hardware support for
some sort of volumetric shadows and lighting which isn't supported by
D3D or openGL. Would you like to take a bet on which API supports it
first?
Minh Truong
Matt Gradwohl wrote in message ...
>Although this has nothing to do with RGP, I have another rant:
[snip]
>Scenario 1: Govt steps in and says "even though this is a free country, and
>you both WANT to do this, we ain't gonna let you"
>
>Scenario 2: Company 1 moves in with Company 2. Public cries "Monopoly, they
>are buying everyone." Like it's any of their business what Company 2
decides.
Come on, Matt. Business interests are sometimes in conflict w/ consumers'
interests. A monopoly business interest is always in conflict w/ the
consumers'.
If there were only 2 companies in your scenarios, the it's the exactly right
thing for the govt to do.
Glenn Corpes
xy...@plugh.com (Daniel Phillips) wrote:
> gle...@cix.co.uk.notthisbityoumoron says...
> >openGL with all it's guaranteed functionality and software fallbacks
> was
> >not a viable option when MS designed an API to support hardware which
> >often had no z-buffer and/or perspective correction, what would be th
> e
> >point of an API that _always_ fell back to software?.
>
> You're relying on a false assumption. OpenGL *can* provide hardware
> acceleration in this case that the hardware has no depth buffer - you
> just have to disable(DEPTH_TEST). There isn't any
> disable(PERSPECTIVE_CORRECTION), as far as I can tell, but you'd also
> have a pretty tough time finding a 3D card than can't do perspective
> correction, today.
>
> >Sure you can sit here in late 97 with cards like Riva and the new
> >relaxed SGI/ARB attitude to partial drivers and argue that openGL is
> >starting to make sense as a game API but you can't argue that D3D
> >never needed to exist.
>
> Can too. :-) Microsoft would have been a lot wiser to commit to
> OpenGL right in the beginning and make the _small_ changes that
> would have been needed at that time to support acceleration on all
> those early, broken cards, for the modes most useful to game
> programmers.
But the ARB wouldn't have allowed those changes, I think they only allow
them now because GL is too few letters to protect legally, remember that
openGL has to pass conformance tests.
> As it is, M$ is fighting a losing battle against
> OpenGL for games,
Continuing to improve D3D isn't fighting against openGL.
> and in the process has already lost a whole pile
> of credibility with game developers.
There you go again, overestimating the importance of this newsgroup and
therefore yourself. Try subscribing to the DirectX and openGL mailing
lists and compare the questions and sigs of real game developers with
better things to do than read all of the crap here.
Paul Miller
David Springer wrote:
> I think you guys should check out the class action lawsuits filed
> by two major stock brokerage firms against SGI. It seems there's
And how does that have ANYTHING to do with graphics technology and games
programming? I personally think Microsoft has them easily beat with this
whole Department of Justice thing.
--
Paul Miller | st...@fxtech.com
Paul Miller
Russ Williams wrote:
> >I'm forwarding this one to Dell, Mr. Springer,
>
> Back to your old tricks again?
I've noticed an interesting thing about you Russ - you never post except
to defend David and/or Microsoft and/or D3D. Have anything original to
contribute about GAME PROGRAMMING perhaps?
--
Paul Miller | st...@fxtech.com
Jon Leech
In article <memo.19971206...@frog.compulink.co.uk>,
Glenn Corpes <gle...@cix.co.uk.notthisbityoumoron> wrote:
>openGL with all it's guaranteed functionality and software fallbacks was
>not a viable option when MS designed an API to support hardware which
>often had no z-buffer and/or perspective correction, what would be the
>point of an API that _always_ fell back to software?.
An intelligent OpenGL implementation on such hardware wouldn't fall back
to SW if you disabled depth testing and perspective correction.
Jon Leech
Silicon Graphics
Daniel Phillips
gle...@cix.co.uk.notthisbityoumoron says...
>>>isn't the point, I suggested that Microsoft would have been better off
>>>(in terms of PR) keeping their wrapper under NDA and letting people talk
>>>about what they (thought they) saw at meltdown.
>>
>>Your right, there. That would have been a _much_ more effective sleazy
>>trick. Maybe they erred in playing the Springer card - thinking it was
>>an Ace, but later find out it was just a joker...
>
>Or maybe you are just a fucking paranoid twat...
Oh, Glenn, you didn't think it was funny - grab a prozac and calm
down. Don't worry about me hurting Springer's feelings - he loves this
kind of abuse.
Daniel Phillips
gle...@cix.co.uk.notthisbityoumoron says...
>
>Of course the wrapper didn't happen by accident, and it's true that
>DrawPrimitive was implimented, at least partialy because of Carmack's
>criticisms, Microsoft were hoping that he would look at it, admit it was
>fixed and write D3DQuake, for some reason he refused, some say because
>it still sucked, others because he knew that supporting D3D would kill
>GL as a game API. The reason is irrelevant to the 'wrapper' incident,
>Microsoft wrote it to make the point that DX5 was capable of running
>Quake. IMO it was far too bug ridden to do this on any level other than
>speed, but it should have at least silenced the morons who claimed that
>GL was 'twice as fast as D3D' but no..., everyone decided they would
>rather ignore the issue and argue about the wording of the reporting in
>a case of spin doctoring that MS themselves wouldn't even have the nerve
>to try and get away with...
What you're looking at, Glenn is a whole industry that's increasingly
pissed at Microsoft. Don't cry for Microsoft, Glenn - they've dished
it out a lot more than they've taken it. That whole wrapper thing
had the classic Microsoft odor of rotting fish - otherwise, the
source code would have simply been released to everybody, instead of
playing all those _wierd_ NDA games.
Daniel Phillips
>>>>The problem is that your Direct3D code gets obsolete along with the
>>card.
>>>benefit from any hardware acceleration available.
>>
>>Does your original D3D code support DrawPrimitive? ;-) Or is it still
>>using the old broken DX3 stuff? Get the idea?
>
>That's like asking if your old 1.0 GL code supports vertex arrays or
>some other feature new to 1.1.
That's weak. OpenGL had a powerful, general API right from the start.
Russ Williams
Daniel Phillips wrote in message <66c9c7$30m$3...@usenet.kornet.nm.kr>...
>gle...@cix.co.uk.notthisbityoumoron says...
>>>Does your original D3D code support DrawPrimitive? ;-) Or is it still
>>>using the old broken DX3 stuff? Get the idea?
>>
>>That's like asking if your old 1.0 GL code supports vertex arrays or
>>some other feature new to 1.1.
>
>That's weak. OpenGL had a powerful, general API right from the start.
Sure. But are you trying to say a dozen function calls per poly is efficent?
Vertex arrays are a big improvement in OpenGL 1.1.
GL 1.1 is better than 1.0, DX5 is better than DX3. It's called progress.
---
Russ
Russ Williams
Paul Miller wrote in message <34899E39...@fxtech.com>...
>Russ Williams wrote:
>
>> >I'm forwarding this one to Dell, Mr. Springer,
>>
>> Back to your old tricks again?
>
>I've noticed an interesting thing about you Russ - you never post except
>to defend David and/or Microsoft and/or D3D. Have anything original to
>contribute about GAME PROGRAMMING perhaps?
Try reading some of my posts. Like, maybe the ones covering sprite
graphics, texture mapping, various DOS coding things. Dejanews might
be helpful.
---
Russ
Daniel Phillips
>>OpenGL for games,
>
>Continuing to improve D3D isn't fighting against openGL.
Continuing to fail to release the DirectDraw bindings _is_. There
were two things the game developer comunittee pleaded with Microsoft
for: 1) Release the MCD driver model 2) release the DirectDraw
bindings. Microsoft's response to this was a big fat zero. I call
that fighting against OpenGL - what do you call it?
Paul Miller
Russ Williams wrote:
> >I've noticed an interesting thing about you Russ - you never post except
> >to defend David and/or Microsoft and/or D3D. Have anything original to
> >contribute about GAME PROGRAMMING perhaps?
> Try reading some of my posts. Like, maybe the ones covering sprite
> graphics, texture mapping, various DOS coding things. Dejanews might
> be helpful.
You're right - after I posted that I saw a bunch of your more
informative posts. I apologize.
--
Paul Miller | st...@fxtech.com
Russ Williams
Paul Miller wrote in message <3489BF57...@fxtech.com>...
Don't worry about it. I keep feeling I'm posting too little content, as
well...
---
Russ
Kent Sandvik
(Daniel Phillips) wrote:
> Simple. With OpenGL it's easy to write you code to use something really
> extravagant, like alpha blending throughout, or stencilled shadows, then
> just flip off a few state variables and have it actually function at a
> reasonable speed on a consumer card. A player can just flip those fancy
> features back on in a simple configuration setup a couple of years from
> now when cards that can actually accelerate those features appear. So
> your OpenGL game is non-obsolescent - you have a chance at establishing
> a real long runner, like Myst.
>
> With Direct3D, you're pretty well stuck with writing to the spec it
> supports today - your game is going to start looking pretty crude,
> pretty fast, as hardware progresses and your game can't automatically
> take advantage of it.
I'm not that good of a marketdroid, just a software engineer, and bad in
this domain as well. So I can't really compete against such discussions
as defined above, all I could use is my simple notion of not talking
about the futures before the future is here. It's plain speculation,
used in combintation with hype, FUD, waporware and similar marketdroid
discussion topics I'm not good with.
I do like OpenGL, but in all fairness it has technologies based on the
notion of a client-server graphics pipeline, and most home systems are
not based on this concept. So there are certainly designs in every API
that assumes something, and misses the boat. And I'm sure that's the
case with D3D as well. But it's better to stick to shipping products and
hardware. So I'm still wondering if DX3 applications break with DX5. If
this is not the case, then your current hypothesis can't be justified.
If you could show it indeed, then you are of course right. --Kent
Wilbur Streett
gle...@cix.co.uk.notthisbityoumoron (Glenn Corpes) wrote:
>> Well, then by your claims, the demonstration of a Quake wrapper at the
>> DX5
>> release had nothing to do with any sort of reality at all..
>>
>> You can't have it both ways..
>
>DX5 was planned way before anyone thought of writing a wrapper, in fact
>it may have been planned before GLQuake was even up and running,
>DrawPrimitive() was indeed a reaction to Carmack's (and many others)
>criticism of the extremely fucked up execute buffers but it wasn't
>designed to make writing a wrapper for the 'GL' game which didn't even
>exist at the time. Or are you suggesting that Microsoft 'cheated' by
>making D3D non shit?, I guess you are, people like you lost their main
>logical argument against it when EBs were taken out of the equation.
Oh come on Glen. I never said that Microsoft "cheated". I didn't even
really have a problem with the EB architecture. I read the documents and
figured out what was going on in one pass. That doesn't mean that I think
that the design of D3D was very good.. and with the other games that were
being played.. OpenGL is just a simpler answer at this time. Check out the
lengthy message that I wrote to David about it.
Wilbur
Paul Shirley
In article <66bcha$cmp$1...@usenet.kornet.nm.kr>, xy...@plugh.com writes
>You're relying on a false assumption. OpenGL *can* provide hardware
>acceleration in this case that the hardware has no depth buffer - you
>just have to disable(DEPTH_TEST). There isn't any
>disable(PERSPECTIVE_CORRECTION), as far as I can tell, but you'd also
>have a pretty tough time finding a 3D card than can't do perspective
>correction, today.
glHint(GL_PERSPECTIVE_CORRECTION_HINT,GL_FASTEST);
...will disable perspective correction IF that would run faster. So on a
decent accelerator it would automagically stay perspective correct.
Typically you would try this during calibration to detect incompatible
settings. (In fact on a broken piece of hardware the default setting is
likely to disable optional features in favour of speed anyway, you would
need glHint(GL_PERSPECTIVE_CORRECTION_HINT,GL_NICEST); to guarantee
perspective correction)
I really wish you would open an OpenGL manual before making assertions
about what it can or cannot do ;)
---
Paul Shirley: my email address is 'obvious'ly anti-spammed
David Springer
Glenn Corpes <gle...@cix.co.uk.notthisbityoumoron> wrote:
> This is complete crap David, I have never spoken to an IHV with any
> intention of hardware EB support, have you recently been brainwashed by
Just because no one told you they were planning hardware EB support
doesn't mean no one is planning to do it.
Are you contesting the argument that hardware EB support would be
faster than a stream of DrawPrimitive calls ?
If you aren't contesting that argument why do you suppose that no
hardware vendors would want to take advantage of the higher
performance ?
I think we first need to slow down and see if we agree on the
performance advantages of EB hardware support. If we agree on that
then it's just a short hop to assuming that IHVs have also realized
the benefit and at least one of them is acting on it.
> one of the halfwits from Rendermorphics who fucked D3D with EBs in the
> first place?. If EBs are the future of D3D, why are MS adding all of the
> new features in to DrawPrimitive.
I don't have my "DirectX 5 Foundations" CD with me but if you have
it you can find on it a brief discussion of why DrawPrimitive was
added. Paraphrasing, it reiterates that EBs are the way to go
for object oriented programming and DrawPrimitive is the way to
go if you don't want to take the time to learn and use the newer
programming paradigm.
OpenGL supports this through "display lists" from what I understand.
David Springer
--
*************** IGAMES INTERNET GAME LOBBY ****************
* *
* NOW SUPPORTING MICROSOFT DirectPlay 3 LOBBY STANDARD ! *
* *
* A real-time game lobby for the internet with many *
* exciting games and thousands of players. Game *
* developers, players, and ISP's can try it out at: *
* *
****************** http://www.igames.com ******************
David Springer
> >j...@curved-logic.com says...
> >>>The problem is that your Direct3D code gets obsolete along with the card.
> >>Hows that? My original D3D code still works with DX5, and it will
> >>benefit from any hardware acceleration available.
> >
> >Does your original D3D code support DrawPrimitive? ;-)
> Doesn't need to. Sure EB's were broken. DrawPrimitive is much easier
> to code, but the functionality (and performance) is essentially the
> same.
EBs aren't "broken". They are merely a programming paradigm that is
strange to developers who've never used high end graphics subsystems
with a native instruction set. The vast majority of game programmers
are only familiar with manipulating bit mapped display devices.
The performance is essentially the same because everyone using EBs
were wrapping them so they could be used in place of the missing
DrawPrimitive. In order to use EBs effectively you have to architect
your code around the complex objects you can build with an EB. This
is not convenient for existing game engines where you just want to
replace your CPU intensive routines that rasterize lit triangles
with something faster. Sure, that takes you up to the next level
in performance, an order of magnitude more than before, but where
is the next bottleneck and how do you address it to gain that next
level of performance increase ?
Consider the Riva 128 benchmarks in 3D Winbench 97. They are
entirely CPU bound with an AGP card. I ran the benchmarks with
PII speeds beginning at 200mhz and the scores continue to get
better in a linear ramp with CPU speed. The conclusion of that
experiment is that current generation CPUs can't pump lit
triangles out to it as fast as the Riva can rasterize them.
The bottleneck is no longer in rasterizastion but rather in
the delivery speed.
One way to optimize delivery speed is
through EBs that can be directly executed from system memory
through a bus mastering AGP chip. Another way is to move
geometry processing over to the grahics chip. However, in the
low end, I think you'll be hard pressed to beat a 300 to 500
megahertz Pentium II with a $200 graphics card when it comes
to geometry. However, it adds little expense or complexity
to give the AGP graphics card the ability to directly execute
an EB residing in system memory.
In the consumer space the
answer that gives the most bang-for-the-buck is usually the
one that wins. If it requires more effort from developers
to use it, there will be resistance, but invariably at least
one developer will go the extra mile, and then competitive
pressure will force the rest of them to do it.
If it were workstations with a $20K+ pricetag I'd say sure,
throw in $2K+ graphics card that beats the snot out of a
Pentium II for geometry processing. It doesn't even have to
beat it by a lot to justify the price tag. Workstations
are running mission critical engineering applications so
it doesn't take a whole lot of saved hours from an expensive
engineer to justify the price.