Carmack on OpenGL vs D3D
Thomas B Valesky
Check out John Carmack's latest finger info (finger jo...@idsoftware.com).
He comes out pretty strongly against D3D and in favor of OpenGL.
--
===========================================================================
Tom Valesky -- This account is closing soon. I'm getting
an ISP account set up. In the interim, mail to tval...@hqlink.cbsi.com.
Frank Pitt
In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
>
>Check out John Carmack's latest finger info (finger jo...@idsoftware.com).
>He comes out pretty strongly against D3D and in favor of OpenGL.
Yeah, but anyone with any sense would.
The point is not whether OpenGL is better, it obviously is,
but whether it will ever get marketshare.
Remember what happened to Beta in relation to VHS.
Frankie
Thomas B Valesky
Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
: In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
: >
: >Check out John Carmack's latest finger info (finger jo...@idsoftware.com).
: >He comes out pretty strongly against D3D and in favor of OpenGL.
: Yeah, but anyone with any sense would.
No argument from this quarter. :-)
: The point is not whether OpenGL is better, it obviously is,
: but whether it will ever get marketshare.
Well, if OpenGL has a friend in Id Software, it might make some card
manufacturers sit up and take notice. Microsoft doesn't win _all_
the time. :-)
David Springer
Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
: In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
: >
: >Check out John Carmack's latest finger info (finger jo...@idsoftware.com).
: >He comes out pretty strongly against D3D and in favor of OpenGL.
: Yeah, but anyone with any sense would.
: The point is not whether OpenGL is better, it obviously is,
: but whether it will ever get marketshare.
: Remember what happened to Beta in relation to VHS.
Exactly right - other examples:
OS/2 and Win 3.x
PGA and VGA
Microchannel and ISA
68K and 80x86
Apple Mac and IBM PC
Amiga and IBM PC
Dvorac and Qwerty
Trackballs and Pointsticks
Trackballs and Touchpads
All the items on the left have a clear technologic edge over
the items on the right but other factors made the technologically
inferior product the big winner in the marketplace.
What appeals to engineers, and we are all fundamentally engineers
in this group, is not necessarily what appeals to the world's
marketing forces and the bottom line is that a product that can't
be sold, no matter how good it is, will fade into obscurity.
David Springer
--
********* WEB GAMES ONLINE ENTERTAINMENT SYSTEM ***********
* *
* A client/server system designed from the ground up to *
* support real time multi-player games on the Internet. *
* Game developers, players, and ISP's check it out at: *
* *
************** http://www.eden.com/~springer **************
David Springer
Thomas B Valesky (tval...@osf1.gmu.edu) wrote:
: Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
: : In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
: Well, if OpenGL has a friend in Id Software, it might make some card
: manufacturers sit up and take notice. Microsoft doesn't win _all_
: the time. :-)
This is the mistake that many of you are making. What makes card
manufacturers sit up and take notice is multimillion dollar purchase
orders from Dell, IBM, Compaq, Gateway, Toshiba, NEC, Packard Bell,
DEC, and HP. Id doesn't buy millions of graphics cards and Id
doesn't make Compaq sit up and take notice. Microsoft makes Compaq
sit up and take notice. If OpenGL is to come out on top something
has to make Microsoft sit up and take notice. Id doesn't have the
influence to do that. SGI might have the influence if they tried
(read spent) hard enough and wanted it bad enough but I don't believe
the upper echelons at SGI want it that bad as they aren't primarily
a Wintel camp follower to begin with.
Stephen Random Wilkinson
On 27 Dec 1996 16:33:25 GMT, tval...@osf1.gmu.edu (Thomas B Valesky)
wrote:
>Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
>: In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
[snip]
>: The point is not whether OpenGL is better, it obviously is,
>: but whether it will ever get marketshare.
>
>Well, if OpenGL has a friend in Id Software, it might make some card
>manufacturers sit up and take notice. Microsoft doesn't win _all_
>the time. :-)
If as many folks as have spoken out about D3D IM would actually band
together and take their issues up w/Microsoft maybe there would be
some hope. I am not looking forward to using D3D (I've already been
spoiled by OpenGL and Glide) and am _really_ not looking forward to
dealing w/D3D as Microsoft incrementally "fixes" it. I'd rather be a
volunteer practice patient at a dental school :)
Someone or some group of people need to feel strongly enough about
this to get together and attempt to change the direction of the MS
efforts. I kinda like the "Direct-GL" idea from Carmack. It would be
best to have some video card manufacturers on-board too. Without
accellerated drivers, the OpenGL issue will go nowhere.
Cheers,
Stephen Wilkinson "Programming is like pinball.
Software Engineer the reward for doing it well
Interactive Creations Inc. is the opportunity to do it
wi...@airmail.net again." (anon - Wizards Bane)
RNDM in Warbirds http://www.icigames.com
Glenn Corpes
spri...@eden.com (David Springer) wrote:
> Thomas B Valesky (tval...@osf1.gmu.edu) wrote:
> : Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
> : : In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
>
> : Well, if OpenGL has a friend in Id Software, it might make some card
> : manufacturers sit up and take notice. Microsoft doesn't win _all_
> : the time. :-)
>
> This is the mistake that many of you are making. What makes card
> manufacturers sit up and take notice is multimillion dollar purchase
> orders from Dell, IBM, Compaq, Gateway, Toshiba, NEC, Packard Bell,
> DEC, and HP. Id doesn't buy millions of graphics cards and Id
I wonder how many people bought a verite based card because it had a
native quake port?
> doesn't make Compaq sit up and take notice. Microsoft makes Compaq
> sit up and take notice. If OpenGL is to come out on top something
> has to make Microsoft sit up and take notice. Id doesn't have the
> influence to do that.
Hopefully you _may_ be wrong here, while MS where originaly evangelising
DirectX (or the games SDK as it was at the time), they made a big deal of
how id were involved: 'we have been working with Id, Origin, etc...' and
the fact that Doom95 would be one of the first games to use it, the
dropping of the D3D version of Quake _must_ be a big deal for them,
hopefully it will push them into taking openGL more seriously. Personally
my attitude has shifted slightly, i'm going to get someone to seriously
look into the feasibility of openGL as soon as possible.
> SGI might have the influence if they tried
> (read spent) hard enough and wanted it bad enough but I don't believe
> the upper echelons at SGI want it that bad as they aren't primarily
> a Wintel camp follower to begin with.
>
> David Springer
Before everyone gets on the standard 'Flame David' bandwagon it's worth
remembering that he has an almost unique perspective on the relationship
in this newsgroup between MS and the hardware guys.
- - --=-==< gco...@ea.com, Head of R&D Bullfrog Productions >==-=-- - -
Terry Sikes
In article <59rf0v$r...@portal.gmu.edu>,
Thomas B Valesky <tval...@osf1.gmu.edu> wrote:
>
>Check out John Carmack's latest finger info (finger jo...@idsoftware.com).
>He comes out pretty strongly against D3D and in favor of OpenGL.
I'm going to repost this in this thread, since the first one I posted
in may be in a lot of killfiles by now. Sorry for the duplication,
please respond here.
In article <memo.9612271...@frog.compulink.co.uk>,
Glenn Corpes <gle...@cix.co.uk> wrote:
>tsi...@netcom.com (Terry Sikes) wrote:
>
>> I suggest you check out John Carmack's .plan file by using:
>> finger car...@idsoftware.com
>>
>> (Wsfinger95 is a nice Win95 finger utility if you happen to be running
>> in that environment).
>>
>> The first paragraph should read:
>>
>> ---begin quote---
>> I am going to use this installment of my .plan file to get up on a
>> soapbox about an important issue to me: 3D API. I get asked for my
>> opinions about this often enough that it is time I just made a public
>> statement. So here it is, my current position as of december '96...
>> ---end quote---
>>
>> If this .plan goes away soon and there's enough interest, I'd be happy
>> to repost it in full. Since its quite long, I won't do that now.
>>
>> To get to the meat of the matter, Carmack feels that D3D is seriously
>> broken, and he's going to do OpenGL instead. He feels that the game
>> developer community can and should influence the chip and card
>> manufacturers to support OpenGL in addition to D3D (and eventually
>> instead of D3D).
>>
>> I'm a senior software engineer and game developer, and I like what he
>> has to say. I'm going to send E-mail to every 3D hardware company I
>> can find requesting Win32 OpenGL drivers. Not only is OpenGL a better
>> designed, more extensible API than Direct3D, it will allow portability
>> to many other platforms including Apple, Be and OS/2. Also your
>> skills will be portable to other areas besides game development. I'd
>> encourage everyone, both at a personal and company level, to pursue
>> this.
>>
>> I hope this was of interest...
>It is interesting but it doesn't really tell me anything I didn't know.
>D3D _is_ broken but it's slowly, card by card, being fixed, the issue for
>me is how do we support 3d hardware?, i'd love to be able to switch to
>openGL but I have no idea what happens at the other end when a customer
>pulls his new Bullfrog game out of the box?, where does he get the
>drivers?
The idea with this is that we, the game developers and companies, tell
the hardware (chip and board) vendors what we'd like to see. It makes
complete sense to not waste all the power of say a Rendition card on
only doing games anyhow. With OpenGL CAD etc. would also be supported
(not to mention 3D Studio Max ;-).
Supporting OpenGL is likely much easier than Direct3D, since it is
well documented, there are reference drivers available, and
conformance tests.
Also, I think Carmack makes a good case for Direct3D being broken at
the API and extensibility level compared with OpenGL. Those aren't
things that can be fixed "card by card".
>, openGL is not being pushed like D3D is. I'll say it again: i'd
>love to switch to openGL but I don't think it's possible *yet*, as usual
>as the discussion is getting religious and people are starting to confuse
>what they want to be true with what they know is true.
Microsoft is intent on supporting OpenGL. They are pushing Direct3D
for games AT THIS TIME. If it becomes clear to them that Direct3D is
broken and the majority of the development community wants OpenGL,
they will change their tune. It really shouldn't matter much to them.
>Game developers think about this: just suppose that the openGL speed
>problems are solved (or the companies making hardware that doesn't fit
>openGL too well somehow, conveniently never sell a card and go bust) and
I'm curious on this point. What kinds of optimizations (or the
opposite ;-) make a card work well under Direct3D and poorly under
OpenGL? Would anyone care to comment on which of the low-cost 3D
chipsets would be likely to work better with Direct3D?
>MS admit defeat and make openGL the standard API, and somebody (MS maybe?)
>gathers together openGL drivers that game developers can freely distribute
>on their CDs (directGL? :-)),
OpenGL support is already in Windows NT, and slated to be in Windows
97. Part of the standard installation, no less. This is really not a
matter of Microsoft admitting defeat, but rather recognizing that
Direct3D doesn't bring much to the table compared with THEIR other
technology.
>I truly hope this will happen but the
>question is when?, 12 months?, 18 months?, two years?, how do we get
>hardware support for all those 3d blasters, matrox mystiques and 3dfx
>cards that people have _now_ without direct 3d?.
From what I understand, several of the card manufacturers are working
on OpenGL drivers right now. Since the current Direct3D
implementation is barely useable on several cards, there really isn't
much of a difference is there?
I'll contact several of the board and chip vendors, take a quick
survey on planned OpenGL support, and summarize the results here
sometime next week.
BTW, in case this subject somehow comes up, I have absolutely no
connection to SGI, and I'm as old as dirt. ;-)
--
Terry Sikes | Software Developer
tsi...@netcom.com | C++, Delphi, Java, Win32 (in alphabetical order ;)
finger for PGP pub key | Objective objects objectify objectivity.
My opinions - mine only! | http://members.aol.com/tsikes
Terry Sikes
In article <851657591...@mundens.welly.gen.nz>,
Frank Pitt <fra...@mundens.welly.gen.nz> wrote:
>In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
>>
>>He comes out pretty strongly against D3D and in favor of OpenGL.
>
>
>but whether it will ever get marketshare.
>
Not a good analogy. OpenGL is a much more widely supported technology
than Direct3D, and Microsoft is already supporting OpenGL.
Actually, perhaps its a good analogy in the other direction, except
that OpenGL is the better technlogy. ;-)
David Springer
Glenn Corpes (gle...@cix.co.uk) wrote:
: spri...@eden.com (David Springer) wrote:
: > Thomas B Valesky (tval...@osf1.gmu.edu) wrote:
: > : Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
: > : : In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
: I wonder how many people bought a verite based card because it had a
: native quake port?
I'm just guessing but I doubt it's more than a tiny fraction of
the number of people who bought a certain graphics card because
Dell or Compaq chose it to bundle into their machines based on
overall excellence.
A good question though since only a tiny fraction of those same
people have played Quake.
I think the bottom line still stands though... a chip vendor is
concerned about a design win that guarantees hundreds of thousands
or millions of chip sales. You can't do that based on one game
from Id that happens to play a little better on your chip. You
CAN do that by having outstanding Windows graphics benchmarks,
a good price, and a track record of reliable delivery of those
kinds of quantity.
: > doesn't make Compaq sit up and take notice. Microsoft makes Compaq
: > sit up and take notice. If OpenGL is to come out on top something
: > has to make Microsoft sit up and take notice. Id doesn't have the
: > influence to do that.
: Hopefully you _may_ be wrong here, while MS where originaly evangelising
: DirectX (or the games SDK as it was at the time), they made a big deal of
: how id were involved: 'we have been working with Id, Origin, etc...' and
: the fact that Doom95 would be one of the first games to use it, the
: dropping of the D3D version of Quake _must_ be a big deal for them,
: hopefully it will push them into taking openGL more seriously. Personally
: my attitude has shifted slightly, i'm going to get someone to seriously
: look into the feasibility of openGL as soon as possible.
By all means never stop evaluating your options. MS is going to say,
within the bounds of truthfulness one hopes, selected pieces of truth
that will help make the sale. They didn't have quite the stranglehold
on the PC design & distribution channels a year ago that they do today
either. No D3D version of Quake will certainly be a new wrinkle but
I don't seriously think it'll be more than a wrinkle at this stage.
If they back away from D3D now they'll sour a lot of people that
trusted them a year ago. Better to push ahead and deliver on that
promise even if it's late. If they got away with Win 95 being so
late I'm sure that covering for D3D is no great hurdle... look how much
practice they have.
: > SGI might have the influence if they tried
: > (read spent) hard enough and wanted it bad enough but I don't believe
: > the upper echelons at SGI want it that bad as they aren't primarily
: > a Wintel camp follower to begin with.
: >
: > David Springer
: Before everyone gets on the standard 'Flame David' bandwagon it's worth
: remembering that he has an almost unique perspective on the relationship
: in this newsgroup between MS and the hardware guys.
Flame David is a "standard" bandwagon ? I'm flattered... but really,
I haven't had a really good flaming since I was beating y'all about the
head & shoulders saying Win 95 would be a better game platform than
DOS what - 18 months ago I started saying that ? Man, I sure took
some lumps then...
I don't think I said _when_ it would be a better platform. I know
better than to try to predict how long it'll take Microsoft to debug
something before they shove it down our throats - or should I say
shove it down our throats and THEN debug it ?
All I knew was that the games market was pretty much the last software
frontier where they hadn't reached in grabbed control so it's only
a matter of time - they have the money. It was the last because
in the big picture games are still pretty small potatoes but I had
confidence that greed would win out in the end and they'd want it all.
Also I think entertainment software is due for another renaissance
what with the coming together of digital networks and broadcast
entertainment and if I see it coming so does Bill Gates... his
plan to rule the digital world therefore must include this industry.
And much as I hate to admit it I think he's unstoppable. Let's just
hope it's a benevolent dictatorship...
David
---
David Springer
Terry Sikes (tsi...@netcom.com) wrote:
: In article <59rf0v$r...@portal.gmu.edu>,
: Thomas B Valesky <tval...@osf1.gmu.edu> wrote:
: Microsoft is intent on supporting OpenGL. They are pushing Direct3D
: for games AT THIS TIME. If it becomes clear to them that Direct3D is
: broken and the majority of the development community wants OpenGL,
: they will change their tune. It really shouldn't matter much to them.
One might think it doesn't matter at first blush but unless you have
all the information, which I'm sure NONE of us here do, you don't know
what all of their motivations are.
: OpenGL support is already in Windows NT, and slated to be in Windows
: 97. Part of the standard installation, no less. This is really not a
: matter of Microsoft admitting defeat, but rather recognizing that
: Direct3D doesn't bring much to the table compared with THEIR other
: technology.
It may seem obvious to some but I keep coming back to the point where
I'm wondering what we don't know about the D3D/OpenGL decision that
caused MS to go the way they did. There's no new information been
revealed here except for the unsurprising bit about the current D3D
implimentation being somewhat less than robust. I don't believe
anyone at MS has been "just plain stupid or disconnected from the
industry"... they hire and retain too many good people for that to
happen. There's a reason that's not being made known.
: From what I understand, several of the card manufacturers are working
: on OpenGL drivers right now. Since the current Direct3D
: implementation is barely useable on several cards, there really isn't
: much of a difference is there?
: I'll contact several of the board and chip vendors, take a quick
: survey on planned OpenGL support, and summarize the results here
: sometime next week.
I'll be glad to help out with any of the vendors that Dell uses.
: BTW, in case this subject somehow comes up, I have absolutely no
: connection to SGI, and I'm as old as dirt. ;-)
Well, if you're old you're probably too set in your ways and that's
not a good thing either... however, no connection to SGI is
looking like a plus to me at this point. ;)
Believe it or not I have absolutely no preference either way which
3D API comes out on top as long as ONE OF THEM does by this summer
when I plan on picking something.
David Springer
Robert Rodgers
spri...@eden.com (David Springer) wrote:
>It may seem obvious to some but I keep coming back to the point where
>I'm wondering what we don't know about the D3D/OpenGL decision that
>caused MS to go the way they did.
I thought that was obvious.
The reference implementation's performance sucks and, as good as MS's
software implementation *is*, it's performance still sucks for SW only
-- it just does not compare to D3d running in ramp. It was even worse
back when MS bought Rendermorphics.
It has taken SGI a long, long time to start work on a version of
OpenGL that has decent *software* performance and they still don't
have it out yet, while MS wanted to *ship* their final version early
this year at the latest (and when they bought RM, pre-GSDK era, they
probably intended to ship it even earlier than that -- we're probably
looking at a decision that was made more than two years ago, and a lot
of the IM problems are the result of adapting to com/directx and the
lack of a hardware architecture plans in the original product).
Their reasoning was "no time to rewrite and tune OpenGL." They were
probably right at the time.
They aren't right anymore and everyone is legitimately hoping they'll
add "DirectGL" to DirectX. OpenGL has better docs, a better API, it's
easier to accellerate, programming manuals are available that are top
notch, it's portable &c&c.
Portable -- actually, is there an implementation of OpenGL for the
Mac? The Mac is the main port target for PC games.
[...]
> There's a reason that's not being made known.
What's the reason for the use of FAT in Win95 and FAT32 (an extension
to an extension to a really bad filesystem)?
A lot of what MS does is the result of inertia.
When they bought RM, OpenGL's performance sucked and it was not clear
at all what their game development strategy would look like. Today
it's a lot clearer and they can see that many developers who *really
crave* a standard are still choosing to write HW-dependent code or are
bitching about the bugs/performance of Direct3d.
A lot depends on how SGI handles the source code to Cosmo OpenGL. If
they make it available and MS can bundle it as a patch and/or with the
OS without negotiating a huge license fee, we're a lot more likely to
see MS release a DirectGL.
[...]
>Believe it or not I have absolutely no preference either way which
>3D API comes out on top as long as ONE OF THEM does by this summer
>when I plan on picking something.
You should. Not only is OpenGL desirable because it's a better API,
but by pushing for OpenGL we'd see better support for OpenGL across
the board, making cards for high-end (non-game) use substantially
cheaper in the long run. This would be a boon to the platform as
Direct3d will **never** be adopted by the 3d apps vendors.
rsr
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
words were just sounds -- genes made them bludgeons
Glenn Corpes
spri...@eden.com (David Springer) wrote:
> Thomas B Valesky (tval...@osf1.gmu.edu) wrote:
> : Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
> : : In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
>
> : Well, if OpenGL has a friend in Id Software, it might make some card
> : manufacturers sit up and take notice. Microsoft doesn't win _all_
> : the time. :-)
>
> This is the mistake that many of you are making. What makes card
> manufacturers sit up and take notice is multimillion dollar purchase
> orders from Dell, IBM, Compaq, Gateway, Toshiba, NEC, Packard Bell,
> DEC, and HP. Id doesn't buy millions of graphics cards and Id
I wonder how many people bought a verite based card because it had a
native quake port?
> doesn't make Compaq sit up and take notice. Microsoft makes Compaq
> sit up and take notice. If OpenGL is to come out on top something
> has to make Microsoft sit up and take notice. Id doesn't have the
> influence to do that.
Hopefully you _may_ be wrong here, while MS where originaly evangelising
DirectX (or the games SDK as it was at the time), they made a big deal of
how id were involved: 'we have been working with Id, Origin, etc...' and
the fact that Doom95 would be one of the first games to use it, the
dropping of the D3D version of Quake _must_ be a big deal for them,
hopefully it will push them into taking openGL more seriously. Personally
my attitude has shifted slightly, i'm going to get someone to seriously
look into the feasibility of openGL as soon as possible.
> SGI might have the influence if they tried
> (read spent) hard enough and wanted it bad enough but I don't believe
> the upper echelons at SGI want it that bad as they aren't primarily
> a Wintel camp follower to begin with.
>
> David Springer
Before everyone gets on the standard 'Flame David' bandwagon it's worth
remembering that he has an almost unique perspective on the relationship
in this newsgroup between MS and the hardware guys.
- - --=-==< gco...@ea.com, Head of R&D Bullfrog Productions >==-=-- - -
Terry Sikes
In article <851657591...@mundens.welly.gen.nz>,
Frank Pitt <fra...@mundens.welly.gen.nz> wrote:
>In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
>>
Terry Sikes
>He comes out pretty strongly against D3D and in favor of OpenGL.
I'm going to repost this in this thread, since the first one I posted
Microsoft is intent on supporting OpenGL. They are pushing Direct3D
for games AT THIS TIME. If it becomes clear to them that Direct3D is
broken and the majority of the development community wants OpenGL,
they will change their tune. It really shouldn't matter much to them.
>Game developers think about this: just suppose that the openGL speed
>problems are solved (or the companies making hardware that doesn't fit
>openGL too well somehow, conveniently never sell a card and go bust) and
I'm curious on this point. What kinds of optimizations (or the
opposite ;-) make a card work well under Direct3D and poorly under
OpenGL? Would anyone care to comment on which of the low-cost 3D
chipsets would be likely to work better with Direct3D?
>MS admit defeat and make openGL the standard API, and somebody (MS maybe?)
>gathers together openGL drivers that game developers can freely distribute
>on their CDs (directGL? :-)),
OpenGL support is already in Windows NT, and slated to be in Windows
97. Part of the standard installation, no less. This is really not a
matter of Microsoft admitting defeat, but rather recognizing that
Direct3D doesn't bring much to the table compared with THEIR other
technology.
>I truly hope this will happen but the
>question is when?, 12 months?, 18 months?, two years?, how do we get
>hardware support for all those 3d blasters, matrox mystiques and 3dfx
>cards that people have _now_ without direct 3d?.
From what I understand, several of the card manufacturers are working
on OpenGL drivers right now. Since the current Direct3D
implementation is barely useable on several cards, there really isn't
much of a difference is there?
I'll contact several of the board and chip vendors, take a quick
survey on planned OpenGL support, and summarize the results here
sometime next week.
BTW, in case this subject somehow comes up, I have absolutely no
connection to SGI, and I'm as old as dirt. ;-)
David Springer
[snip]
: > There's a reason that's not being made known.
[snip]
: A lot depends on how SGI handles the source code to Cosmo OpenGL. If
: they make it available and MS can bundle it as a patch and/or with the
: OS without negotiating a huge license fee, we're a lot more likely to
: see MS release a DirectGL.
Bingo. My #1 suspicion. Do you know what's going on in closed
door meetings between SGI and Microsoft ?
A rhetorical question. I don't know either. I don't think anyone
posting here knows and I suspect that if we did know there wouldn't
be so much wonderment about MS continuing to push D3D.
: You should. Not only is OpenGL desirable because it's a better API,
: but by pushing for OpenGL we'd see better support for OpenGL across
: the board, making cards for high-end (non-game) use substantially
: cheaper in the long run. This would be a boon to the platform as
: Direct3d will **never** be adopted by the 3d apps vendors.
I don't necessarily agree. In my experience there have always been
tradeoffs in speed/flexibility/accuracy when selecting algorithms
to perform specific jobs. Are you saying there's no benefit to
tuning one 3D API for flawless precision at CAD/CAM and other
serious applications where mistakes might mean cutting a millimeter
too deep into someone's brain because the model wasn't *quite*
true to scale or a $100M satellite's antenna failing to unfold because
the model wasn't *quite* perfect --- and tuning another API to
sacrifice that kind of accuracy to get a big increase in rendering
speed ?
Another example - I'm evaluating voice codecs to add real-time
voice chat to my internet game framework. I can get a 2400 bps
codec that will work fine with one speaker at a time, no music
or anything else, just voice... or I can use one that has no
limitation but won't work in real time... or I can use one that
will work in real time but the data rate is too high for people
using modems on the internet. The last thing I want is to have
just one codec that does a mediocre job at all the above instead
of excellence in a more narrowly focused job.
David
Brian Hook
David Springer wrote:
> It may seem obvious to some but I keep coming back to the point where
> I'm wondering what we don't know about the D3D/OpenGL decision that
> caused MS to go the way they did. There's no new information been
> revealed here except for the unsurprising bit about the current D3D
> implimentation being somewhat less than robust. I don't believe
> anyone at MS has been "just plain stupid or disconnected from the
> industry"... they hire and retain too many good people for that to
> happen. There's a reason that's not being made known.
What we don't know about is the internal politics that are going at
Microsoft, which are pretty extensive. Keep in mind that Microsoft is a
huge organization, and while it has lots of good people, it also has a
lot of political animals who aren't really into their job but are more
interested in climbing up the corporate ladder. So it's entirely
conceivable that some people like this are/were associated with the DX
project and bungled it.
I could easily imagine someone being involved in DX and thinking "You
know, we could easily do our own 3D API and not have to share control
with those pukes in the OpenGL group".
The rumors of the OpenGL vs. D3D rivalry internal to Microsoft are
legendary.
> Believe it or not I have absolutely no preference either way which
> 3D API comes out on top as long as ONE OF THEM does by this summer
> when I plan on picking something.
This is the EXACT same attitude many game developers had UNTIL they
started using Direct3D. If you recall, Michael Abrash said quite
clearly at last year's CGDC that id would ship a Direct3D version of
Quake. Neither he nor John Carmack had really worked with D3D at that
point -- they just assumed that it would be bad. But it couldn't be
THAT bad.
They, and many other developers, were wrong. *I* fell into the same
trap. I had heard stories about how bad D3D was, but I figured it was
mostly the whining of putzes who couldn't code. I mean, Win32 wasn't
THAT bad, and everyone said Win32 was horrible, so how bad could D3D
be? Pretty bad, as I found out. Inability to work with D3D has little
to do with how good you are. If APIs aren't documented, don't work, or
aren't implemented (there are several VERY important pieces of D3D that
aren't implemented right now, including execute buffer validation), then
the programmer simply has to program empirically -- muck with things
until they work. Which is a VERY bad way of programming practice,
because they may hit upon some undefined or undocumented behaviour that
HAPPENS to work right now, but will later quit working.
This is why a SOLID specification is a must have for an API to really be
useful.
I've seen, time and time again, developers commit to D3D without ever
having seen it or written to it, and really end up hating life
afterwards.
Brian
--
+-----------------------------------------------------------------+
+ Brian Hook, b...@wksoftware.com +
+ WK Software, http://www.wksoftware.com +
+ Consultants specializing in 3D graphics hardware and software +
+ --------------------------------------------------------------- +
+ For a list of publications on 3D graphics programming, +
+ including Direct3D, OpenGL, and book reviews: +
+ http://www.wksoftware.com/publications.html +
+-----------------------------------------------------------------+
Brian Hook
Robert Rodgers wrote:
> The reference implementation's performance sucks and, as good as MS's
> software implementation *is*, it's performance still sucks for SW only
> -- it just does not compare to D3d running in ramp. It was even worse
> back when MS bought Rendermorphics.
Well, that's not exactly fair -- comparing ramp performance vs. a pure
RGB mode? What about comparing OpenGL vs. the D3D RGB software driver?
> Their reasoning was "no time to rewrite and tune OpenGL." They were
> probably right at the time.
As it turns, they were WRONG, because now MS OpenGL's performance is
very close to D3D's performance, and in some cases faster. And this is
a VERY small group working on OpenGL.
> Portable -- actually, is there an implementation of OpenGL for the
> Mac? The Mac is the main port target for PC games.
Yes, there is, but it's not part of the MacOS distribution.
Brian Hook
Glenn Corpes wrote:
> Before everyone gets on the standard 'Flame David' bandwagon it's worth
> remembering that he has an almost unique perspective on the relationship
> in this newsgroup between MS and the hardware guys.
He has A perspective, but some of us work quite routinely VERY closely
with a lot of 3D hardware vendors, so I don't think his perspective is
necessarily very accurate. While he claims he's under blanket NDAs, he
IS a laptop BIOS programmer, which means he probably (I'm sure he'll
correct me if I'm wrong) doesn't get to many meetings involving 3D
acceleration and Microsoft and all the various semiconductor companies.
Mark Kilgard
In article <851657591...@mundens.welly.gen.nz>,
fra...@mundens.welly.gen.nz (Frank Pitt) writes:
|> In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
|> >
|> >He comes out pretty strongly against D3D and in favor of OpenGL.
|>
|>
|> The point is not whether OpenGL is better, it obviously is,
|> but whether it will ever get marketshare.
OpenGL is ahead in market share. I assure you that the total
dollar value of 3D hardware and software for OpenGL sold is
significantly higher than that spent on Direct3D hardware and
software. The only thing Direct3D has ever been ahead in is
hype share.
|> Remember what happened to Beta in relation to VHS.
Yes, Beta companies switched to VHS. Just like previously
Direct3D oriented hardware developers are now refocusing
their efforts on OpenGL. (I do admit that many of these
3D hardware vendors are not admitting this change publically
to avoid killing whatever demand exists for their current
generation 3D cards. In private, most will admit that
they've been royally hosed by Direct3D.)
- Mark
Mark Kilgard
In article <5a0tol$7...@portal.gmu.edu>, tval...@osf1.gmu.edu (Thomas B Valesky)
writes:
|> Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
|> : In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
|> : >
|> : >Check out John Carmack's latest finger info (finger
|> jo...@idsoftware.com).
|> : >He comes out pretty strongly against D3D and in favor of OpenGL.
|>
|> : Yeah, but anyone with any sense would.
|>
|>
|> : but whether it will ever get marketshare.
|>
|> manufacturers sit up and take notice. Microsoft doesn't win _all_
|> the time. :-)
Look, Microsoft (as a whole, obviously not the small Direct3D organization
within Microsoft) definitely wins when OpenGL comes to dominate the PC 3D
market. Realize that Microsoft has been supporting OpenGL for the last
five years. I'd guess that Microsoft's Softimage organization which
depends totally on OpenGL for its 3D modeling software products is both
bigger than the Direct3D organization and brings in substantially more
revnue. Direct3D is actually a pretty small operation within Microsoft
that simply managed to generate large amount sof hype due to those within
Microsoft whose personal ambitions within the company depended on making
the Rendermorphics merger look like it made sense.
- Mark
Mark Kilgard
In article <5a18d3$4ig$1...@boris.eden.com>, spri...@eden.com (David Springer)
writes:
|> Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
|> : In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
|> : >
|> : >Check out John Carmack's latest finger info (finger
|> jo...@idsoftware.com).
|> : >He comes out pretty strongly against D3D and in favor of OpenGL.
|>
|> : Yeah, but anyone with any sense would.
|>
|> : but whether it will ever get marketshare.
|>
|>
|> Exactly right - other examples:
|>
|> OS/2 and Win 3.x
|> PGA and VGA
|> Microchannel and ISA
|> 68K and 80x86
|> Apple Mac and IBM PC
|> Amiga and IBM PC
|> Dvorac and Qwerty
|> Trackballs and Pointsticks
|> Trackballs and Touchpads
|>
|> All the items on the left have a clear technologic edge over
|> the items on the right but other factors made the technologically
|> inferior product the big winner in the marketplace.
I agree with David here. The issue is clearly not simply technical
superiority. While OpenGL's technical superiority is really unquestioned
by everyone that knows the issues (definitely read John Carmack's .plan
if you need to understand the technical issues), OpenGL is ALSO a superior
product in ALL ways compared to Direct3D. Here's what I mean:
o Vendor support:
OpenGL: Microsoft, Intel, SGI, DEC, IBM, Integraph, Evans &
Sutherland, HP, Sun, 3Dlabs, & most PC graphics card vendors.
Direct3D: Microsoft, & most PC graphics card vendors.
o Publically available implementation:
OpenGL: Yes (Mesa 2.1, full source code, portable to all 32-bit
computers with a C compiler).
Direct3D: No (never will be).
o OS support:
OpenGL: Windows 95, Windows NT, X11/Unix, Linux, BeOS, MacOS, NextStep.
Direct3D: Windows 95 (NT support in beta).
o Conformance suite:
OpenGL: Yes, all OpenGL implementations MUST pass.
Direct3D: No.
o Proven extension mechanism:
OpenGL: Yes, OpenGL 1.1 and scores of extensions designed and implemented.
Direct3D: No.
o Maturity:
OpenGL: Over 5 years old; OpenGL 1.1 compatibly updated API.
Direct3D: Under 1 year old; updates have historically broken compatibility.
o Acceptance in university-level teaching of 3D graphics courses:
OpenGL: Extremely widespread; excellent OpenGL-based textbook exists.
Direct3D: No.
o Image processing support:
OpenGL: fully functional pixel path combined with texture mapping
allows powerful pixel manipulation and image warping; extensions
support convolution and histograms.
Direct3D: None.
o Maximum demonstrated performance:
OpenGL: 11+ million polygons/second, 800 million pixels updated/second
(SGI's InfiniteReality OpenGL implementation).
Direct3D: Well under 1 million polygons/second, under 100 million pixels
updated/second.
o Demonstrated application domain:
OpenGL: ALL interactive 3D application domains.
Direct3D: Games, demos, 3D viewers.
o Books:
OpenGL: 7 (Programming Guide, Reference Manual, OpenGL for Windows,
OpenGL for X, textbook, Super Bible, ...)
Direct3D: 2 (neither have usable coverage of the Direct3D immediate
mode API).
o Web available information:
OpenGL: Several thousand pages.
Direct3D: Around a hundred (no detailed info).
o True open standard (public specification, controlled by a vendor-neutral
standards body):
OpenGL: Yes.
Direct3D: No.
Honestly, anyone believing that Direct3D is "ahead" in any area but hype
needs to look at the facts.
- Mark
Mark Kilgard
In article <5a19ld$4qi$1...@boris.eden.com>, spri...@eden.com (David Springer)
writes:
|> : : In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
|>
|> : manufacturers sit up and take notice. Microsoft doesn't win _all_
|> : the time. :-)
|>
|> manufacturers sit up and take notice is multimillion dollar purchase
|> orders from Dell, IBM, Compaq, Gateway, Toshiba, NEC, Packard Bell,
|> DEC, and HP. Id doesn't buy millions of graphics cards and Id
|> sit up and take notice. If OpenGL is to come out on top something
What you fail to understand is that OpenGL *is* a Microsoft endorsed
standard and has been for the last five years. Microsoft has taken
notice for quite a long time. Microsoft's current OpenGL 1.1
implementation is probably the best 3D software renderer in existance
and it works on ALL PC cards and OpenGL 1.1 has hardware driver models
for both Windows 95 and NT. Microsoft's Softimage organization
relies totally on OpenGL for its products. Said another way, OpenGL
support generates real revnue for Microsoft; Direct3D doesn't.
|> Id doesn't have the
|> influence to do that.
You'd be surprised how much influence Id has.
Also, when Andy Grove talks about truly fast 3D PC graphics at Comdex
that he shows OpenGL based demos. As for Compaq, you might be surprised
how much interest they are showing about OpenGL.
- Mark
Mark Kilgard
In article <32c6759a...@news.cris.com>, wi...@airmail.net (Stephen "Random"
Wilkinson) writes:
|> On 27 Dec 1996 16:33:25 GMT, tval...@osf1.gmu.edu (Thomas B Valesky) wrote:
|> >Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
|> >: In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
|> >: The point is not whether OpenGL is better, it obviously is,
|> >: but whether it will ever get marketshare.
|> >
|> >manufacturers sit up and take notice. Microsoft doesn't win _all_
|> >the time. :-)
|>
|> together and take their issues up w/Microsoft maybe there would be
|> some hope. I am not looking forward to using D3D (I've already been
|> spoiled by OpenGL and Glide) and am _really_ not looking forward to
|> dealing w/D3D as Microsoft incrementally "fixes" it. I'd rather be a
|> volunteer practice patient at a dental school :)
|> Someone or some group of people need to feel strongly enough about
|> this to get together and attempt to change the direction of the MS
|> efforts.
This has been done on numerous occasions. Microsoft is well aware
how lousy Direct3D is. Id Software's position is just a higher
profile, public version of what Microsoft has understood for a long time.
There are those at Microsoft who really understand the technical issues
involved with 3D and understand OpenGL to be superior and they are
working to make OpenGL succeed as they have been doing for the last five
years. OpenGL for Windows 95 and NT is more stable, better documented,
and generally faster (despite the hype) than Direct3D. OpenGL also has
a substantially wider application domain (not just games). I am happy to
see that the one previous major problem for OpenGL on Windows 95 (the lack
of an easy hardware driver model for Windows 95) is basically resolved now.
Then, there are those fairly "non-3D savvy" political types at
Microsoft who have garnered a high-profile through the aquistition
of Rendermorphics (their RealityLab API was the basis of Direct3D).
They did what they knew how to do (generate hype). They are now reaping
what they have sown. While the pro-Direct3D types within Microsoft
did a splendid hype job, they can't hype over lousy technology.
- Mark
Mark Kilgard
In article <5a4boa$cs5$1...@boris.eden.com>, spri...@eden.com (David Springer)
writes:
|>
|> I'm wondering what we don't know about the D3D/OpenGL decision that
|> caused MS to go the way they did. There's no new information been
|> revealed here except for the unsurprising bit about the current D3D
|> implimentation being somewhat less than robust.
John Carmack's .plan analysis in particular made it quite clear
that it is not simply the implementation of Direct3D that he
objected to. It was the actual interface itself. He didn't just
say that he wasn't interested in the current implementation of
Direct3D, but that he wasn't even interested in watching Microsoft
try to gradually evolve what they have into something better
(assuming that's even worthwhile).
Here's the pertinent .plan quote:
"Direct-3D IM is a horribly broken API. It inflicts great pain and
suffering on the programmers using it, without returning any
significant advantages. I don't think there is ANY market segment
that D3D is apropriate for, OpenGL seems to work just fine for
everything from quake to softimage. There is no good technical reason
for the existance of D3D."
The problem is not simply the implementation of Direct3D, but the
actual programming interface itself AND its general inappropriateness
for any market segment.
|> Believe it or not I have absolutely no preference either way which
|> 3D API comes out on top as long as ONE OF THEM does by this summer
|> when I plan on picking something.
I think that's a very wise position. You are right that there
should only be one 3D API. I hope you'll make that desire clear to
Microsoft.
You are also wise to want to consider the thinking of a well-respected,
experienced, and unbiased expert on the issue (obviously I'm not talking
about myself here). I hope that you feel that the Carmack .plan is that
kind of analysis.
- Mark
Robert Rodgers
m...@woodsy.asd.sgi.com (Mark Kilgard) wrote:
>OpenGL is ahead in market share. I assure you that the total
>dollar value of 3D hardware and software for OpenGL sold is
>significantly higher than that spent on Direct3D hardware and
What does the sale of high-end hardware have to do with games?
Planning on limiting your market to people who bought $2k video cards
or SGI boxes, Mark?
>software. The only thing Direct3D has ever been ahead in is
>hype share.
Today: Direct3d's performance as a software renderer is better.
Direct3d's hardware support is better.
We'll see what Cosmo OpenGL looks like when it ships, but for now,
OpenGL is unusable for games. The hottest 3d cards on the market
don't support it at all (3dfx) and *do* support Direct3d.
The politicization of this issue is ridiculous.
Carmack made two very strong points: Direct3d IM is hard to program
and OpenGL is easy to program, Direct3d lacks conformance tests (and
let me add that driver support varies so much that it's a pain in the
ass).
That's it. He said nothing about the current market situation because
the current market situation is *what gave us Direct3d in the first
place*.
If people want to lean on someone, lean on MS to give us DirecgOpenGL
-- and lean on SGI to release Cosmo OpenGL.
rsr b a l a n c e
- - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - - -
please discontinue use of r...@msn.com
David Springer
Mark Kilgard (m...@woodsy.asd.sgi.com) wrote:
: In article <5a19ld$4qi$1...@boris.eden.com>, spri...@eden.com (David Springer)
: writes:
: |> Thomas B Valesky (tval...@osf1.gmu.edu) wrote:
: |> : Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
: |> : : In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu writes:
: What you fail to understand is that OpenGL *is* a Microsoft endorsed
: standard and has been for the last five years. Microsoft has taken
With all due respect here Mark, OpenGL is NOT endorsed by Microsoft
for *entertainment* applications and if that weren't the case we wouldn't
be having this conversation.
: notice for quite a long time. Microsoft's current OpenGL 1.1
: implementation is probably the best 3D software renderer in existance
: and it works on ALL PC cards and OpenGL 1.1 has hardware driver models
: for both Windows 95 and NT. Microsoft's Softimage organization
: relies totally on OpenGL for its products. Said another way, OpenGL
: support generates real revnue for Microsoft; Direct3D doesn't.
I've no doubt whatsoever that OpenGL is successful, profitable, and
well established for what it was designed for - precision 3D. I
used to own a graphics board company and we had one of the fastest
2D vector engines around in the late 80's. We supported AutoDesk
products. Using this as an analogous situation I don't think anyone
at AutoDesk is considering D3D for inclusion in a CAD/CAM package
but OpenGL would be quite suitable (if it isn't already being used
by AutoDesk).
: |> Id doesn't have the
: |> influence to do that.
: You'd be surprised how much influence Id has.
If John Carmack has personal sit-down meetings with Bill Gates
and Andy Groves like Michael Dell does then yes I would be
surprised.
I wouldn't be surprised if someone from Id managed to meet
and talk to Groves or Gates briefly at a Comdex party but
sit down with them at a pre-arranged meeting to talk about
strategic relationships between Microsoft, Intel, and Id ?
I don't think so but I've been wrong before.
Still think I'd be surprised ?
: Also, when Andy Grove talks about truly fast 3D PC graphics at Comdex
: that he shows OpenGL based demos. As for Compaq, you might be surprised
: how much interest they are showing about OpenGL.
SGI is another story. I believe that the CEO of SGI could arrange
and get a meeting with either Groves or Gates and probably already
has.
David
Javier Arevalo
gle...@cix.co.uk (Glenn Corpes) wrote:
>Before everyone gets on the standard 'Flame David' bandwagon it's worth
>remembering that he has an almost unique perspective on the relationship
>in this newsgroup between MS and the hardware guys.
I'd add to this that, even if David Springer's posts are not always
nice to read (i.e. are offensive), he often comes up with ideas and
insights about the whole business. Many other posts are quite polite
and nice but don't add anything to our bags. It may or may not be an
excuse, but I'd rather get angry at any of David's posts, than feel
relaxed at 90% of the rest.
Javier Arevalo
Arvirago Entertainment
(but my opinions are mine)
David Springer
Mark Kilgard (m...@woodsy.asd.sgi.com) wrote:
: In article <5a4boa$cs5$1...@boris.eden.com>, spri...@eden.com (David Springer)
: writes:
: |> Believe it or not I have absolutely no preference either way which
: |> 3D API comes out on top as long as ONE OF THEM does by this summer
: |> when I plan on picking something.
: I think that's a very wise position. You are right that there
: should only be one 3D API. I hope you'll make that desire clear to
: Microsoft.
Only one 3D API targeted towards *games*. I think there should be
two API's if their missions are significantly different. My
experience tells me that precision CAD/CAM applications and
entertainment applications take different approaches regards
tradeoffs in speed, accurcy, and ease of use. I'm not convinced that
a single 3D API will address both with the same efficiency.
: You are also wise to want to consider the thinking of a well-respected,
: experienced, and unbiased expert on the issue (obviously I'm not talking
: about myself here). I hope that you feel that the Carmack .plan is that
: kind of analysis.
I consider *everyone's* opinions. There are a number of people
who've come forward and said that while D3D isn't a bowl of
cherries to work with they *are* working with it and planned to
continue on that course.
If everyone were of exactly the same opinion and as vehemently
opposed to D3D as Carmack then I'd have no doubt that D3D is
well and truly broken. However, some people disagree, if not
in principle then at least in degree and I don't ignore those
other voices anymore than I ignore yours, Brian Hook's, or
John Carmack's.
Personally I'd rather see OpenGL come out on top if it's the
best choice and if you do a DejaVu search on my writing in
this group looking for references to OpenGL you'll find that
18 months ago I was arguing FOR OpenGL's inclusion in Win 95
and AGAINST some Rendermorphics derivative.
More to the point you'll find that it was Brian Hook and/or
Kendall Bennet from SciTech that I was arguing with about it.
At that point in time I thought that surely OpenGL would be
included in the initial release of Win 95 as it made too
much sense not to pursue that course. To my chagrin, that
was one argument Brian ended up winning. My apology to
Brian if it was Kendall Bennet I'm thinking of, they were
both arguing with me about video driver and API issues in
Win 95 at that time.
Lastly, to re-use one of Brian or Kendall's points from 18
months ago - Rendermorphics Reality Lab was the engine of
choice among many game developers then and apparently continues
to be acceptable to at least some of them in its D3D incarnation
now. Please forgive me if I've mixed up any names - brand
names of products I've never used don't stick well.
Many voices - lots of history - no concensus. I'm remaining
neutral until it's sorted out. All I want is a clean set of
hardware device drivers for whatever API eventually comes out
on top for entertainment applications under Win 95. Since
I'm starting fresh with whatever API it happens to be I
don't have any religious attachment to a specific one and
have no doubt I'll manage to learn how to effectively use
whichever one happens to win this war.
Look on the bright side - at least it's down to 2 contenders.
A year ago at this time I was looking closely at Intel's
3DR package and that didn't seem so bad either but what
do I know ? They're all the same to me at this point.
David Springer
--
************ IGAMES ONLINE ENTERTAINMENT SYSTEM ***********
* *
* A client/server system designed from the ground up to *
* support real time multi-player games on the Internet. *
* Game developers, players, and ISP's check it out at: *
* *
***************** http://www.igames.com *******************
Dale Pontius
In article <32c7c80...@news.wam.umd.edu>,
rsro...@wam.umd.edu (Robert Rodgers) writes:
>
> Today: Direct3d's performance as a software renderer is better.
> Direct3d's hardware support is better.
>
obviously where the games are. But over a broad spectrum the OpenGL
hardware support is better. -->There are many people in the world
and in the industry who know how to do fast OpenGL hardware.<--
That level of maturity just isn't there, yet, for D3D. At ANY price.
The only think keeping the low cost market away from OpenGL is the
lack of drivers. There is no technical reason why it can't be done.
There is no compelling reason whatsoever for D3D other than the
Microsoft name.
> We'll see what Cosmo OpenGL looks like when it ships, but for now,
> OpenGL is unusable for games. The hottest 3d cards on the market
> don't support it at all (3dfx) and *do* support Direct3d.
>
NO the HOTTEST cards on the market are OpenGL. It just happens that
they're too hot for my wallet. Your's too, even though SGI probably
feels that they have reasonable prices. When you said HOT you forgot
to specify $$$.
Dale Pontius
(NOT speaking for IBM)
Dale Pontius
> manufacturers sit up and take notice. Microsoft doesn't win _all_
> the time. :-)
>
need some 'friendly convincing' of which department needs more
funding. Perhaps they could be pursuaded to move D3D Retained Mode
over on top of OpenGL, and keep D3D IM for 'compatibility'.
Dale Pontius
In article <5a18d3$4ig$1...@boris.eden.com>,
spri...@eden.com (David Springer) writes:
>
> All the items on the left have a clear technologic edge over
> the items on the right but other factors made the technologically
> inferior product the big winner in the marketplace.
>
> in this group, is not necessarily what appeals to the world's
> marketing forces and the bottom line is that a product that can't
> be sold, no matter how good it is, will fade into obscurity.
>
And let's keep making inferior choices, we're all so GOOD at it.
Look at the sick state of the world today and see how WONDERFUL
making expedient, short-term choices has made it.
Yeah, it's programming, not solving world hunger. But it's our
little corner of the world. Isn't it time to do something RIGHT?
Robert Rodgers
spri...@eden.com (David Springer) wrote:
>Only one 3D API targeted towards *games*. I think there should be
>two API's if their missions are significantly different.
I'm just curious, but *why*?
The basics of Direct3d are the same as those for OpenGL. Since the
actual API content (as opposed to the interface) is the same, it would
be silly to argue that Direct3d should exist because it makes sense to
have an API for games and an API for high end rendering* as the two
have very different requirements and goals.
Some of what *is* missing in OpenGL is also missing in Direct3d.
Heighfields, true primities, &c&c&c and other game-specific tasks are
not easily possible in either, so presumably a true Games 3d API would
offer this, and other tools. But Direct3d doesn't.
Instead of two APIs, it makes a lot more sense to have *if necessary*
two different rendering engines that cater to the assumptions of the
kinds of programmers. But the interface to the different models
(e.g., support for colored lighting in the "slow" renderer) need not
be different (and indeed, if you look at D3d, *is not* different).
*(Actually, it would be nice if, in addition to the shaded renderers
like OpenGL, the systems offered a standard API for high-end rendering
into which third party renderers could be plugged in. A standard API
for high-quality [by which I mean actual rendering of the
scanline/rt/rad variety] rendering would be a real boon. Of course,
it should be renderman, but wont be.)
Robert Rodgers
pon...@btv.ibm.com (Dale Pontius) wrote:
>In article <32c7c80...@news.wam.umd.edu>,
> rsro...@wam.umd.edu (Robert Rodgers) writes:
>> Today: Direct3d's performance as a software renderer is better.
>> Direct3d's hardware support is better.
>>
>Both points are debatable. You're focusing at the low end,
Worse, I'm focusing on the PC low-end, and the Windows PC low-end, at
that (though the alternaties aren't much better: OpenGL for OS/2 is
even less adequate than MS's implementation, and most of the third
parties don't yet have HW support on 95),
But, as you say, this is -
> which is
>obviously where the games are.
[...]
>The only think keeping the low cost market away from OpenGL is the
>lack of drivers. There is no technical reason why it can't be done.
>There is no compelling reason whatsoever for D3D other than the
>Microsoft name.
The other reason is the lack of integration with DirectDraw. This is
why it's important for MS to realize that they need to put work into
D3d and that arguments that D3d is crap are missing the point: d3d is
orthogonal to getting support for DirectDraw into OpenGL for Win95/NT.
>> We'll see what Cosmo OpenGL looks like when it ships, but for now,
>> OpenGL is unusable for games. The hottest 3d cards on the market
>> don't support it at all (3dfx) and *do* support Direct3d.
>>
>NO the HOTTEST cards on the market are OpenGL.
The first line in my message made the domain of which I was speaking
quite plain. Since I'm clearly speaking of games, PC platform, this
last bit is clear flamebait.
Gavin Bell
Terry Sikes wrote:
> I'm curious on this point. What kinds of optimizations (or the
> opposite ;-) make a card work well under Direct3D and poorly under
> OpenGL? Would anyone care to comment on which of the low-cost 3D
> chipsets would be likely to work better with Direct3D?
Boards that don't have hardware z-buffers might work better. Microsoft
is warning developers to be SURE to call the BeginScene/EndScene D3D-IM
API calls (even though they don't do anything on a lot of current
boards), I assume because boards that sort or split the polygons in each
frame will be coming out.
You could extend OpenGL with similar calls, but OpenGL programs tend to
do things that would be likely to break a polygon-sorting
implementation. Like turning on/off the z-buffer to do 2D/3D layering.
Or drawing all transparent polygons "last" in rough back-to-front order
with z-buffer compare turned on but z-buffer update turned off (with
alpha-blending). Or drawing certain objects multiple times with
alpha-blending (and the zbuffer test set to GL_EQUAL).
You can do all that with D3D... but there's no guarantee that those
features will be supported by any given D3D hardware. Given the
difficulty of supporting "D3DRENDERSTATE_ZFUNC" on non-zbuffered
hardware, and assuming that boards without z-buffers will be appearing
(why else would Microsoft tell developers to be sure to call
BeginScene/EndScene?), you can be pretty certain that there WILL be
hardware that won't support that API call. So you'd better not use it
in your game if you care about your game working on all the 3D boards
out there.
I believe that one of OpenGL's greatest strengths is its clean,
orthogonal design that treats polygons and images as both first-class
citizens. Being able to assume that you can do things like put multiple
textures onto a polygon with alpha-blending, or control the z-buffer
test allow you to do amazing, devious, wicked-cool effects. With OpenGL
you have no guarantees that your game will be fast enough if you use
these features, but that's true of D3D, too-- neither have minimum
performance requirements. I think that one of the great weaknesses of
D3D is that it will force game developers to create games that all
pretty much look the same, because the lowest common denonimator of
features that you can reasonably expect to work on all implementations
is way too low.
Disclaimer: I used to work for SGI and have used OpenGL or IRIS GL for
close to ten years now. My only experience with Direct3D is reading the
Microsoft documentation, thinking about how I'd implement something like
Open Inventor or a game using it, and keeping track of the Direct3D vs
RenderWare vs OpenGL vs BRender vs Pex debates that regularly occur.
--
Gavin Bell -- ga...@acm.org -- http://www.mailbag.com/users/gavin
Work: (608) 231-0282 Home: (608) 238-3974
VRML info: http://www.sdsc.edu/vrml http://vag.vrml.org/
Paul Shirley
In article <5a85h1$d...@fido.asd.sgi.com>, Mark Kilgard
<m...@woodsy.asd.sgi.com> writes
> Realize that Microsoft has been supporting OpenGL for the last
>five years. I'd guess that Microsoft's Softimage organization which
>depends totally on OpenGL for its 3D modeling software products is both
>bigger than the Direct3D organization and brings in substantially more
>revnue.
Its probably worth reminding you that Microsoft *bought* Softimage (and
not that many years ago - 3 sounds about right)
I'm not convinced M$ had (or have) any interest in OpenGL beyond
migrating Softimage to Windoze. I somehow doubt there is any strategic
planning behind any of this (or we would be using DirectGL now ;) M$ are
just doing there normal 'buy everything | everyone and hope some of it
sticks' routine.
---
Paul Shirley: shuffle chocolat before foobar for my real email address
Stephen Random Wilkinson
On 30 Dec 1996 10:56:37 GMT, m...@woodsy.asd.sgi.com (Mark Kilgard)
wrote:
>In article <5a18d3$4ig$1...@boris.eden.com>, spri...@eden.com (David Springer)
>writes:
[snip]
>I agree with David here. The issue is clearly not simply technical
>superiority. While OpenGL's technical superiority is really unquestioned
[snip]
> o Vendor support:
> OpenGL: Microsoft, Intel, SGI, DEC, IBM, Integraph, Evans &
> Sutherland, HP, Sun, 3Dlabs, & most PC graphics card vendors.
> Direct3D: Microsoft, & most PC graphics card vendors.
I've gotta disagree on this point... My biggest issue is hardware
accelleration. I disagree that "most PC graphics card vendors" have an
accellerated OpenGL driver. Since our games only run on "most PC
graphics card vendors", that is what matters the most to me in the
short term :) That is where the Microsoft people have pushed the hype.
Joe-Bob the consumer is going to want hardware accellerated graphics
for his new computer and he will note those buzzwords on the boxes or
web-pages of any games he wants. These are the requirements placed
upon us (the developers) by the Marketing department of MS and until
"most PC graphics card vendors" have implementations that can take
advantage of hardware I'm almost forced by market forces to use the
inferior solution that has been marketed better :)
If SGI (or whomever) can get board manufacturers to write accellerated
OpenGL drivers I bet you'd see a huge exodus from D3D IM.
[snip lots of good examples]
> o Maximum demonstrated performance:
> OpenGL: 11+ million polygons/second, 800 million pixels updated/second
> (SGI's InfiniteReality OpenGL implementation).
> Direct3D: Well under 1 million polygons/second, under 100 million pixels
> updated/second.
Ummm. This is kind of a strange comparison. I can't see what this
really tells us about the API (yes, it can gague overhead, but only on
the same OS?)
> o Demonstrated application domain:
> OpenGL: ALL interactive 3D application domains.
> Direct3D: Games, demos, 3D viewers.
As far as I'm concerned, I'm only really interested in "games, demos
and 3D viewers" :) I have not yet seen an OpenGL game (except those
that came bundled w/my Indy).
[snip other valid points]
>Honestly, anyone believing that Direct3D is "ahead" in any area but hype
>needs to look at the facts.
Agreed. The Marketing of D3D has made it easier for folks to recognize
a given buzzword and this leads to folks not having as many options
when selecting a 3D API.
Guess we're mostly in violent agreement here :)
Cheers,
Stephen Wilkinson "Programming is like pinball.
Software Engineer the reward for doing it well
Interactive Creations Inc. is the opportunity to do it
wi...@airmail.net again." (anon - Wizards Bane)
RNDM in Warbirds http://www.icigames.com
Stephen Random Wilkinson
Hello!
On 28 Dec 1996 23:50:34 GMT, spri...@eden.com (David Springer) wrote:
[huge snip]
>Believe it or not I have absolutely no preference either way which
>3D API comes out on top as long as ONE OF THEM does by this summer
>when I plan on picking something.
Hmmm. I bet this is why your posts have the flavor they do. I was
pretty ambivalent too until I had to actually sit down and crank out
some D3D IM code. To paraphrase another post (B. Hook?) I thought "How
hard can it be? What's one more API?" Then I sat down with the
documentation (ha!) and tried to slap out some D3D IM code. I had the
slides from the CGDC presentations and some help from the wksoftware
web page and it was still a bitch. I am not old as dirt but I've
programmed for ~7 years and I've never seen code like D3D IM. It's
just nasty and doesn't match up with practices I'm familiar with from
coding Defense apps, Embedded Training apps, or Simulation apps (and
now games). It's just it's own kind of oddness. This is _not_ an
advantage in my book. Anyway, enough ranting from me. Break out your
compiler and slap together a rotating cube in D3D IM. Then get an
OpenGL book and try the same w/OpenGL. This won't resolve the
arguments about market-share and HW vendor support, but you can choose
based on _your_ experiences rather than what others are arguing here.
Good luck!
Robert Rodgers
Brian Hook <b...@wksoftware.com> wrote:
>Robert Rodgers wrote:
>> The reference implementation's performance sucks and, as good as MS's
>> software implementation *is*, it's performance still sucks for SW only
>> -- it just does not compare to D3d running in ramp. It was even worse
>> back when MS bought Rendermorphics.
>
>Well, that's not exactly fair -- comparing ramp performance vs. a pure
>RGB mode? What about comparing OpenGL vs. the D3D RGB software driver?
How many games using SW rendering are using the RGB driver?
This is my objection to the debate, this sort of odd comparisons.
Ramp in D3d is usable. RGB is not. Nobody is using RGB (sw), nobody
is going to use it, nobody is suggesting that anyone use it.
So what does it matter that D3d RGB (sw) can be as slow as OpenGL?
That's like saying, "Sure, your Pentium is fast, but what if we only
use floppies? Then my 486 is about as fast as your Pentium."
I am not defending d3d here, my point was merely that this issue has
become too politicized (either OpenGL s. D3d or whoever vs. Microsoft)
which totally misses the issue which is that all we want is for MS to
Directdraw enable OpenGL.
I couldn't care less what MS does with D3d if COGL is as fast as D3d
(ramp!) and can be used in fsexclusive. If COGL doesn't do DDraw,
it's worthless. If it does and has good performance, use it. Neither
has anything do do with D3d (or, for that matter, Microsoft).
[...]
>> Their reasoning was "no time to rewrite and tune OpenGL." They were
>> probably right at the time.
>
>As it turns, they were WRONG, because now MS OpenGL's performance is
Now? Really? Where?
Mark Kilgard
|> >software. The only thing Direct3D has ever been ahead in is
|> >hype share.
|>
|> Direct3d's hardware support is better.
That's just not true. Only Microsoft hype says Direct3D's performance
as a software renderer is better, but the truth is rather more complicated.
Actually, Microsoft is completely unwilling to do side-by-side
comparisons between their OpenGL and Direct3D software renderers.
What would be revealed is how badly they've overhyped Direct3D. The
truth is that Microsoft's OpenGL 1.1 transformation speed is about
the same or _better_ than Direct3D. The rasterization (pixel update)
speed slightly lags in several cases. The only area Direct3D really
does much better is in Direct3D's "ramp" or color index mode. What that
comes down to is tuning; SGI's Cosmo OpenGL shows that that path too
can be equivalently tuned if desired. There's no inherent reason Direct3D
in software is slower or faster than OpenGL in software.
Microsoft has hyped Direct3D's performance but refused to do head-to-head
benchmarks of its softward Direct3D vs. software OpenGL 1.1
implementations for risk of embarrasing their Direct3D group.
For all the resource thrown into Direct3D, it would simply be
too embarassing to admit that Direct3D was only a small fraction
faster than software OpenGL 1.1 in some cases. If equivalent
resources were put into OpenGL, there'd be no performance reason
for Direct3D at all.
As for the driver model, OpenGL now has good hardware driver
models for Windows 95 (NT has had these for a long, long time).
The only significant advantage Direct3D once had is basically
gone now. For info on the OpenGL driver models for Windows 95
see:
http://www.wksoftware.com/publications/mcd_icd.html
|> The politicization of this issue is ridiculous.
I agree. It is amazing that Microsoft's internal politics can so
badly distort Microsoft's technology roadmap and how the
need by those within Microsoft reponsible for the RenderMorphics
purchase to justify the money they threw at RenderMorphics.
- Mark
Mark Kilgard
In article <5a8nlv$cuk$1...@boris.eden.com>, spri...@eden.com (David Springer)
writes:
|> Mark Kilgard (m...@woodsy.asd.sgi.com) wrote:
|> : In article <5a4boa$cs5$1...@boris.eden.com>, spri...@eden.com (David
|> Springer)
|> : writes:
|> : |> Believe it or not I have absolutely no preference either way which
|> : |> 3D API comes out on top as long as ONE OF THEM does by this summer
|> : |> when I plan on picking something.
|>
|> : should only be one 3D API. I hope you'll make that desire clear to
|> : Microsoft.
|>
I'll remind you that John Carmack's conclusion was that he doesn't
"think there is ANY market segment that D3D is apropriate for"
(the capitalized ANY is Carmack's emphasis). Direct3D is even a
failure in the game market it is targeted for.
|> If everyone were of exactly the same opinion and as vehemently
|> opposed to D3D as Carmack then I'd have no doubt that D3D is
|> well and truly broken. However, some people disagree, if not
|> in principle then at least in degree
Honestly, who are these people? Give me and others a chance to
answer their analysis. In what way do these "some people that
disagree" believe that is Direct3D superior to OpenGL?
|> Personally I'd rather see OpenGL come out on top if it's the
|> best choice and if you do a DejaVu search on my writing in
|> this group looking for references to OpenGL you'll find that
|> 18 months ago I was arguing FOR OpenGL's inclusion in Win 95
|> and AGAINST some Rendermorphics derivative.
That's great to know. I hope that you'll review the analysis
from John Carmack and others today and see that your analysis
18 months ago is even more true today.
- Mark
Mark Kilgard
In article <32c83a5a...@news.cris.com>, wi...@airmail.net (Stephen "Random"
Wilkinson) writes:
|> On 30 Dec 1996 10:56:37 GMT, m...@woodsy.asd.sgi.com (Mark Kilgard)
|> wrote:
|> >I agree with David here. The issue is clearly not simply technical
|> >superiority. While OpenGL's technical superiority is really unquestioned
|> [snip]
|> > o Vendor support:
|> > OpenGL: Microsoft, Intel, SGI, DEC, IBM, Integraph, Evans &
|> > Sutherland, HP, Sun, 3Dlabs, & most PC graphics card vendors.
|> > Direct3D: Microsoft, & most PC graphics card vendors.
|>
|> I've gotta disagree on this point... My biggest issue is hardware
|> accelleration. I disagree that "most PC graphics card vendors" have an
|> accellerated OpenGL driver. Since our games only run on "most PC
|> graphics card vendors", that is what matters the most to me in the
|> short term :)
Actually, yes, I agree with you. Probably today, it would be better
to say "numerous" PC graphics card vendors as compared to "most". Let
me please correct myself on this point.
The problem is that the dynamic is actually changing as we speak,
but I'll be fair and just say what's true today.
|> That is where the Microsoft people have pushed the hype.
|> Joe-Bob the consumer is going to want hardware accellerated graphics
|> for his new computer and he will note those buzzwords on the boxes or
|> web-pages of any games he wants. These are the requirements placed
|> upon us (the developers) by the Marketing department of MS and until
|> "most PC graphics card vendors" have implementations that can take
|> advantage of hardware I'm almost forced by market forces to use the
|> inferior solution that has been marketed better :)
|> If SGI (or whomever) can get board manufacturers to write accellerated
|> OpenGL drivers I bet you'd see a huge exodus from D3D IM.
Yep, the exodus is coming. I've known John Carmack's opinions on OpenGL
and Direct3D for about two months now; it just took a while for those
comments to be written up and known publically. Honestly, even I was
surprised just how pro-OpenGL his analysis ended up being. Obviously, he
talks to a lot of 3D graphics card vendors and I believe he's been letting
them know for some time what his conclusions are about OpenGL vs.
Direct3D. The graphics card vendors know the truth.
I think your huge exodus from D3D IM is a pretty solid prediction
at this point. It just takes a bit of time for the 3D graphics
card vendors to prepare drivers and announce new hardware. It's kind of
been lost in the noise, but Microsoft DOES have a reasonable hardware
driver model for OpenGL on Windows 95 now. Of course Microsoft and
the card vendors certainly don't want to admit the reality of the
situation prematurely. Doing so would just kill their current board
sales. They aren't stupid.
- Mark
Clinton Keith
Brian Hook <b...@wksoftware.com> wrote in article <32C70D...@wksoftware.com>...
> Robert Rodgers wrote:
> > The reference implementation's performance sucks and, as good as MS's
> > software implementation *is*, it's performance still sucks for SW only
> > -- it just does not compare to D3d running in ramp. It was even worse
> > back when MS bought Rendermorphics.
>
> Well, that's not exactly fair -- comparing ramp performance vs. a pure
> RGB mode? What about comparing OpenGL vs. the D3D RGB software driver?
The performance for either one sucks. That's really the whole point to
vendors writing games *now*. If the MS OpenGL had a ramp-mode and
also ran accelerated on alot of low cost 3D boards, then I'd go for it.
OpenGL is a superior API.
>
> > Their reasoning was "no time to rewrite and tune OpenGL." They were
> > probably right at the time.
>
> As it turns, they were WRONG, because now MS OpenGL's performance is
> very close to D3D's performance, and in some cases faster. And this is
> a VERY small group working on OpenGL.
Maybe compared to D3D's RGB mode. Though even ramp's performance
is so bad, we're considering a custom software renderer.
Clinton Keith
Mark Kilgard <m...@woodsy.asd.sgi.com> wrote in article <5a9qd8$b...@fido.asd.sgi.com>...
>
> Microsoft has hyped Direct3D's performance but refused to do head-to-head
> benchmarks of its softward Direct3D vs. software OpenGL 1.1
> implementations for risk of embarrasing their Direct3D group.
> For all the resource thrown into Direct3D, it would simply be
> too embarassing to admit that Direct3D was only a small fraction
> faster than software OpenGL 1.1 in some cases. If equivalent
> resources were put into OpenGL, there'd be no performance reason
> for Direct3D at all.
>
saying how MS's implementation of OpenGL was too slow:
"Do it yourself, SGI"
Why wait for Microsoft to publish benchmarks. Perform them yourself.
Publish the code and your results. Prove your point.
> As for the driver model, OpenGL now has good hardware driver
> models for Windows 95 (NT has had these for a long, long time).
> The only significant advantage Direct3D once had is basically
> gone now. For info on the OpenGL driver models for Windows 95
> see:
>
> http://www.wksoftware.com/publications/mcd_icd.html
>
What does another hardware model do unless it is currently
supported by the vendors? Unless you can tell me that all the current
D3D accelerator cards will accelerate OpenGL as fast, or faster,
then it's just vaporware. I don't want to hear about the > $500 cards
either.
> |> The politicization of this issue is ridiculous.
Yup. The cow-pies are flying in both directions. I'd prefer to
use OpenGL personally. My first IM example I wrote, to rotate
a triangle, took 10 pages of source. However I'm not going to wait for
"hardware model" to be adopted when millions of D3D compatible
cards are being sold.
--
Clinton Keith
Angel Studios
(W) cl...@angel.com
(H) clinto...@worldnet.att.net
Mark Kilgard
In article <01bbf6e6$e11e7040$010000c8@clint>, "Clinton Keith"
<Clinto...@worldnet.att.com> writes:
|> Mark Kilgard <m...@woodsy.asd.sgi.com> wrote in article
|> <5a9qd8$b...@fido.asd.sgi.com>...
|> > models for Windows 95 (NT has had these for a long, long time).
|> > The only significant advantage Direct3D once had is basically
|> > gone now. For info on the OpenGL driver models for Windows 95
|> > see:
|> >
|> > http://www.wksoftware.com/publications/mcd_icd.html
|>
|> What does another hardware model do unless it is currently
|> supported by the vendors?
The MCD model is already supported for Windows NT. The fact that
the model exists means that you can start reasonably demanding
that 3D PC card vendors accelerate OpenGL. Before MCD, Microsoft's
Direct3D group had effectively stacked the deck in their favor, but
that's no longer true.
|> > |> The politicization of this issue is ridiculous.
|>
|> Yup. The cow-pies are flying in both directions. I'd prefer to
|> use OpenGL personally. My first IM example I wrote, to rotate
|> a triangle, took 10 pages of source.
Does that seem reasonable?
|> However I'm not going to wait for
|> "hardware model" to be adopted when millions of D3D compatible
|> cards are being sold.
The fact is even with 10 pages of source code, a real application still
can't guarantee that it will really run correctly (not just some toy
demo) on those millions of D3D compatible cards you claim to exist.
Direct3D means you have to buy multiple vendors cards to for reasonable
testing and even then you can probably still expect returns.
- Mark
Suzy Deffeyes
I would like to echo some of Mark Kilgard's comments on this
never ending discussion.
All of Mark's comments on Vendor support, OS support are important
ones (next time remember to include OS/2 to your list, please Mark).
OpenGL also has standardized benchmarking which is voted on by
members of the OpenGL Performance Characterization Committee.
Performance numbers are published only on approved benchmarks,
and are subject to peer review before publication. You can check
out the widespread results from many of the vendors Mark mentioned
by going to http://www.specbench.org/gpc/opc
Another point that should be mentioned is the widespread range of
performance that OpenGL provides a device independant interface to.
OpenGL runs on low end PCs on up to big beasts (like SGI Infinite
Reality). Direct3D is targeted at a smaller set of uses.
As for college acceptance, and ease of learning, the graphics prof
has all his students use OpenGL in their graphics programming class.
I frequently see the Addison-Wesley OpenGL books listed on the
'best selling computer books' list at the University bookstore.
I have helped a 15 year old high school kid program OpenGL on
his PC running OS/2.
And as for API maturity, OpenGL was almost named GL 5.0, following
on to many previous versions of GL done by SGI. They've been
designing 3D APIs for a while. They certainly understand how
to exploit graphics hardware, which is what defining a 3D API
is all about: defining an interface to graphics hardware.
Just another Vendor supporting OpenGL-
Suzy Deffeyes
--
====================================================================
Suzy Deffeyes su...@austin.ibm.com deffeyes@ausvm6
IBM OpenGL Development (512)838-2818 I speak for myself.
============= OpenGL: It's Everywhere. Do the Math. ================
Paul Shirley
In article <5a8nlv$cuk$1...@boris.eden.com>, David Springer
<spri...@eden.com> writes
>Lastly, to re-use one of Brian or Kendall's points from 18
>months ago - Rendermorphics Reality Lab was the engine of
>choice among many game developers then and apparently continues
>to be acceptable to at least some of them in its D3D incarnation
>now. Please forgive me if I've mixed up any names - brand
>names of products I've never used don't stick well.
Reality Lab was certainly *preferred* (over BRender).
That does *not* mean it was actually used for released games.
AFAIK almost no games have been written in any of the 3d toolkits.
Paul Shirley
In article <32c7c80...@news.wam.umd.edu>, Robert Rodgers
<rsro...@wam.umd.edu> writes
>We'll see what Cosmo OpenGL looks like when it ships, but for now,
>OpenGL is unusable for games. The hottest 3d cards on the market
>don't support it at all (3dfx) and *do* support Direct3d.
You should really check the 3DFX pages more often... OpenGL is coming.
Given that the voodoo chipset currently wipes the floor with all the
others, most of the cards sold will likely have an OpenGL driver ;)
Robert Rodgers
Paul Shirley <Pa...@foobar.co.uk.chocolat> wrote:
>In article <32c7c80...@news.wam.umd.edu>, Robert Rodgers
><rsro...@wam.umd.edu> writes
>>We'll see what Cosmo OpenGL looks like when it ships, but for now,
>>OpenGL is unusable for games. The hottest 3d cards on the market
>>don't support it at all (3dfx) and *do* support Direct3d.
>
>You should really check the 3DFX pages more often... OpenGL is coming.
For the Rush. Is 3dfx now announced that OpenGL *will* arrive for
the voodooo (fullscreen only) board? And note also the tense: "do"
vs. "will."
>Given that the voodoo chipset currently wipes the floor with all the
>others, most of the cards sold will likely have an OpenGL driver ;)
I'm not talking about the potential changes in the future.
I'd be perfectly happy to see D3d go the way of DOS4, and happier
still if it's replacement was QD3d or OpenGl, and happier yet more if
cheap OpenGl hardware became commonplace. All of these things are
great. But they'll only happen if MS and SGI decide to make them
happen, and that's independent of what we say here, and independent of
whether D3dIM is the ugliest API on the market.
Brian Hook
David Springer wrote:
> Only one 3D API targeted towards *games*. I think there should be
> experience tells me that precision CAD/CAM applications and
> entertainment applications take different approaches regards
> tradeoffs in speed, accurcy, and ease of use. I'm not convinced that
> a single 3D API will address both with the same efficiency.
Then your experience is flawed. Those of us who actually work on 3D
graphics are quite aware that fundamentally OpenGL and D3D do not differ
in functionality and feature set. The same types of operations occur in
both APIs. The difference (to a developer) is simply in ease of use and
minor features that could easily be added to either.
People have this weird impression that OpenGL is somehow like
Renderman....
Brian
--
+-----------------------------------------------------------------+
+ Brian Hook, b...@wksoftware.com +
+ WK Software, http://www.wksoftware.com +
+ Consultants specializing in 3D graphics hardware and software +
+ --------------------------------------------------------------- +
+ For a list of publications on 3D graphics programming, +
+ including Direct3D, OpenGL, and book reviews: +
+ http://www.wksoftware.com/publications.html +
+-----------------------------------------------------------------+
Chris Altmann
Paul Shirley <Pa...@foobar.co.uk.chocolat> wrote:
>In article <5a8nlv$cuk$1...@boris.eden.com>, David Springer
><spri...@eden.com> writes
>>Lastly, to re-use one of Brian or Kendall's points from 18
>>months ago - Rendermorphics Reality Lab was the engine of
>>choice among many game developers then and apparently continues
>>to be acceptable to at least some of them in its D3D incarnation
>>now. Please forgive me if I've mixed up any names - brand
>>names of products I've never used don't stick well.
>Reality Lab was certainly *preferred* (over BRender).
>That does *not* mean it was actually used for released games.
>AFAIK almost no games have been written in any of the 3d toolkits.
Warhammer used the Reality Lab engine. I think one or more of those
weird 'drive around inside a cylinder playing rollerball sport' games
used it as well.
__________________________________________________
Except for the historical information
contained herein, the matters discussed
in this communication are forward looking
statements that involve risks and
uncertainties that could cause actual
results to differ materially from those
projected. Have a nice day.
Chris Altmann - alt...@netcom.com
__________________________________________________
Glenn Corpes
m...@woodsy.asd.sgi.com (Mark Kilgard) wrote:
Does MS's openGL system work with under direct draw, without direct draw
or both?
- - --=-==< gco...@ea.com, Head of R&D Bullfrog Productions >==-=-- - -
Michael I. Gold
gle...@cix.co.uk (Glenn Corpes) writes:
| Does MS's openGL system work with under direct draw, without direct draw
| or both?
Currently MS OpenGL does not use ddraw. I understand this will addressed
Real Soon Now.
--
Michael I. Gold Silicon Graphics Inc. http://reality.sgi.com/gold
And my mama cried, "Nanook a no no! Don't be a naughty eskimo! Save your
money, don't go to the show!" Well I turned around and I said, "Ho! Ho!"
George MacDonald
Robert Rodgers wrote:
> I am not defending d3d here, my point was merely that this issue has
> become too politicized (either OpenGL s. D3d or whoever vs. Microsoft)
> which totally misses the issue which is that all we want is for MS to
> Directdraw enable OpenGL.
>
> I couldn't care less what MS does with D3d if COGL is as fast as D3d
> (ramp!) and can be used in fsexclusive. If COGL doesn't do DDraw,
> it's worthless. If it does and has good performance, use it. Neither
> has anything do do with D3d (or, for that matter, Microsoft).
Vendor support can be reduced to a very few chip sets in the mass
market. Assume that the installed base turns over every few years. If
all new machines have some kind of 3d accelerator, then by x-mass '97
a minority of the installed base will have some kind of 3D accelerator.
By x-mass '98 it will be about even. Only by x-mass '99 will the
accelerated crew be in the majority.
But even then, most of the installed base will have relatively low
performance 3d accelerators. These will be the cards installed by OEMs,
not the higher performance cards that users will buy as upgrades. Most
will only have 2MB memory and will combine 2D and 3D functionality on
the cheapest possible chipset. What is the status of OpenGL with these
major OEM suppliers:
1) Does the S3 ViREG graphics chips have drivers to support
OpenGL?
If not now, when?
2) Does the ATI Rage/Rage II have drivers to support OpenGL?
If not now, when?
3) Does NEC PowerVR have drivers to support OpenGL?
If not now, when?
4) Does the Rendition Verite have drivers to support OpenGL?
If not now, when?
Given OpenGL's strict functionality requirements, can all these
consumer level cards support OpenGL without a massive investment
in writing drivers?
If not, then we may never see OpenGL support on a majority of consumer
level boards. But we know we will see every consumer level board
supporting D3d.
I don't have the expertise to answer these questions. But market
penetration will be an important factor in determining which API gets
used by the majority of software developers.
George MacDonald
Brian Hook
David Springer wrote:
>
> Brian Hook (b...@wksoftware.com) wrote:
> : David Springer wrote:
> : > Only one 3D API targeted towards *games*. I think there should be
> : > two API's if their missions are significantly different. My
> : > experience tells me that precision CAD/CAM applications and
> : > entertainment applications take different approaches regards
> : > tradeoffs in speed, accurcy, and ease of use. I'm not convinced that
> : > a single 3D API will address both with the same efficiency.
>
> : Then your experience is flawed. Those of us who actually work on 3D
> : graphics are quite aware that fundamentally OpenGL and D3D do not differ
> : in functionality and feature set. The same types of operations occur in
> : both APIs. The difference (to a developer) is simply in ease of use and
> : minor features that could easily be added to either.
> from you of all people but when you use it on me it only serves to
> validate my using one on you.
That's amusing, since I don't see in the above where I had a
condescending attitude. It's not like I called you a "bright kid" or
something.
> Third - the same types of operations occur on a MacIntosh as on a
> IBM PC clone, the devil is in the details. Your "same type" of
> operations argument is flawed. :-)
This is classic misdirection.
Instead of telling me that my argument is flawed, tell me WHY my
argument is flawed. The same types of operations DO occur in both APIs,
with minor differences (D3D has a no clipping hint, D3D has a separate
specular color, D3D has per vertex fog, OpenGL has image processing and
histogram functions, OpenGL supports a stencil buffer and
overlay/underlay planes, etc. etc.).
My argument still stands.
> Lastly, you point out that the differences between OpenGL and D3D
> are (I quote) "simply in ease of use and minor features that can be added
> to either". Without going into detail about how this contradicts
> things you've said before, the answer is then clearly for Microsoft to
> simply fix D3D instead of replacing it.
I've quoted my original post at the beginning of this post. Please
don't quote me out of context. Specifically, I said (emphasis mine):
"The difference (TO A DEVELOPER) is simply in ease of use and minor
features that could easily be added to either."
If you leave off the "TO A DEVELOPER" part and insert "between OpenGL
and D3D" instead, which you did, you radically change the meaning of the
sentence. When "quoting" me, please try and keep the quotes intact.
Brian
Gil Colgate
David Springer wrote:
>
> Suzy Deffeyes (su...@austin.ibm.com) wrote:
>
> [snip]
>
> : I have helped a 15 year old high school kid program OpenGL on
> : his PC running OS/2.
>
> His name wasn't Brian Hook by any chance was it ? ;)
>
> Seriously, are you suggesting that we judge software API's by how
> readily children can program effectively with them ?
>
> David
I'm sure he isn't, but remember the bottom line of API design:
KISS.(Keep it simple, stupid).
Game developers shied away from Windows before Direct X not because fast
action games can't be written in Windows, but because it was too
complicated.
The part of the code that is complicated, oddly enough, is the
part that is slow.......
Thus if an API is more complicated than a competing one, and they
both have the same functionality, than the simpler one is better.
Of course, the simpler one doesn't always win.....
David Springer
Stephen "Random" Wilkinson (wi...@airmail.net) wrote:
: Hello!
: On 28 Dec 1996 23:50:34 GMT, spri...@eden.com (David Springer) wrote:
: [huge snip]
: >Believe it or not I have absolutely no preference either way which
: >3D API comes out on top as long as ONE OF THEM does by this summer
: >when I plan on picking something.
: Hmmm. I bet this is why your posts have the flavor they do. I was
: pretty ambivalent too until I had to actually sit down and crank out
: some D3D IM code. To paraphrase another post (B. Hook?) I thought "How
: hard can it be? What's one more API?" Then I sat down with the
: documentation (ha!) and tried to slap out some D3D IM code. I had the
: slides from the CGDC presentations and some help from the wksoftware
: web page and it was still a bitch. I am not old as dirt but I've
: programmed for ~7 years and I've never seen code like D3D IM. It's
: just nasty and doesn't match up with practices I'm familiar with from
: coding Defense apps, Embedded Training apps, or Simulation apps (and
: now games). It's just it's own kind of oddness. This is _not_ an
: advantage in my book. Anyway, enough ranting from me. Break out your
: compiler and slap together a rotating cube in D3D IM. Then get an
: OpenGL book and try the same w/OpenGL. This won't resolve the
: arguments about market-share and HW vendor support, but you can choose
: based on _your_ experiences rather than what others are arguing here.
Very good points Stephen and probably an accurate guess to what
I'd think of D3D IM at that time. However, by the time I get around
to actually using it those problems will have been ironed out and
I'll use someone elses C++ wrapper classes with much more abstraction
than the D3D IM COM object has to offer.
In fact for my purposes I'll probably be able to skip right on
over D3D IM and use D3D RM. The sacrifice is probably, what, less than
comparing D3D IM on a P75 and a P133 ? So I get last year's state
of the art 3D performance this year in other words. My games depend
on a broad mix of technologies, not just one, so I don't need to
kill myself trying to be bleeding edge in just one area.
David Springer
--
************ IGAMES ONLINE ENTERTAINMENT SYSTEM ***********
* *
* A client/server system designed from the ground up to *
* support real time multi-player games on the Internet. *
* Game developers, players, and ISP's check it out at: *
* *
****************** http://www.igames.com ******************
David Springer
Mark Kilgard (m...@woodsy.asd.sgi.com) wrote:
: In article <5a8nlv$cuk$1...@boris.eden.com>, spri...@eden.com (David Springer)
: writes:
: |> Mark Kilgard (m...@woodsy.asd.sgi.com) wrote:
: |> : In article <5a4boa$cs5$1...@boris.eden.com>, spri...@eden.com (David
: |> Springer)
: |> : writes:
: |> If everyone were of exactly the same opinion and as vehemently
: |> opposed to D3D as Carmack then I'd have no doubt that D3D is
: |> well and truly broken. However, some people disagree, if not
: |> in principle then at least in degree
: Honestly, who are these people? Give me and others a chance to
Activision and Electronic Arts employees have posted in this thread.
Pay attention.
They've both indicated that they will publish with D3D so it apparently
isn't quite as broken as Carmack, SGI, et all, would have you believe.
My *personal* opinion is that the people posting from Activision
and EA are just as equally qualified as Carmack. Quake doesn't impress
me a whole lot as being much more than an incremental improvement
of DOOM. More hype than substance in other words. So it's higher
resolution and you can look up and down now - big deal - one would
expect some improvement considering I first saw DOOM on a 386 and I
first saw Quake on a P166 for gawd's sake.
: answer their analysis. In what way do these "some people that
: disagree" believe that is Direct3D superior to OpenGL?
For Christ's sake Mark - how many different ways can it be
spelled out to you ? They believe D3D has better marketing
support, for games, than OpenGL. They don't need any more
reason than that. Microsoft is pushing D3D and that's that.
They don't want to buck Microsoft and no one concerned with the
bottom line can blame them.
: |> 18 months ago I was arguing FOR OpenGL's inclusion in Win 95
: |> and AGAINST some Rendermorphics derivative.
: That's great to know. I hope that you'll review the analysis
: from John Carmack and others today and see that your analysis
: 18 months ago is even more true today.
It's still true today but it's too late for it to come true. D3D
won the contest... it's time to accept that and move on. I'm not
one to cry over spilt milk. I'm just as lazy as the next programmer,
probably more so, and if you say D3D IM is gonna make me miserable
compared to OpenGL I'm prepared to believe that since I figured
as much 18 months ago. BUT ... there's not a damn thing I can do about
it.
Moreover, I think any attempt to slow down MS at this point will only
make things worse. I'd rather have them fix the known problems in
D3D then reverse course and try to replace it with OpenGL this close
to Win 97.
David Springer
Brian Hook (b...@wksoftware.com) wrote:
: David Springer wrote:
: > Only one 3D API targeted towards *games*. I think there should be
: > two API's if their missions are significantly different. My
: > experience tells me that precision CAD/CAM applications and
: > entertainment applications take different approaches regards
: > tradeoffs in speed, accurcy, and ease of use. I'm not convinced that
: > a single 3D API will address both with the same efficiency.
: Then your experience is flawed. Those of us who actually work on 3D
: graphics are quite aware that fundamentally OpenGL and D3D do not differ
: in functionality and feature set. The same types of operations occur in
: both APIs. The difference (to a developer) is simply in ease of use and
: minor features that could easily be added to either.
Brian, Brian, Brian... I'm sure I deserve a condescending attitude
from you of all people but when you use it on me it only serves to
validate my using one on you.
With that in mind let me start out by saying that those of us who
actually design PC's that are sold in the millions every year are
quite aware of which standards to support and when to support them.
Secondly, my experience isn't flawed. It's history. I'll concede
it may not be relevant today but flawed is not an applicable term.
Third - the same types of operations occur on a MacIntosh as on a
IBM PC clone, the devil is in the details. Your "same type" of
operations argument is flawed. :-)
Lastly, you point out that the differences between OpenGL and D3D
are (I quote) "simply in ease of use and minor features that can be added
to either". Without going into detail about how this contradicts
things you've said before, the answer is then clearly for Microsoft to
simply fix D3D instead of replacing it.
David Springer
David Springer
Suzy Deffeyes
David Springer wrote:
>
> Seriously, are you suggesting that we judge software API's by how
> readily children can program effectively with them ?
>
Why isn't it valid? I wasn't insinuating that it is a child's
API, just showing the scope of use OpenGL is getting. OpenGL
is usable, and a growing number of people (the 15 year old
is probably close to 18 by now) are aquiring OpenGL programming
skills.
Suzy
Brandon Van Every
>
> |> If everyone were of exactly the same opinion and as vehemently
> |> opposed to D3D as Carmack then I'd have no doubt that D3D is
> |> well and truly broken. However, some people disagree, if not
> |> in principle then at least in degree
>
> Honestly, who are these people? Give me and others a chance to
> disagree" believe that is Direct3D superior to OpenGL?
I've argued at great length with Mark and Brian for several months now.
People who have followed comp.graphics.api.opengl and rec.games.programmer
know that I don't ever *volunteer* to agree with them! :-) But I
completely agree with their technical assessments of D3D vs. OpenGL. D3D
basically sucks.
The only difference between their outlook and mine, is that we work for
different companies. DEC will implement D3D because it will make some
money. But it certainly seems like a lot of busywork, and I'd be just as
happy to see D3D fail, as that would save us development time. It's sort
of a "win-win" situation, as far as we're concerned. We're just playing
both sides of the fence.
Cheers,
--
Brandon J. Van Every | Free3d: old code never dies! :-)
| Starter code for GNU Copyleft projects.
DEC Graphics & Multimedia |
Windows NT Alpha OpenGL | vane...@blarg.net www.blarg.net/~vanevery
Brandon Van Every
>
> I could easily imagine someone being involved in DX and thinking "You
> know, we could easily do our own 3D API and not have to share control
> with those pukes in the OpenGL group".
>
> The rumors of the OpenGL vs. D3D rivalry internal to Microsoft are
> legendary.
I think you're missing the boat as far as who's causing the confusion.
It's not technical group vs. technical group. It's technical groups vs.
marketing/management groups. The marketroids simply said "we want to own
3d API-dom," bought RenderMorphics to do that job, and then tried to tell
the OpenGL folks to go do silly things like implement their stuff on top of
D3D.
David Springer
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^
: That's amusing, since I don't see in the above where I had a
: condescending attitude. It's not like I called you a "bright kid" or
: something.
Brian, I was developing drivers and graphics hardware for AutoDesk
products, including 3D oriented products, in 1987. That's ten years
ago. I OWNED the company that was doing it in fact and my products
were receiving great reviews. What were you doing in 1987 Brian ?
Your algebra homework in high school perhaps ?
You have no idea what all my background includes yet you choose to
cop an attitude with me. That's fine, I deserve it, but just be
aware that when you do I'll take that as consent to needle you in
return about your shortcomings.
: > Third - the same types of operations occur on a MacIntosh as on a
: > IBM PC clone, the devil is in the details. Your "same type" of
: > operations argument is flawed. :-)
: This is classic misdirection.
Really. Defend your statement then. I believe it's analogy rather
than misdirection.
: Instead of telling me that my argument is flawed, tell me WHY my
: argument is flawed. The same types of operations DO occur in both APIs,
: with minor differences (D3D has a no clipping hint, D3D has a separate
: specular color, D3D has per vertex fog, OpenGL has image processing and
: histogram functions, OpenGL supports a stencil buffer and
: overlay/underlay planes, etc. etc.).
I did say why your argument is flawed by way of analogy. To make it
plainer, your statement about same types of operations is so generalized
as to render it pointless. As pointless as trying to compare a
MacIntosh and an IBM PC clone by saying they perform the same type
of operations. Yeah, they do - so what ? The devil is plainly in
the details.
: My argument still stands.
It will always stand in your eyes. Unilateral declarations of victory
are valid to only one of the contestants.
: > Lastly, you point out that the differences between OpenGL and D3D
: > are (I quote) "simply in ease of use and minor features that can be added
: > to either". Without going into detail about how this contradicts
: > things you've said before, the answer is then clearly for Microsoft to
: > simply fix D3D instead of replacing it.
: I've quoted my original post at the beginning of this post. Please
: don't quote me out of context. Specifically, I said (emphasis mine):
: "The difference (TO A DEVELOPER) is simply in ease of use and minor
: features that could easily be added to either."
I stand corrected. You're excellent at spotting yourself being
taken out of context... the expression "it takes a thief to know
one" comes to mind. I'll ask that you quote me correctly in the
future too including the context.
: If you leave off the "TO A DEVELOPER" part and insert "between OpenGL
: and D3D" instead, which you did, you radically change the meaning of the
: sentence. When "quoting" me, please try and keep the quotes intact.
Anyhow - so you're saying the difference, from a developer's standpoint,
is trivial ? Then why does Carmack say D3D is fundamentally broken from
an API point of view ? Your assertion of a trivial difference viewed
from the outside doesn't square with Carmack's.
Brandon Van Every
> In article <5a19ld$4qi$1...@boris.eden.com>, spri...@eden.com (David
Springer)
> writes:
> |> Thomas B Valesky (tval...@osf1.gmu.edu) wrote:
> |> : Frank Pitt (fra...@mundens.welly.gen.nz) wrote:
> |> : : In article <59rf0v$r...@portal.gmu.edu> tval...@osf1.gmu.edu
writes:
> |>
> |> : Well, if OpenGL has a friend in Id Software, it might make some card
> |> : manufacturers sit up and take notice. Microsoft doesn't win _all_
> |> : the time. :-)
> |>
> |> This is the mistake that many of you are making. What makes card
> |> manufacturers sit up and take notice is multimillion dollar purchase
> |> orders from Dell, IBM, Compaq, Gateway, Toshiba, NEC, Packard Bell,
> |> DEC, and HP. Id doesn't buy millions of graphics cards and Id
> |> doesn't make Compaq sit up and take notice. Microsoft makes Compaq
> |> sit up and take notice. If OpenGL is to come out on top something
> |> has to make Microsoft sit up and take notice.
Ok I hope that Real Soon Now we can make you sit up and take notice.
And hopefully, we can make you notice why you should all ditch this stupid
Intel CISC thingy. :-)
Brandon Van Every
> |> but whether it will ever get marketshare.
>
> OpenGL is ahead in market share. I assure you that the total
> dollar value of 3D hardware and software for OpenGL sold is
> significantly higher than that spent on Direct3D hardware and
> software. The only thing Direct3D has ever been ahead in is
> hype share.
Be fair. The market share under discussion is commodity graphics for the
Intel-based PC, *not* all 3d graphics hardware everywhere.
Paul Shirley
In article <5af33p$ot6$1...@boris.eden.com>, David Springer
<spri...@eden.com> writes
>In fact for my purposes I'll probably be able to skip right on
>over D3D IM and use D3D RM. The sacrifice is probably, what, less than
>comparing D3D IM on a P75 and a P133 ? So I get last year's state
>of the art 3D performance this year in other words. My games depend
>on a broad mix of technologies, not just one, so I don't need to
>kill myself trying to be bleeding edge in just one area.
I,for one, will keep this in mind while reading any of your future
posts. You just seriously devalued your credibility ;)
Andrew Matthew Theurer
In article <32CC6E...@student.hk-r.se>,
Mats Grahm <pt9...@student.hk-r.se> wrote:
>David Springer wrote:
>
>> : answer their analysis. In what way do these "some people that
>> : disagree" believe that is Direct3D superior to OpenGL?
>>
>> spelled out to you ? They believe D3D has better marketing
>> support, for games, than OpenGL. They don't need any more
>> reason than that. Microsoft is pushing D3D and that's that.
>> They don't want to buck Microsoft and no one concerned with the
>> bottom line can blame them.
>
>still are people out there who love to debate how much better
>the world would have been should Betamax had won the VCR battle.
>But people actually INVOLVED in video production have stopped
>grieving decades ago.
Actually, people involved in video production use ONLY betamax
today. Next time you see your local news camera on the scene,
check it out, it's a betacam.
The problems with D3D are already starting. The customer is
not getting a clear picture of what their card does or does not
do well. Example: go check out the scuffle between ATI and
Matrox. Since there are no conformance tests, we have card
makers who will naturally skimp on features. The typical
customer wll have no idea until they play their game, and it performs
like total shit.
At least when there was no 3D hardware, game makers could
put on the box "486/66 and above". What will we do with D3D?
With OGL, we can say "OGL viewperf xx.x and above for good game play"
>
>Same with D3D/OGL. Which one is "better" is, how sad it might
>seem to some people, not all that relevant. The one that gets
>critical mass first will probably take it all. The jury is
sadly, this is true. PC industry motto: hack, fix, hack, fix.
Hey, that sounds like another product -WIN95!
>still out, maybe OGL can get a share of the home market too.
>But if serious developers think MS will win, then they just
>accept fact, learn how to use the shit, and stop treating the
>subject like a religios war!
If we just accept that microsoft will stick only w/ D3D, you
might as well send them a blank check. If people don't at least
say "I prefer OGL because I have better time to market, etc.
and I would like to see more OGL" OGL will never survive.
I sure hope OGL works out. Most of the new cards coming out
will support OGL. I know, I know, there still kind of expensive,
but like all harware, give it 6 months, and they'll be dirt cheap.
They also happen to blow away most 3D3 cards today, and support
much higher resolutions.
OGL's greatest chance will probably be when win9x is cut loose.
At least OGL has a bigger market on NT.
Andrew Theurer.
Tim Beckmann
David Springer wrote:
> Lastly, you point out that the differences between OpenGL and D3D
> are (I quote) "simply in ease of use and minor features that can be added
> to either". Without going into detail about how this contradicts
> things you've said before, the answer is then clearly for Microsoft to
> simply fix D3D instead of replacing it.
>
David,
From the example code I've seen of D3D and OpenGL, the only way to "fix"
D3D is to kill it - by the time you fix it, you've got OpenGL :)
Tim
Brian Hook
> : his PC running OS/2.
>
> His name wasn't Brian Hook by any chance was it ? ;)
Christ David, how low do you plan on sinking? It amazes me that you've
even found two people who have bothered defending you, but if they have
any sense, I imagine that they'll stop soon enough after seeing comments
like that.
Brian
--
Brian Hook
Mats Grahm wrote:
> Same with D3D/OGL. Which one is "better" is, how sad it might
> seem to some people, not all that relevant. The one that gets
> critical mass first will probably take it all. The jury is
> But if serious developers think MS will win, then they just
> accept fact, learn how to use the shit, and stop treating the
> subject like a religios war!
The crux of the matter is that D3D has NOT gained critical mass yet.
Developers have a vested interest in seeing D3D die and OpenGL succeed
-- no developer I've ever spoken has said that they prefer D3D to
OpenGL. For this reason it's very important to see awareness of OpenGL
raised.
Stephen Random Wilkinson
On 2 Jan 1997 01:30:33 GMT, spri...@eden.com (David Springer) wrote:
[snip]
>I'd think of D3D IM at that time. However, by the time I get around
>to actually using it those problems will have been ironed out and
>I'll use someone elses C++ wrapper classes with much more abstraction
>than the D3D IM COM object has to offer.
Hmm. IMHO some things will not/cannot be fixed at the D3D IM API level
but you can layer whatever you want on top. But think of all the
layers you're adding to this problem. Do you really want that many
more layers on top of your graphics functions? Think of the overhead!
Yuk. Why not have a lowest level that is clean enough that you might
want to use it directly?
>In fact for my purposes I'll probably be able to skip right on
>over D3D IM and use D3D RM. The sacrifice is probably, what, less than
>comparing D3D IM on a P75 and a P133 ? So I get last year's state
>of the art 3D performance this year in other words. My games depend
>on a broad mix of technologies, not just one, so I don't need to
>kill myself trying to be bleeding edge in just one area.
There is a fair bit of overhead in using RM and the kind of
performance you will get out of it depends a lot on the kind of app.
Maybe this is enough for you. But you are skirting another issue here.
We say that D3D API sucks as a lowest level. You state that no, it's
fine _if_ we layer some other stuff on top of it.. So in effect,
you've thrown your choice as to ease of use (without writing any D3D
code yet that I can recall) to the OpenGL crowd :) Hehe. Enough bait
from me.
Cheers,
Stephen Wilkinson "Programming is like pinball.
Software Engineer the reward for doing it well
Interactive Creations Inc. is the opportunity to do it
wi...@airmail.net again." (anon - Wizards Bane)
RNDM in Warbirds http://www.icigames.com
Stephen Random Wilkinson
On 2 Jan 1997 01:50:51 GMT, spri...@eden.com (David Springer) wrote:
[snip]
>and EA are just as equally qualified as Carmack. Quake doesn't impress
>me a whole lot as being much more than an incremental improvement
>of DOOM. More hype than substance in other words. So it's higher
>resolution and you can look up and down now - big deal - one would
[snip]
OK. I'll bite.
Surely this must be the biggest piece of flame bait I've seen in this
whole thread. The Quake and Doom engines are two wildly different
beasts. Abrash gives talks around and about town here (Dallas) maybe
you should get the tape from last years CGDC and read up a bit on it.
I'll agree that others here are as qualified as Carmack but the
preceeding paragraph is a bit of a stretch.
Anyway, this is all far afield from the debate at hand.
[snip]
>D3D won the contest... it's time to accept that and move on. I'm not
>one to cry over spilt milk. I'm just as lazy as the next programmer,
>probably more so, and if you say D3D IM is gonna make me miserable
>compared to OpenGL I'm prepared to believe that since I figured
>as much 18 months ago. BUT ... there's not a damn thing I can do about
>it.
This is a bit fatalistic. Why must you be saddled w/the sorry lot
given to you by MS? Why not take up the issue with MS over it? Why not
propose alternatives -- especially when they have merit? Seems a bit
premature to me.
>Moreover, I think any attempt to slow down MS at this point will only
>make things worse. I'd rather have them fix the known problems in
>D3D then reverse course and try to replace it with OpenGL this close
>to Win 97.
Hmmm. I don't quite see your point here. Why is it bad to "slow down
MS at this point" ? Better to slow them down earlier in the process so
you have a better product in the end IMHO. Think about all the
problems that just grow worse with time (installed base of drivers,
software base, support, etc) that will be much worse if you delay by
"tweeking" D3D then reversing to OpenGL. Sounds like a bad deal to me.
Mats Grahm
David Springer wrote:
> : answer their analysis. In what way do these "some people that
> : disagree" believe that is Direct3D superior to OpenGL?
>
> For Christ's sake Mark - how many different ways can it be
> spelled out to you ? They believe D3D has better marketing
> support, for games, than OpenGL. They don't need any more
> reason than that. Microsoft is pushing D3D and that's that.
> They don't want to buck Microsoft and no one concerned with the
> bottom line can blame them.
Exactly! This is just another VHS/Betamax debate. Sure, there
still are people out there who love to debate how much better
the world would have been should Betamax had won the VCR battle.
But people actually INVOLVED in video production have stopped
grieving decades ago.
Same with D3D/OGL. Which one is "better" is, how sad it might
seem to some people, not all that relevant. The one that gets
critical mass first will probably take it all. The jury is
still out, maybe OGL can get a share of the home market too.
But if serious developers think MS will win, then they just
accept fact, learn how to use the shit, and stop treating the
subject like a religios war!
/mats
Mark Kilgard
In article <5af49r$p6f$1...@boris.eden.com>, spri...@eden.com (David Springer)
writes:
|> : In what way do these "some people that
|> : disagree" believe that is Direct3D superior to OpenGL?
|>
|> For Christ's sake Mark - how many different ways can it be
|> spelled out to you ? They believe D3D has better marketing
|> support, for games, than OpenGL.
This is your claim, but check out the reality of Microsoft's
current Direct3D marketing efforts. They've slowed considerably.
It used to be (February & March '96) that Microsoft had every PC graphics
hardware vendor releasing near simultaneous press releases saying
how great Direct3D was. The vendors didn't even know what Direct3D was
back then.
Now, the truth is coming out. The John Carmack .plan concludes "no
technical reason" for Direct3D. DirectX 4 got scrapped. You continue
to have more Web interest in OpenGL (3,600 OpenGL URLs vs 458 Direct3D
URLs in December). 3Dfx and 3Dlabs are now doing low-end OpenGL drivers.
The wide consensus is that Direct3D IM is a considerably more difficult
API to develop and debug than OpenGL. The performance limitations of
Direct3D's execute buffer scheme are now well known. Microsoft's
3D evangelist and head marketeer for Direct3D hasn't been quoted in
press releases for months now (she used to be quoted in *every*
one); rumor is she's been moved aside.
If Direct3D is such a marketing steamroller, how do you explain the
above? If the marketing is so strong, where's the evidence?
- Mark
Patrick Earl
> : I have helped a 15 year old high school kid program OpenGL on
> : his PC running OS/2.
>
> His name wasn't Brian Hook by any chance was it ? ;)
>
> readily children can program effectively with them ?
>
I think that it is a good thing if an inexperienced programmer can
easily learn how to use an API. This post was not to say that we should
judge anything based on if children can use them, but if children can
use them, that's all the better. He was merely demonstrating OpenGL's
ease of use.
Patrick Earl - pat...@freenet.edmonton.ab.ca
Michael I. Gold
Patrick Earl <pat...@freenet.edmonton.ab.ca> writes:
| He was merely demonstrating OpenGL's ease of use.
^^
Sorry to nitpick, but you're the second person to refer to Suzy as "he".
(I agree with your points about ease of use, however. ;-)
George T. Talbot
In article <5af613$pj2$2...@boris.eden.com>, spri...@eden.com (David Springer) wrote:
>Suzy Deffeyes (su...@austin.ibm.com) wrote:
>
>
>: I have helped a 15 year old high school kid program OpenGL on
>: his PC running OS/2.
>
>His name wasn't Brian Hook by any chance was it ? ;)
For Christ's sake...lighten up. Putting the ;) doesn't make it any better.
>Seriously, are you suggesting that we judge software API's by how
>readily children can program effectively with them ?
I've been watching this debate for a while...you guys have degenerated it
(the debate) into a name calling match.
______________________________________________________________________
George T. Talbot <geo...@phat.com>
<http://knucklehead.phat.com/~george>
Finger geo...@knucklehead.phat.com for PGP public key.
John Bigboote
>For Christ's sake...lighten up. Putting the ;) doesn't make it any better.
>of operations. Yeah, they do - so what ? The devil is plainly in
>the details.
> For Christ's sake Mark - how many different ways can it be
>Christ David, how low do you plan on sinking? It amazes me that you've
....so tell me, which API does Jesus recommend?
Thatcher Ulrich
Whew, I can't believe I'm taking the bait, but here goes.
David Springer wrote:
> Activision and Electronic Arts employees have posted in this thread.
>
> Pay attention.
>
> They've both indicated that they will publish with D3D so it apparently
> isn't quite as broken as Carmack, SGI, et all, would have you believe.
Maybe it's time to look at D3D from the user's perspective. I'll
pretend to be an ordinary user and give my observations on using D3D
games.
I've tried a few different D3D games recently with various
hardware/driver configurations, and it hasn't been much fun. Starting
with driver installation: The primary adapter in my computer is an S3
Virge based card with a Diamond driver. In order to run a D3D game
(even with software rendering) I have to shuffle video drivers around,
or trust an install program to do it for me. This generally causes my
Win95 desktop to be screwed up in various ways (I ordinarily like to run
at 1280x1024 with 60Hz refresh for doing most Windows stuff, but all of
the DirectX drivers I've installed so far force the display into an
unusable 43Hz interlaced mode at 1280x1024). Some games force one or
another color depth for the desktop as well -- please explain why this
is necessary for a full-screen game. I realize these driver woes are
more the fault of DirectX and Win95 brain-damage in general than D3D in
particular, but most users won't know this -- the upshot is that D3D is
perceived as a pain in the neck, reminiscent of juggling TSR's to make
enough real mode memory in the bad old days of 16-bit DOS.
So once I got the drivers correctly installed, some of the games worked
and some didn't. Activision's Hyperblade was probably the best
behaved. I tried the software renderer (worked OK, decent frame rate on
PPro200), S3 Virge driver (performed a lot like the software renderer),
and 3Dfx Voodoo (got some bonus blended textures instead of
flat-shading, plus the annoyance of bonus blurry text on my Windows
desktop due to the required VGA pass-through connector).
SegaSoft/Rocket Science's Rocket Jockey demo offered more grief. Using
the 3Dfx driver (recommended by the install program), I got a mostly
blank main-menu screen with an invisible mouse pointer (I guess some
vestiges of RJ's Myst-like-puzzle-game origins still remain in the
demo). After getting past that one, I got to the actual game,
consisting of a completely blank screen, requiring a hard reset. On to
the next driver. I'm not completely sure whether the other option
offered by the install program represented the software driver or the S3
Virge hardware driver (not like the performance is distinguishable
anyway). It worked somewhat better, but far from flawlessly. The main
menu was sometimes visible, and the 3D view in the game usually worked,
although the options overlay rarely did. Frame rate was on par with
what I'd expect from a mediocre software renderer drawing the same
imagery.
I also tried a bunch more SegaSoft demos, with a similar mixture of
success, failure and aggravation.
Overall, none of these games, even with the 3Dfx Voodoo card, rivaled
the software-only Quake for overall visual quality and playability. All
of them were far more trouble to install and configure than Quake. I
would advise my game-playing friends to buy 3D hardware and/or D3D games
only if they're willing and able to deal with annoying configuration
problems (some of them are willing and able, but that's another story).
> My *personal* opinion is that the people posting from Activision
> and EA are just as equally qualified as Carmack. Quake doesn't impress
> me a whole lot as being much more than an incremental improvement
> of DOOM. More hype than substance in other words. So it's higher
> resolution and you can look up and down now - big deal - one would
> expect some improvement considering I first saw DOOM on a 386 and I
> first saw Quake on a P166 for gawd's sake.
My *personal* opinion is that Carmack is the Michael Jordan of 3D
programmers, and if the people from Activision and EA are smart they're
listening very carefully to what he's saying. I'll also say this for Id:
they consistently produce quality, bug-free, engaging products that tend
to work as advertised, unlike the majority of game companies.
I realize that my above observations do little to recommend OpenGL, as
it would also seem susceptible to the same kinds of problems, but I
think it's worth remembering during this debate the following points (my
personal opinions only, of course):
1) The combination of existing 3D hardware with D3D drivers is not
stable and not compelling, yet, and given the perceived waning of
enthusiasm for D3D-IM in the developer community, may never be.
2) OpenGL's stability and decent conformance tests may be its decisive
advantages as an API. Ease of use and scalability would be very welcome
side benefits that would be good for the industry as a whole.
3) The traditional PC gamer market has been dealing with driver/hardware
stupidity for a long time now, and may more-or-less-happily muddle
through the current mess. However (and this is important even if you
work for Dell), it will be extremely difficult for the industry to make
inroads into new markets using this strategy.
-Thatcher Ulrich
http://world.std.com/~ulrich
Timothy Field
<snip>
>I have helped a 15 year old high school kid program OpenGL on
>his PC running OS/2.
<snip>
I do hope you're not casting aspertions on this 'kid.'
;)
>====================================================================
>Suzy Deffeyes su...@austin.ibm.com deffeyes@ausvm6
>IBM OpenGL Development (512)838-2818 I speak for myself.
>============= OpenGL: It's Everywhere. Do the Math. ================
--
Timothy Field <mailto:nob...@ihug.co.nz> ...... Wellington, New Zealand
Brandon Van Every
> Mats Grahm wrote:
> > But if serious developers think MS will win, then they just
> > accept fact, learn how to use the shit, and stop treating the
> > subject like a religios war!
Once upon a time, I thought pretty much like this. "Hey, it's just 2
different 3D API's, right? Quit your whining and learn what you need to
learn."
Well, part of the difficulty is that D3D is so damn painful to learn and
use. Not to mention to test for correctness.
When people like Brian Hook and John Carmack tell me that D3D sucks to use,
I listen to them. Brian, because he seems to be a sharp guy and has
extensive online materials to back up his claims. Carmack, because he's
widely regarded in the industry as "not being some chump." :-) The
difficulties in learning, using, and testing D3D are non-trivial, make no
mistake about it.
This is an area where OpenGL clearly demonstrates more maturity as an API.
Figure that OpenGL has been "ironed out" for 4 years now, and that it was
proceeded by IrisGL which was ironed out for many years before that.
Whereas D3D is comparable to a beta release.
Gil Colgate
David Springer wrote:
> Gil Colgate wrote:
> >
> > You should read the fine book "The Design of Everyday Things". The main
> > thing an
>
> I find most books offering insight to be less insightful than my own
> thinking.
Yes, you know everything. Noone can teach you anything.
> Just because you can find agreement in print with your
> own opinion doesn't make it right.
Contrarily, I was just recommending the book as a good read. It
discusses a
lot of details, with good pictures. It's not about computers, it's about
door handles, nuclear reactor control rooms, and programmable VCR's.
It's
one of the best books on industrial design around.
>
> Inane philosophical tripe... it needs to get the job done in the most
> straightforward manner possible. How well the job is defined is
> paramount to assessing the validity of any judgement of how well
> or poorly it succeeds in its mission.
Wrong. A good example of a bogus API is one that has a routine called
AllocRecord that allocates memory, and a routine called FreeRecord that
marks it as freeable, and another call RefreshControlBlocks, in another
file documented in another book, that actually frees it. The poor user
will probably run out of memory before he understands that he has to
call
RefreshControlBlocks to actually free the memory.
>
> I don't think there's much basic agreement on what D3D's job is except
> on a very abstract level. My personal opinion, which no one at SGI
> seems willing to accept but was commonly accepted two years ago is that
> OpenGL is a precision 3D API for mission critical applications. Games
> are neither precision nor mission critical...
I think Microsoft would have been better off writing "OpenGL-Lite"
rather
than direct 3D.
>
> If you can't agree on the mission parameters of an API then you can't
> compare judgements on how well it accomplishes the mission. QED
For writing games that consumers use, I'd use the API's from the
manufacturers and get a bundling deal.
Currently I'm on Playstation. That has a real simple API for
3D work. Lacking, but simple. It's nice to have control over the math
directly! I'd say, except for harware deficiencies (no Z buffer, for
example, no perspective correction, etc) I think the Playstation
API does a better job at providing a 3D environment for games than
Direct3D does.
Bernd Kreimeier
In article <5a8c8a$d...@fido.asd.sgi.com> m...@woodsy.asd.sgi.com (Mark Kilgard) writes:
> From: m...@woodsy.asd.sgi.com (Mark Kilgard)
> Newsgroups: rec.games.programmer
> Date: 30 Dec 1996 12:23:38 GMT
> Organization: Silicon Graphics, Inc.
> |> Someone or some group of people need to feel strongly enough about
> |> this to get together and attempt to change the direction of the MS
> |> efforts.
>
> This has been done on numerous occasions. Microsoft is well aware
> how lousy Direct3D is. Id Software's position is just a higher
> profile, public version of what Microsoft has understood for a long time.
> Then, there are those fairly "non-3D savvy" political types at
> Microsoft who have garnered a high-profile through the aquistition
> of Rendermorphics (their RealityLab API was the basis of Direct3D).
> They did what they knew how to do (generate hype). They are now reaping
> what they have sown. While the pro-Direct3D types within Microsoft
> did a splendid hype job, they can't hype over lousy technology.
Please do not overlook one important aspect of John Carmack's statement:
the ONLY issue is solving the problem
The problem being: there are two API's (D3D IM and OpenGL) for the
very same purpose, one is mature, the other is important for some
careers.
While I take some delight in MickeySoft bashing now and then (having
suffered myself from their products occasionally), this is not going
to solve anything.
Face-keeping is (unfortunately) part of any viable solution. John's
remark (keep and improve D3D retained mode, implement it on top of
MS OpenGL, dump D3D IM in favor of OpenGL) is not only a technical one.
two cents spent for 1997
b.
Toshi Morita
David Springer (spri...@eden.com) wrote:
: Mark Kilgard (m...@woodsy.asd.sgi.com) wrote:
: : In article <5a8nlv$cuk$1...@boris.eden.com>, spri...@eden.com (David Springer)
: : writes:
: : |> Mark Kilgard (m...@woodsy.asd.sgi.com) wrote:
: : |> : In article <5a4boa$cs5$1...@boris.eden.com>, spri...@eden.com (David
: : |> Springer)
: : |> : writes:
: : |> If everyone were of exactly the same opinion and as vehemently
: : |> opposed to D3D as Carmack then I'd have no doubt that D3D is
: : |> well and truly broken. However, some people disagree, if not
: : |> in principle then at least in degree
: : Honestly, who are these people? Give me and others a chance to
: Activision and Electronic Arts employees have posted in this thread.
: Pay attention.
: They've both indicated that they will publish with D3D so it apparently
: isn't quite as broken as Carmack, SGI, et all, would have you believe.
They probably haven't programmed Direct3D yet. :)
I've programmed Direct3D, and I think it's a pretty horrible API.
This is how you draw a triangle in Direct3D:
(You should take notes if you want need to program Direct3D IM - this
will save you a lot of hassle trying to figure it out)
Step 1: Initializing Direct3D
Initialize DirectDraw
Enumerate resolutions available and bit depths
Select a valid resolution and bit depth
Create a surface of the resolution and bit depth marked as a primary surface
with the the 3DSURFACE feature bit set
Create a surface of Z-buffer type with your choice of bit depth
Attach the Z-buffer to your primary surface
Initialize Direct3D
Enumerate the list of renderers available in the system
Find a renderer in the system which can accept your texture bit depth
and render to the output bit depth and also the feature bits you want.
(feature bits = Z bit depth you've used, bilinear shading, trilinear,
gouraud, point sample, perspective correct, fogging, etc)
(if you can't find one, either error out or go to step 2 or accept
something less than what you want, like maybe no perpective correct)
Get a pointer to the renderer
Step 2: Loading the texture
Grab the caps bits from the renderer and find out whether it wants
textures in system or video memory
If in video memory:
Create a surface in video memory of the correct size
Get the texture interface to the video memory surface
Get the handle for the video memory texture
Create a surface in system memory of the correct size
Get the texture interface to system memory surface
Get the handle for the system memory texture
Execute a texture load call on the video memory surface
to load the texture from system memory
else if system memory
Create the surface in system memory
Get the texture interface to the system memory surface
Get the texture handle for the system memory texture
Step 3: Rendering the triangle
Read the caps bits once again, determine the maximum number of vertices
and instructions in the ExecuteBuffer
Create an ExecuteBuffer
Lock the ExecuteBuffer
Put in a default light and render viewport for viewing the triangle
(Direct3D by default will give you no lights...der)
Put in the three vertices for the triangle
Put in a D3DOP_TEXTUREHANDLE command for selecting a texture handle
Put in a D3DOP_TRIANGLELIST command for the triangle face
Put in the face for the triangle
Set the face data offset from the start of the ExecuteBuffer
Set the number of vertices in the ExecuteBuffer
Unlock the ExecuteBuffer
Clear the Z buffer
Call BeginScene
Call Execute
Call EndScene
I'm actually glossing over a few steps here - such as setting up
the lights and the viewport - because I don't remember all the gory
details.
Anyway, this is not a joke - it takes about 700 lines of code to properly
render a single triangle on the screen. And you wondered why Microsoft
products were always so large...
It took me about four months to figure out most of this stuff,
and even then I had to fly to Microsoft for a week to find the
last few bugs (triangle vertices weren't counterclockwise and
Z-buffer wasn't cleared to correct value).
I consider myself a pretty good programmer too - I've been programming
in the games industry for ten years now.
: My *personal* opinion is that the people posting from Activision
: and EA are just as equally qualified as Carmack. Quake doesn't impress
: me a whole lot as being much more than an incremental improvement
: of DOOM. More hype than substance in other words. So it's higher
: resolution and you can look up and down now - big deal - one would
: expect some improvement considering I first saw DOOM on a 386 and I
: first saw Quake on a P166 for gawd's sake.
I think you need to play Quake.
: : answer their analysis. In what way do these "some people that
: : disagree" believe that is Direct3D superior to OpenGL?
: For Christ's sake Mark - how many different ways can it be
: spelled out to you ? They believe D3D has better marketing
: support, for games, than OpenGL. They don't need any more
: reason than that. Microsoft is pushing D3D and that's that.
: They don't want to buck Microsoft and no one concerned with the
: bottom line can blame them.
Microsoft lemmings!
: : |> 18 months ago I was arguing FOR OpenGL's inclusion in Win 95
: : |> and AGAINST some Rendermorphics derivative.
: : That's great to know. I hope that you'll review the analysis
: : from John Carmack and others today and see that your analysis
: : 18 months ago is even more true today.
: It's still true today but it's too late for it to come true. D3D
: won the contest... it's time to accept that and move on. I'm not
: one to cry over spilt milk. I'm just as lazy as the next programmer,
: probably more so, and if you say D3D IM is gonna make me miserable
: compared to OpenGL I'm prepared to believe that since I figured
: as much 18 months ago. BUT ... there's not a damn thing I can do about
: it.
Been hanging around alt.gothic?
Save this post. You'll need it later.
: Moreover, I think any attempt to slow down MS at this point will only
: make things worse. I'd rather have them fix the known problems in
: D3D then reverse course and try to replace it with OpenGL this close
: to Win 97.
Which parts need to be fixed to turn a Ford Pinto into a Porsche 911?
: David Springer
Toshi
Brandon Van Every
Toshi Morita <t...@nntp.best.com> wrote in article
<5aqp5o$5...@nntp1.best.com>...
>
> David Springer (spri...@eden.com) wrote:
> : Mark Kilgard (m...@woodsy.asd.sgi.com) wrote:
> : : In article <5a8nlv$cuk$1...@boris.eden.com>, spri...@eden.com (David
Springer)
> : : writes:
> : : |> Mark Kilgard (m...@woodsy.asd.sgi.com) wrote:
> : : |> : In article <5a4boa$cs5$1...@boris.eden.com>, spri...@eden.com
(David
> : : |> Springer)
> : : |> : writes:
>
> : : |> If everyone were of exactly the same opinion and as vehemently
> : : |> opposed to D3D as Carmack then I'd have no doubt that D3D is
> : : |> well and truly broken. However, some people disagree, if not
> : : |> in principle then at least in degree
>
> : : Honestly, who are these people? Give me and others a chance to
>
> : Activision and Electronic Arts employees have posted in this thread.
>
> : Pay attention.
I've been paying attention for many months. I've followed most "D3D vs.
OpenGL" threads, because we have a vested market interest in the issue.
I've not seen *one* person step forwards and say that they cared to work
with D3D. Quite frankly, "people don't like it." All reporting on D3D has
been uniformly bad, as far as I can tell. And some key people whom I
consider to be "not chumps" have said why they feel that way. Until
someone announces that D3D has undergone a complete redesign, that's enough
for me.
> : They've both indicated that they will publish with D3D so it apparently
> : isn't quite as broken as Carmack, SGI, et all, would have you believe.
>
> They probably haven't programmed Direct3D yet. :)
Indeed. Saying you will *support* D3D, from either the application side or
the driver side, is an entirely different comment than saying you *like*
D3D, or that D3D is a good tool to use. "Support" just means that you're
afraid of losing money if Microsoft's marketroids carry the day. People
will support D3D in the short term, because they're afraid of the economic
consequences of guessing wrong. Once they get experience with D3D, they'll
start thinking about the economic consequences of being late to market with
their products, because it's such a bear to work with. Not to mention
incompatibility issues due to the lack of a conformance suite.
Smart people are already making the technical decision now: ditch it.
Were that technology was the only decision to make. The show would already
be over. John Carmack and Id Software can afford to make very public
"ditch it" choices, because they lead the industry. Most games companies
don't have that luxury, they have to wait and see how the wind blows.
We don't really have the luxury of making choices either, but at least
we're not tremendously affected by the results.
Tony Di Croce
A friend of mine once "Broke Down" the internal operation of Microsoft
for me once and it went something like this:
1. Groups within Microsoft work separately on often competing
products.
2. Company backing is won through sales.
So the obvious answere to this debate is that Microsoft will support
whatever standard either Makes the most money Or has LOTS of potential
to make monry. Its entirely Money driven.
Just my 2cents
Tony
David Springer
Brandon Van Every (vane...@blarg.net) wrote:
: >
: > : : |> If everyone were of exactly the same opinion and as vehemently
: > : : |> well and truly broken. However, some people disagree, if not
: > : : |> in principle then at least in degree
: >
: > : : Honestly, who are these people? Give me and others a chance to
: >
: > : Activision and Electronic Arts employees have posted in this thread.
: >
: > : Pay attention.
: I've been paying attention for many months. I've followed most "D3D vs.
: OpenGL" threads, because we have a vested market interest in the issue.
: I've not seen *one* person step forwards and say that they cared to work
: with D3D. Quite frankly, "people don't like it." All reporting on D3D has
: been uniformly bad, as far as I can tell. And some key people whom I
: consider to be "not chumps" have said why they feel that way. Until
: someone announces that D3D has undergone a complete redesign, that's enough
: for me.
You didn't pay much attention to what I wrote in the first paragraph
above Brandon. Restating: Carmack's opinion is that D3D is a "broken"
API. Others have come forward and said it is difficult but they are
managing to use it effecitively nonetheless.
Please don't put words into my mouth that I didn't utter. Building
straw men to tear down is the mark of someone who is desperate to
win their argument at any cost, including deceit.
: > : They've both indicated that they will publish with D3D so it apparently
: > : isn't quite as broken as Carmack, SGI, et all, would have you believe.
: >
: > They probably haven't programmed Direct3D yet. :)
: Indeed. Saying you will *support* D3D, from either the application side or
: the driver side, is an entirely different comment than saying you *like*
: D3D, or that D3D is a good tool to use. "Support" just means that you're
: afraid of losing money if Microsoft's marketroids carry the day. People
: will support D3D in the short term, because they're afraid of the economic
: consequences of guessing wrong. Once they get experience with D3D, they'll
: start thinking about the economic consequences of being late to market with
: their products, because it's such a bear to work with. Not to mention
: incompatibility issues due to the lack of a conformance suite.
I never said those same people "like" D3D, only that they are successfully
using it and intend to publish with it. More straw men...
: Smart people are already making the technical decision now: ditch it.
Smart peoople base their decision on more than technical merits.
: Were that technology was the only decision to make. The show would already
: be over. John Carmack and Id Software can afford to make very public
: "ditch it" choices, because they lead the industry. Most games companies
: don't have that luxury, they have to wait and see how the wind blows.
John Carmack and Id software are neither Gods nor the top revenue
earners in the industry. Get a grip.
Rob Barris
In article <5b35da$agg$3...@boris.eden.com>, spri...@eden.com (David
Springer) wrote:
> Please don't put words into my mouth that I didn't utter. Building
> straw men to tear down is the mark of someone who is desperate to
> win their argument at any cost, including deceit.
> : Were that technology was the only decision to make. The show would already
> : be over. John Carmack and Id Software can afford to make very public
> : "ditch it" choices, because they lead the industry. Most games companies
> : don't have that luxury, they have to wait and see how the wind blows.
>
> John Carmack and Id software are neither Gods nor the top revenue
> earners in the industry. Get a grip.
Please don't put words into people's mouths that they didn't utter. I
don't believe the words "gods" or "top revenue earners" were used, only
"lead".
Id's work is *highly* influential. Who in their right mind would disagree?
Rob Barris
Quicksilver Software Inc.
rba...@quicksilver.com
* Opinions expressed not necessarily those of my employer *
John Slagel
David Springer wrote:
>> John Carmack and Id Software can afford to make very public
>> "ditch it" choices, because they lead the industry. Most games
>> companies don't have that luxury, they have to wait and see how
>> the wind blows.
> John Carmack and Id software are neither Gods nor the top revenue
> earners in the industry. Get a grip.
But, they do have enough clout to set standards... For example,
since they used DOS4GW in Quake, other game companies began using
it also without fear also, and other extenders such as PharLap
were basically ignored by the entire game industry. VESA 2.0
and DJGPP are more examples of them adding credibility by
way of a huge installed base to some game development products.
So if they are really going to not use D3D and only use OpenGL,
this is going to hurt D3D in a big way, like it or not. Or else
some customer is going to write a fast OpenGL layer on top of
D3D for them, and the D3D vs OpenGL problem will be solved for
everyone.
--
-John
Glenn Mandelkern
In article <32CC6E...@student.hk-r.se>,
Mats Grahm <pt9...@student.hk-r.se> wrote:
>David Springer wrote:
>
>> : disagree" believe that is Direct3D superior to OpenGL?
>>
>> For Christ's sake Mark - how many different ways can it be
>> spelled out to you ? They believe D3D has better marketing
>> support, for games, than OpenGL. They don't need any more
>> reason than that. Microsoft is pushing D3D and that's that.
>> They don't want to buck Microsoft and no one concerned with the
>> bottom line can blame them.
>
>still are people out there who love to debate how much better
>the world would have been should Betamax had won the VCR battle.
>But people actually INVOLVED in video production have stopped
>grieving decades ago.
>
Or 80x86 vs 680x0.
>Same with D3D/OGL. Which one is "better" is, how sad it might
>seem to some people, not all that relevant. The one that gets
>critical mass first will probably take it all. The jury is
>still out, maybe OGL can get a share of the home market too.
>But if serious developers think MS will win, then they just
>accept fact, learn how to use the shit, and stop treating the
>subject like a religios war!
>
We have this argument here all the time, degreed vs non-degreed, etc.
Incidentally, on another thread within this newsgroup about interpreting
games as an art form that may or may not get recognized, I brought up
the point about two types of workers named "Sustainers" and "Achievers."
I happened to have this debate of OpenGL vs. Direct3D on my mind so
I made a connection there. Basically, "Sustainers" are people who believe
like in school that all do is have something that is technically superior
and the world will beat a path to their door. It's like handing in a paper
and waiting to be graded. "Achievers", on the other hand, do the work and
realize one has to do a lot more to get recognized and promoted.
(Incidentally, that's why some companies like to snag young college graduates
who are still operating in "Sustainer" mode, ready made for sweatshops.)
Some may wish to view Microsoft marketing as a monster, yet others view it
as something very inspirational -- you may have the best mousetrap in the
world, but if your competition lets many others know about one that works
85% of the time, that may indeed be what gets them your reward. If the
customer who does the grading feels that 85% is just right, although the
"Sustainer" may sob about lowering standards, meeting a bottom line with
some promises generates the "Achiever" revenue today and tomorrow.
Another book that covers this is "Straight A's Never Made Anybody Rich."
Although I used to lament how some of the technically superior solutions
would make life easier for some of us if they were just advertised better,
more assertively, etc., in another way, now that the technically inferior
product has made so much headway, it does provide another opportunity.
All the time for cleanup has a matching dollar -- are you going to sulk
about it, saying how it is not "fair", or are you going to find a way to say,
"Hey, I can actually make some money because of this sanitary software
engineering!"? I'm not even talking about planned obsolescence.
Making sense out of chaos is actually a new form of lucrative business
in the 1990's. I see it all the time when a solution that appears that
may outlast another one is addressed by one savvy vendor who says (inspired
by one 4-letter word in the above poster), "You may choose to hold on to
Product A which may be great, but its support seems uncertain. I specialize
in reformulating doodoo. Forecasts show Product B lasting much longer,
possibly taking over some of Product A's marketshare. I realize a forecast
does not hold guarantees, but I think about your present AND future costs.
Yes, this processing of Product B's doodoo may cost you more upfront, but it
will be an investment that you may feel confident about for the next 3 to 5
years. Product B's manufacturer has even recognized the problem and has
a new department specifically dedicated to reducing doodoo. This newsletter
graph shows the decline of Product B's doodoo over the past 3 years. I helped
people deal with worse doodoo back then." This vendor will now get a
reputation of being the Profitable Product B Doodoo Transformer.
Is it manipulative to make a buck out of chaos? A Sustainer would say so,
an Achiever finds how to do it ethically. A Sustainer may appoint himself
judge and accuse the Achiever of making use of other people's body parts
when the Achiever is just using his and other people's brains (street smarts?)
I once heard one quality expert say that the problem with the quality
movement is that it is possible to do so well a job that you actually
reengineer yourself out of it. I do not want to promote mediocrity,
but if we made the games and other tools so complete, we would destroy
our own adjacent market of add-on's and plug-in's. In terms of API's,
we always see "extensions." Regardless, someone is always unsatisfied
and somone else sees how to make a new market out of another market's
shortcomings. So actually, all this complaining is great, as a form
of release and a form of monetary capture.
--
---------------------------------------------------------------------------
Glenn Mandelkern "Hee, hee, hee, hee!" -- Questor the Elf
gma...@netcom.com "When passion runs deep,
San Jose, CA you're playing for keeps" -- Keith Emerson
Games, GUI's and Entertainment What does Motif sound like in the key of C++?
Andreas Maurer
bigb...@yoyodyneps.com (John Bigboote) wrote:
>
>
>>For Christ's sake...lighten up. Putting the ;) doesn't make it any better.
>
>>of operations. Yeah, they do - so what ? The devil is plainly in
>>the details.
>
>> For Christ's sake Mark - how many different ways can it be
>
>>Christ David, how low do you plan on sinking? It amazes me that you've
>
>
>....so tell me, which API does Jesus recommend?
>
AFAIK the most realistic threedimesional Liveforms can be created
using his fathers LPL (Liveform-Programming-Language), DNA.
It is known to be rather low level, though. ;)
James Shaw
Gavin Bell <ga...@mailbag.com> wrote:
>Terry Sikes wrote:
>> I'm curious on this point. What kinds of optimizations (or the
>> opposite ;-) make a card work well under Direct3D and poorly under
>> OpenGL? Would anyone care to comment on which of the low-cost 3D
>> chipsets would be likely to work better with Direct3D?
Probably none make it work more optimally. Optimizations for Direct3D
would stem almost totally from faster rendering, provided you can
supply polygons quickly enough!
>Boards that don't have hardware z-buffers might work better. Microsoft
>is warning developers to be SURE to call the BeginScene/EndScene D3D-IM
>API calls (even though they don't do anything on a lot of current
>boards), I assume because boards that sort or split the polygons in each
>frame will be coming out.
This is almost certainly to deal with Talisman, which requires all the
scene data to be available before rendering (at least that's how the
specs read). EndScene() then becomes a kind of 'commit' function to
tell the hardware all the scene data is now ready, go and do it.
> I think that one of the great weaknesses of
>D3D is that it will force game developers to create games that all
>pretty much look the same, because the lowest common denonimator of
>features that you can reasonably expect to work on all implementations
>is way too low.
The other way to look at this statement is to say that D3D will force
hardware manufacturers to at least support many of the features. Any
board not supporting all of
gouraud/textured/zbuffering/transparencies/bilinear-filtering/fog in
hi-res, hi-colour or true colour will fall by the wayside. Besides,
do all games look the same now? No. The look of a game comes from
graphic artists, it's the speed of a game that will be affected by
hardware acceleration, notwithstanding the API that's used to get that
speed. The end user could hardly care less if I used OpenGL, Direct3D
or any other API - that's our job. They care about look and feel (and
whizzy graphics :).
Jim.
Brandon Van Every
> In article <5b35da$agg$3...@boris.eden.com>, spri...@eden.com (David
> Springer) wrote:
>
> > : Were that technology was the only decision to make. The show would
already
> > : "ditch it" choices, because they lead the industry. Most games
companies
> > : don't have that luxury, they have to wait and see how the wind blows.
> >
> > John Carmack and Id software are neither Gods nor the top revenue
> > earners in the industry. Get a grip.
When your game company makes the cover of WIRED magazine, then we'll talk.
;-)
David Springer
John Slagel (jo...@volition-inc.com) wrote:
: David Springer wrote:
: >> John Carmack and Id Software can afford to make very public
: >> "ditch it" choices, because they lead the industry. Most games
: >> companies don't have that luxury, they have to wait and see how
: >> the wind blows.
: > John Carmack and Id software are neither Gods nor the top revenue
: > earners in the industry. Get a grip.
: But, they do have enough clout to set standards... For example,
: since they used DOS4GW in Quake, other game companies began using
Piffle - merely a matter of being in the right place at the right
time. IBM used to set a LOT of standards in the PC industry for
many years and look what happened there... will you now try to
compare Id to IBM ?
How about Hercules ? There's a nice one time wonder for comparison
purposes.
: So if they are really going to not use D3D and only use OpenGL,
: this is going to hurt D3D in a big way, like it or not. Or else
: some customer is going to write a fast OpenGL layer on top of
: D3D for them, and the D3D vs OpenGL problem will be solved for
: everyone.
We shall see. I disagree with your conclustion. It's speculation on
both our parts right now.
David Springer
Rob Barris (rba...@quicksilver.com) wrote:
: In article <5b35da$agg$3...@boris.eden.com>, spri...@eden.com (David
: Springer) wrote:
: Id's work is *highly* influential. Who in their right mind would disagree?
I would certainly argue that Electronic Arts or Activision is just as
*highly* influential. Believe it or not the entire world does not
revolve around Quake and there are many programmers the equal of Carmack
who've been less fortunate in having been in the right place at the
right time.
Who in their right mind would disagree with that ?
David Springer
Brandon Van Every (vane...@blarg.net) wrote:
: > In article <5b35da$agg$3...@boris.eden.com>, spri...@eden.com (David
: > Springer) wrote:
: >
: already
: > > : be over. John Carmack and Id Software can afford to make very public
: > > : "ditch it" choices, because they lead the industry. Most games
: companies
: > > : don't have that luxury, they have to wait and see how the wind blows.
: > >
: > > John Carmack and Id software are neither Gods nor the top revenue
: > > earners in the industry. Get a grip.
: When your game company makes the cover of WIRED magazine, then we'll talk.
: ;-)
What does my startup have to do with Id ? Let's compare my employer
and where I have the larger personal stake - Dell Computer. When Id
surpasses Dell to become the world's best performing stock (see Wall
Street Journal - no disrespect to Wired) some year - then we'll talk.
When Id goes from nothing to $1 billion in sales in three years like
the notebook division of Dell has done since I've joined it then
we'll talk.
As long as Id is little more than a glitzy shareware producer whose
pricetag ($0) did more for winning a place in the hearts of game
players than anything else - we won't talk...
David Springer
Laptop BIOS Programmer
Dell Computer Corporation
David Springer
Andreas Maurer (ama...@sime.com) wrote:
: bigb...@yoyodyneps.com (John Bigboote) wrote:
Yes, but it's simple and fun to "program" with... provided that you
avoid the "dating and relationship" extensions. Those can very
complex...
David Springer
Pete Olson
David Springer (spri...@eden.com) wrote:
:
: What does my startup have to do with Id ? Let's compare my employer
: and where I have the larger personal stake - Dell Computer. When Id
: surpasses Dell to become the world's best performing stock (see Wall
: Street Journal - no disrespect to Wired) some year - then we'll talk.
Id is privately held, and plans to stay that way.
: When Id goes from nothing to $1 billion in sales in three years like
: the notebook division of Dell has done since I've joined it then
: we'll talk.
Id is privately held, so you won't know.
: As long as Id is little more than a glitzy shareware producer whose
: pricetag ($0) did more for winning a place in the hearts of game
: players than anything else - we won't talk...
Hmm, yeah, I guess I can't consider MS as serious on
browsers as they give theirs away.
Clinton Keith
Brandon Van Every <vane...@blarg.net> wrote in article <01bbfed3$3214d040$1c90...@hammurabi.blarg.net>...
> > In article <5b35da$agg$3...@boris.eden.com>, spri...@eden.com (David
> > Springer) wrote:
> >
> > > John Carmack and Id software are neither Gods nor the top revenue
> > > earners in the industry. Get a grip.
>
> When your game company makes the cover of WIRED magazine, then we'll talk.
> ;-)
The same magazine that put Rocket Science on the cover? ;) ;)
--
Clinton Keith
Angel Studios
(W) cl...@angel.com
(H) clinto...@worldnet.att.net