There is a limited pool of *good* engineers in the Bay Area. There are plenty of companies out here with a bunch of bad engineers, there are very few engineers who are really good at what they are doing.
I have no trouble thinking that the pool of good engineers out here is only 500 - 1000 big.
Jamie Zawinski has a list of rude words which had to be removed from the Netscape client code before it could be open-sourced. Microsoft's probably looks a bit like this.
So how is this architecture so revolutionary and amazing when compared to the processor in the Cray X1? (Which has a MIPS [like PPC, but without broken IO] core, and multiple vector units configured in much the same way as this seems to be).
In the United Kingdom you do *not* own the copyright of your face if you're in a public place (which is why papparazzi are allowed to sell the photos they take for money). This has been tested in court.
More recently there have been rules set out to discourage photographers from stalking famous people, but the basic law still stands.
It's way simpler to implement a windowsystem in software than on silicon.
It's also not like driving 3D chips is impossible. Normally you have acceleration for doing the regular 2D stuff (fill, copy and line with the standard 256 raster ops). If you want to do something like accelerated alpha blit then it's normally just drawing two textured triangles with the 3D engine (the hardest part of that is engine setup).
The hard stuff is in texture memory "VM" design (i.e.: efficiently using texture memory when allocated textures [or windows] are growing/shrinking).
I would be interested in seeing the numbers for MS Office on Mac. I have personally found development on Mac to be _much_ more expensive (longer development cycles, poor api documentation) than on either Windows or X11.
Remembering that the Mac represents less than 2% of market share, and even less in annual sales I would be very suprised if MS Office on Mac is profitable or even break even for Microsoft.
Re:They're improving the file dialogs...
on
GTK 2.6.0 Released
·
· Score: 1
Composite will make things seem smoother, because of the way things get drawn in off-screen memory before being blitted on screen.
Composite doesn't really have much to do with drop shadows, per se. It allows windows to be moved off screen and for a compositing agent to draw them onto the screen as it likes.
Most (all?) drop shadow implementations currently use the RENDER extension to draw the shadows. All drivers (except the NVidia binary ones) implement RENDER acceleration through the XAA (X Acceleration Architecture) hooks. Currently XAA only really accelerates RENDER if you're compositing text (so xterm with AA fonts is faster). There aren't any hooks for "Blend this region onto the screen" yet -- it has been rumoured that those will be added soon.
Note that the Xserver with KDrive DDX does have hooks for doing blending and compositing onto the screen, and thus can draw drop shadows (amongst other things) fast on supported cards... It doesn't support GLX (or a bunch of other useful things like XvMC) though.
Actually, VIA (who own S3) were very nice to me. I told them that I wanted to write an X driver for their graphics chip (the CLE266 northbridge with integrated graphics). They sent me an NDA and then the register documentation.
And they did actually already write their own driver which was released as opensource (although I'm not sure of the license) for XFree86 including all of the "GL stuff".
IMHO S3/VIA are very appreciative of opensource work and are very supportive of opensource developers.
SGI had a product called "MineSet" which did this kind of stuff, only a long long time ago. Originally it was inspired by the 3D filemanager SGI did for Jurassic Park. Cool idea, but old hat:).
The whole environment is in a single X window with a GLXContext & GLXWindow associated.
Other X windows appear in there because (AFAIK) sun have written an OpenGL output driver for the Kdrive DDX (in English: sun wrote a device driver for a small X server which draws using OpenGL, rather than writing to registers on a graphics chip). They then used this to draw X windows into the 3D environment.
So sure, you could write some 3D user interfaces if you didn't go through X (using whatever API they've written to do 3D stuff in lookingglass instead).
I'd be suprised if OpenGL in X windows which existed in the 3d environment existed at all (given that GLX doesn't yet work on kdrive, and given the poor OpenGL pbuffer support on Linux systems).
A few years back they put them on London buses, to catch people who drove in bus lanes. The main problem was determining when the bus was in a bus lane.
demonstrates an incredible ignorance of the technical reasons why Windows requires no single hardware manufacturer. This amazing technical feat has nothing to do with Microsoft, and more to do with IBM and the use of a central (reverse-engineered) bios, and Intel's ubiquity and the reverse-engineering of their instruction set. You would think that the chairman of AMD would realize this.
Actually, I guess that Sanders wasn't talking about CPUs there. They make networking boards, too. So I imagine that he was talking about how difficult it would be to write device drivers (tough enough already on Windows) for multiple distributions of the OS.
> its claims of 56k transfer rates are highly optimistic.
From ARM.com: The secret to the Pogo's hi-speed wireless web connection lies in its unique proprietary compression technology. It is powered by an ARM7 core, which was selected for its excellent code density and speed.
There's always SGI's OpenGL Shader, an evaluation of which is available for free for Linux 2.2, and IRIX 6.5 (currently at IRIX 6.5.14). Get it here: http://www.sgi.com/software/shader/
The MIPS architecture still has a long way to go. SGI have recently opened up a development centre in Boston to continue work on the R10000 upwards chips, whilst MIPS continue developing all other MIPS cores.
Even if you don't have a lovely O2 sitting on your desk, your playstation(1|2) and N64 are both MIPS machines.
Slashdot Mirror
There is a limited pool of *good* engineers in the Bay Area. There are plenty of companies out here with a bunch of bad engineers, there are very few engineers who are really good at what they are doing.
I have no trouble thinking that the pool of good engineers out here is only 500 - 1000 big.
The article says that the northbride is a CN400, but the photos have a CLE266 northbridge on... What's up with that?
--marmite
Jamie Zawinski has a list of rude words which had to be removed from the Netscape client code before it could be open-sourced. Microsoft's probably looks a bit like this.
http://www.jwz.org/doc/censorzilla.html
So how is this architecture so revolutionary and amazing when compared to the processor in the Cray X1? (Which has a MIPS [like PPC, but without broken IO] core, and multiple vector units configured in much the same way as this seems to be).
In the United Kingdom you do *not* own the copyright of your face if you're in a public place (which is why papparazzi are allowed to sell the photos they take for money). This has been tested in court.
More recently there have been rules set out to discourage photographers from stalking famous people, but the basic law still stands.
It's way simpler to implement a windowsystem in software than on silicon.
It's also not like driving 3D chips is impossible. Normally you have acceleration for doing the regular 2D stuff (fill, copy and line with the standard 256 raster ops). If you want to do something like accelerated alpha blit then it's normally just drawing two textured triangles with the 3D engine (the hardest part of that is engine setup).
The hard stuff is in texture memory "VM" design (i.e.: efficiently using texture memory when allocated textures [or windows] are growing/shrinking).
Who cares about WiFi on digital cameras?
I want GPS coordinates in an exif tag (or something).
I would be interested in seeing the numbers for MS Office on Mac. I have personally found development on Mac to be _much_ more expensive (longer development cycles, poor api documentation) than on either Windows or X11.
Remembering that the Mac represents less than 2% of market share, and even less in annual sales I would be very suprised if MS Office on Mac is profitable or even break even for Microsoft.
Composite will make things seem smoother, because of the way things get drawn in off-screen memory before being blitted on screen.
Composite doesn't really have much to do with drop shadows, per se. It allows windows to be moved off screen and for a compositing agent to draw them onto the screen as it likes.
Most (all?) drop shadow implementations currently use the RENDER extension to draw the shadows. All drivers (except the NVidia binary ones) implement RENDER acceleration through the XAA (X Acceleration Architecture) hooks. Currently XAA only really accelerates RENDER if you're compositing text (so xterm with AA fonts is faster). There aren't any hooks for "Blend this region onto the screen" yet -- it has been rumoured that those will be added soon.
Note that the Xserver with KDrive DDX does have hooks for doing blending and compositing onto the screen, and thus can draw drop shadows (amongst other things) fast on supported cards... It doesn't support GLX (or a bunch of other useful things like XvMC) though.
--ralpht
Actually, VIA (who own S3) were very nice to me. I told them that I wanted to write an X driver for their graphics chip (the CLE266 northbridge with integrated graphics). They sent me an NDA and then the register documentation.
And they did actually already write their own driver which was released as opensource (although I'm not sure of the license) for XFree86 including all of the "GL stuff".
IMHO S3/VIA are very appreciative of opensource work and are very supportive of opensource developers.
SGI had a product called "MineSet" which did this kind of stuff, only a long long time ago. Originally it was inspired by the 3D filemanager SGI did for Jurassic Park. Cool idea, but old hat :).
--ralpht
The whole environment is in a single X window with a GLXContext & GLXWindow associated.
Other X windows appear in there because (AFAIK) sun have written an OpenGL output driver for the Kdrive DDX (in English: sun wrote a device driver for a small X server which draws using OpenGL, rather than writing to registers on a graphics chip). They then used this to draw X windows into the 3D environment.
So sure, you could write some 3D user interfaces if you didn't go through X (using whatever API they've written to do 3D stuff in lookingglass instead).
I'd be suprised if OpenGL in X windows which existed in the 3d environment existed at all (given that GLX doesn't yet work on kdrive, and given the poor OpenGL pbuffer support on Linux systems).
--ralpht
The chip has a full speed FPU. More information on the AES support can be found in this interview, previously reported on slashdot...
l
http://linuxdevices.com/articles/AT2656883479.htm
--ralpht
I think that the Intergraph ZX10 had 2 AGP slots on. Sadly SGI bought the product line and destroyed it.
--ralpht
A few years back they put them on London buses, to catch people who drove in bus lanes. The main problem was determining when the bus was in a bus lane.
Taxi cameras are a good idea, though.
There are several secure closed source operating systems (Trusted...). Don't just think as far as windows wrt closed-source operating systems.
No it's not.
ROX-Filer is written in C. There are various extras which are written in python.
demonstrates an incredible ignorance of the technical reasons why Windows requires no single hardware manufacturer. This amazing technical feat has nothing to do with Microsoft, and more to do with IBM and the use of a central (reverse-engineered) bios, and Intel's ubiquity and the reverse-engineering of their instruction set. You would think that the chairman of AMD would realize this.
Actually, I guess that Sanders wasn't talking about CPUs there. They make networking boards, too. So I imagine that he was talking about how difficult it would be to write device drivers (tough enough already on Windows) for multiple distributions of the OS.
Ralph.
SGI are still designing new MIPS chips, and recently opened a new microprocessor design office in Boston.
> its claims of 56k transfer rates are highly optimistic.
From ARM.com:
The secret to the Pogo's hi-speed wireless web connection lies in its unique proprietary compression technology. It is powered by an ARM7 core, which was selected for its excellent code density and speed.
marmite
If you check MS RTF documentation, you will see that .doc is actually just the binary format of RTF.
There's always SGI's OpenGL Shader, an evaluation of which is available for free for Linux 2.2, and IRIX 6.5 (currently at IRIX 6.5.14). Get it here:
http://www.sgi.com/software/shader/
The MIPS architecture still has a long way to go. SGI have recently opened up a development centre in Boston to continue work on the R10000 upwards chips, whilst MIPS continue developing all other MIPS cores.
Even if you don't have a lovely O2 sitting on your desk, your playstation(1|2) and N64 are both MIPS machines.