I dealt with SGI at the academic level. In the mid 1990s, their hardware was unbelievable. Even their mid-range machine, the Indigo2 IMPACT had far better 3D than just about anything else on the market. Their Onyx RealityEngine2 was almost an order of magnitude better and was quickly replaced by InfiniteReality which was far better yet. The stuff we worked on was horrendously expensive, but probably worth every penny.
By the late 1990s their upgrade cycle came to a stop for awhile. The CPUs we'd been told were in the pipline were suddenly canceled or put on hold. The next huge graphics update was also halted. Meanwhile their sales drones tried to see us snazzy but overpriced and underperforming PCs and PC servers. At least the LCD for the PCs was wicked cool (and worked on O2, yay... but why not Octane??) WTF? A year later, they came back to us trying to see us an even crappier line of totally generic PCs. Models that weren't even compatible with their neat widescreen LCD without the use of a $500 adapter. Double-WTF?? Shortly after I heard they bought Intergraph but halted future development. Triple-WTF???
I think they finally realized the PC thing was silly and started work on their traditional iron again. But this time the updates came slow and were not all that large. Instead of a totally new CPU, they just made the old model faster. Instead of totally new graphics, they just did a 25% speed boost and doubled the RAM. Granted it's hard to complain about 1024 MB texture RAM and 10240 frame buffer RAM (yes, 10 GB). But was still based on original 1995 design and OpenGL 1.1.
One of the last straws for us is when they started to honk the Linux horn again, telling us how much better thier Altix was over the Origin, even though they had been pimping the Origin 3000 to us just a few months earlier. Their new graphics engine was now just a bunch of V12 cards ripped out of Octanes and grafted into an Origin. A little later they began using ATI FireGL cards.
SGI has been slipping downhill ever since about 1997. As time goes on, they slide faster and make even more bizzare moves. They always seem to have at least one really impressive piece of kit, but these days they bounce around so much that they don't keep a customer for very long. I just recently heard about some impressive sales numbers for the (old but fast) InfiniteReality4 graphics to gov't vizsim clients (simulators and such) and to Discreet for Inferno HD/Film editing and effects workstations. Yet I also heard that these will be discontinued in a few months.
These days SGI likes to jump from one bandwagon to the next and they really love to tell their customers what they should want and what they should buy. It might have worked back when all of their gear was top shelf, but it sure doesn't work anymore.
Maybe they'll stick with Altix long enough to regain some ground. Time will tell...
If you talk to any of the animators from Star Wars Episode 1 you'll hear them bitch and moan about having to use SGI O2 workstations. They were running commerical apps like Maya and Softimage, but also ILM's own compositing and animation software on what was basiclly SGI's very lowest end machine. Many even had very low end CPUs that wern't upgraded until long into the project. Meanwhile over at Pixar, they were doing much simpler animation on Octane/Octane2 systems, machines that were an easy 2x faster for CPU (4x faster if you did dual proc) and an easy 4x faster for 3D. A far better fit.
BZFlag is a very fun game. It's not pretty, but then, it's about 13 years old.I have a great deal of fun with bolo, nethack, and xpilot as well and their graphics are even more dated!
The original release of Xandros was Corel Linux. They spruced up a Debian-based distro with a custom configuration of KDE with a few extra utilitites and a nice set of custom icons. They even ported a few of the Windows apps. Corel Draw suite and Corel Word Perfect were both available for Corel Linux. WP was a native port, while Draw used WINE libraries, but was still an impressive piece of work.
Though it was unrelated, anyone remember the Corel Netwider?
I have personal biases, and I think we all do. But the fact remains, the best choice is to use the right tool for the job.
Mac diehards will flame me, but a Mac is not always the best tool.
If you live or work in a MS world, then no amount of utilities and features will ever make your Mac fit perfectly. However, if you only have to grab mail and swap calendar data from an exchange server and open/edit files from Office, then perhaps a Mac will be OK. If you need to access SMB shares and printers, Mac OS X should work fine out of the box. If your office has a screwy setup then you might need to use Dave. Then again, it might also be configured in such a way that your Mac may never be able to connect. A Mac isn't a Windows PC, so there's no guarantee.
If you need to work on a "pure" Unix system, you should be OK. It's POSIX compliant and has pretty much every OS feature found on FreeBSD (with some bits from OpenBSD) though implementation may be different.
If you need to use a Windows-only application, you could probably use Virtual PC, but performance will suck. Badly. Even on an overclocked 9 GHz G7.
You need to figure out what exactly your needs are and pick the right hardware and software for the job. I have friends and coworkers that use 8-year-old Silicon Graphics (SGI) workstations for both work and home use with great success!! I have friends who are highly talented software engineers that use nothing but Wnidows. I've had a college professor who does GIS R&D on a second generation iMac. There are all types.
As for me, these days I develop software for internal company use. My platform of choice is Linux, though I still run a few Solaris servers for various projects. I also have a company-issue ThinkPad running Win2K for the corp stuff, but it gets little use. At home I have a variety of machines... new, old, and ancient. Win, Mac, Linux, Unix, Prodos.:) The only Mac I use these days is my PowerBook G4 I use for most of my personal web surfing, personal email, digital photography, digital home videos, and other light duty experimenting. It works like a charm and I love it. But I also love my other toys, and it's a PC and a PS2 that I play my games on. I don't bother running anything other than Enigmo, SimCity 4, and UT2004 on the PowerBook.
This has been addressed in several interviews with Win Game -> Mac porting houses. It seems that while raw OpenGL performance is about the same between a Mac and a similar PC, the differences come in the game design / optimization and other libraries used. Many coding decisions made in the original game were done with the Windows / x86 crowd in mind. Even many PC "OpenGL" games still use other DirectX libraries for sound, user I/O, and other tasks. Even some very basic considerations, such as data file read-in, is done in little endian. Differences like this are almost nill in a small academic test case, but are significant in a complex game. Virtual machines, such as those used for AI in games like Quake3 are also optimized for x86.
For an example of a "pretty good" port, try the latest Mac build of UT2004. It has a few more tweaks and makes uses of a newer version of OpenAL. It's still slower than the native Windows version, but it's not too bad.
How does Firefox prevent them from patching Windows software? It doesn't. It's just an excuse for lazy MCSE admins who don't want to add an additional step to their daily advisory-reading / patch-installing cycle.
My point is this: in an established MS shop, it's often very hard to get the admins to approve usage of non-MS software. At my previous job we had many people using MS Publisher and that MS photo suite when InDesign and Photoshop would have been far better for their needs.
I'm not agreeing with the original poster's admin, I'm just saying that MS shops are often set in their ways.
Just install it anyway. There's no way that they can tell you're using it, unless they're looking over your shoulder. That kind of attitude will get you fired. Management is edgy these days and support/admin money is tight. There just isn't room for someone who doesn't want to go along with the flow. It's not 1998 anymore. The Aeron chairs and the foosball table have been auctioned off and there are many other people just waiting to take your job. Seriously. I've seen several people canned in 2004 by doing things "their own way" despite being told not to.
I worked in an all-Windows shop for awhile. It wasn't too bad and the network and server admins were *very* tuned into the security notices from Microsoft. They would have every machine patched within one business day of the announcement. Maybe your company is the same way, and introducting non-Microsoft software may upset that cycle.
Does Cray still have its Museum in Chippewa Falls, Wisconsin? I know these days a lot of the Cray buildings are now used by Silicon Graphics and Celestia, but I think Cray still does a fair amount of work in CF.
There are recycling methods for the spent fuel rod assemblies. There's still going to be a significant amount of waste left over, but at least it will reduce the amount of waste that needs to be processed and burried. The French have been recycling portions of their nuclear waste for years. That practice has been outlawed in the USA since the Carter administration to keep the waste away from "the terrorists".
Yucca Mountain is not the end-all, save-all solution. There is already such a huge backlog of waste to be send down there that the current massive tunnel system will be filled up all the waste generated by the year 2013. As soon as work finishes on Yucca Mtn, they're going to have to start drilling another complex elsewhere.
Some of the "newer", safer reactors, such as the Pebble Bed type the Chinese are starting to use, actually produce far less waste than the 1960s style reactors used in the USA. If we update our reactors, we will produce less waste. But then we're still going to have to find a resting place for the old reactors.
Nuclear might be our only hope for gigawatt-scale power production by the end of the century. At the very least it's going to employ many engineers to work on solutions to these problems.
I'm no Mac guy, but isn't that G4 choked by a 133Mhz FSB while the P4 is running 800Mhz(200x4)?
Yep.
But the 1.5 Ghz P4 cited in the parent post would only have a 400 Mhz FSB. I don't think even 533 Mhz FSB came into play until the 2.0 Ghz P4.
This is one of the reasons why many G4 systems had L3 cache, sort of a stop-gap measure between Apple and Motorola to improve performance a tad. Thankfully IBM has much higher FSB on the G5. The dual 2.5 Ghz G5 system has a dedicated 1.25 Ghz bus for *each* CPU. Compare this to a dual or quad Xeon, where all of the CPUs share the same bus. This is one of the reasons why you'll the G5 so far ahead on RAW image conversion and BLAST bioinfometric applications.
Those were the days! I remember my old 21" trinitron had BNC in, but I've never seen a PC equipped with BNC out. I've never seen BNC on a PC graphics card, or even that many worstations. More common were DB15 and HD15 "VGA" connectors with three (sync on green), four (composite sync), or five (seperate H and V sync) coax cables, each with a BNC connector.
Even better, and much more handy is 13w3, a single D shell with 3 mini-coax connectors and 10 data lines. Most unix machines, including Silicon Graphics worstations used this connector for years. Even some PC cards used it. SGI only recently dropped 13w3 when they moved to DVI. http://www.foxns.com/cables/1302.jpg
The guys at Best Buy are worse. They'll just spout off nonsense The best example of this was when Best Buy was selling the original blue iMac. I thought I had heard it all until I overheard a sales goon tell a potential customer, "Bill Gates had a virus on his network, the only way he could remove it was by adding an iMac".
I don't like how I cannot use other resolutions than its native resolution. The image looks weird especially with pixels. I also don't want those black sides. I use resolutions from 800x600 to 1280x1024 depending on games. That's why Barco still sells CRT projectors, some of their clients demand that sort of contrast ratio. Even some home theater buffs shell out the money for the HT version from Runco. (A modified Barco projector and a monster image processor box, unmatched quality, but you need a fairly dark room).
I don't like how I cannot use other resolutions than its native resolution. The image looks weird especially with pixels. I also don't want those black sides. I use resolutions from 800x600 to 1280x1024 depending on games.
One solution is to get a 19" LCD with a native 1600x1200 resolution and a graphics card with plenty of fillrate. For the simpler games, you can run at native resolution. For more demanding games, you can run pixel-doubled at 800x600.
Of course if you're going to get a beefy card, you may as well just get a 17" LCD and run the game at the display's native 1280x1024...
I've played 3D shoot-em-up games on active matrix TFT-LCDs since the high end models hit 40ms (13ms rise, 27ms fall). I noticed some "ghosting" on those LCDs, just as I notice "ghosting" on the latest wiz-bang LCDs. The Apple Studio Display and Silicon Graphics 1600SW were both excellent LCD monitors for DTP, photo, and games back in 1998.
I think there are other issues that make LCDs a turn-off to some:
Not used to LCDs. If you've been a CRT user for more than a solid year, an LCD will seem strange at first. It's hard to describe, but the image just looks a bit different, a bit, "strange".
Pixel density & fixed pixels vs multisync CRT phosphors. While there is really no perfect display for using a variety of resolutions, CRTs are still better than LCDs in this respect. A high quality 19" CRT will display 1024x768 just as nice as it will display 1280x1024 and 1600x1200. Most modern LCDs have very good interpolation circuitry to display non-native resolutions, but you can still spot the fuzzyness.
Adapting to a poor video signal. If you're using an LCD, use DVI, period. LCDs tend to be much less tolerant of a poor analog video signal (typically from a cheap gamer card and/or a poor vga cable). What looks nearly perfect on a CRT might be blurry or even flicker little "sparks" of dropped pixels on a LCD, even at native resolution.
Pixel density. I've noticed this most often on 18.1" LCDs. At 1280x1024, the native resolution for most 18" LCDs, I can usually notice the "screendoor" grid that outlines the pixels. Even with subpixel rendering (which to my eyes reminds me of the awful Apple II "almost-white" text on a color monitor days) and/or various amounts of anti-aliasing, the onscreen text just doesn't look right at first.
That said, I think monitor preference is generally based on one's time with a certain tech, be it LCD or CRT. Both can produce good quality images, both can be adjusted for color correctness. Buy whatever fits best on your desk (and/or budget) and spend a month with it.
Desktop LCD displays may never be fast enough for shutter glasses stereo (3D). However there are many LCD displays with glasses-free stereo features made by companies like DDD, ThreeDeep, and Deep Video Imaging. The biggest problem with most of these is that they tend to be designed for one user. Plus the models that actually work (as opposed to just being a gimmick) are usually very expensive.
From what I gather, OLED monitors will be fast enough for traditional shutter glasses and will also support Z Screen bi-polarization filters.
Nah, USB is so 1998. Ahh, the days when you could count the number of popular USB devices on one hand... Umax 1220U scanner, Epson 740 inkjet printer, Iomega Zip250...
It seemed like every USB "early adopter" had the exact same suite of USB peripherals back in late 1998 / early 1999.
Just how simple is your device? Will you be able to use existing chipsets or microcontrollers? Even a very simple ASIC can quickly cost you $175,000 in engineering and die costs. Low-gate FPGAs may be an option, but the larger models will drain too much current to run off USB power alone, especially if connected to a bus-powered hub.
Even the plastic enclosure for the device will cost you significant money. Expect to pay for CAD work, molds, and factory setup.
Then there's the USB vendor ID registration, device driver, packaging, sales, support...
I'm not surprised Apple chose #3. Now, why Apple didn't design the iPod photo to download photos directly from a digital camera via a USB 2.0 cable, that's another question entirely...
USB Host support is much more complex than just being a USB device. Plus they would have to deal with multiple protocols and even some device drivers. Just look at the size of the code that makes up libgphoto!
That being said, there is the hardware issue. Although I cannot confirm with OS 9 or previous, but OS X was built on Unix (correct me if I am wrong).
"Classic" Mac OS was originally based on UCSD p-System, but has been overhauled numerous times since 1982. System 6 was a huge overhaul, as was 7. The last *massive* under-the-hood updates were Mac OS 7.6, 8.5, and 8.6. Updates leading up to 9.2.2 were mostly to the various support frameworks, hardware support/optimizations, GUI look-and-feel, and other features of the OS, not so much kernel work.
And yes, Mac OS X (10.x) are based on Unix. As with NeXT OPENSTEP, it's a Mach microkernel with a BSD userland and lots of custom frameworks and libraries. GUI is based on Quartz, which is heavily PDF based (but is not Display Postscript nor "Display PDF").
Or maybe I should ask, does anyone still use AFP in *new* installations? We have a mix of Win/Unix/Mac(OSX). The Unix/Linux workstations and Macs automount several servers via NFS when a user logs in. The Win PCs use a Samba server (ugh). More "important" data is sent via scp. Telnet and ftp have been pretty much abandoned.
I thought that AFP was only used to support old legacy Macs running 9.2.2 or older.
Granted there are NFS clients for Windows and for "Classic" Mac OS 9.2.2 and earlier, but most are pretty ugly.
Slashdot Mirror
I dealt with SGI at the academic level. In the mid 1990s, their hardware was unbelievable. Even their mid-range machine, the Indigo2 IMPACT had far better 3D than just about anything else on the market. Their Onyx RealityEngine2 was almost an order of magnitude better and was quickly replaced by InfiniteReality which was far better yet. The stuff we worked on was horrendously expensive, but probably worth every penny.
By the late 1990s their upgrade cycle came to a stop for awhile. The CPUs we'd been told were in the pipline were suddenly canceled or put on hold. The next huge graphics update was also halted. Meanwhile their sales drones tried to see us snazzy but overpriced and underperforming PCs and PC servers. At least the LCD for the PCs was wicked cool (and worked on O2, yay... but why not Octane??) WTF? A year later, they came back to us trying to see us an even crappier line of totally generic PCs. Models that weren't even compatible with their neat widescreen LCD without the use of a $500 adapter. Double-WTF?? Shortly after I heard they bought Intergraph but halted future development. Triple-WTF???
I think they finally realized the PC thing was silly and started work on their traditional iron again. But this time the updates came slow and were not all that large. Instead of a totally new CPU, they just made the old model faster. Instead of totally new graphics, they just did a 25% speed boost and doubled the RAM. Granted it's hard to complain about 1024 MB texture RAM and 10240 frame buffer RAM (yes, 10 GB). But was still based on original 1995 design and OpenGL 1.1.
One of the last straws for us is when they started to honk the Linux horn again, telling us how much better thier Altix was over the Origin, even though they had been pimping the Origin 3000 to us just a few months earlier. Their new graphics engine was now just a bunch of V12 cards ripped out of Octanes and grafted into an Origin. A little later they began using ATI FireGL cards.
SGI has been slipping downhill ever since about 1997. As time goes on, they slide faster and make even more bizzare moves. They always seem to have at least one really impressive piece of kit, but these days they bounce around so much that they don't keep a customer for very long. I just recently heard about some impressive sales numbers for the (old but fast) InfiniteReality4 graphics to gov't vizsim clients (simulators and such) and to Discreet for Inferno HD/Film editing and effects workstations. Yet I also heard that these will be discontinued in a few months.
These days SGI likes to jump from one bandwagon to the next and they really love to tell their customers what they should want and what they should buy. It might have worked back when all of their gear was top shelf, but it sure doesn't work anymore.
Maybe they'll stick with Altix long enough to regain some ground. Time will tell...
If you talk to any of the animators from Star Wars Episode 1 you'll hear them bitch and moan about having to use SGI O2 workstations. They were running commerical apps like Maya and Softimage, but also ILM's own compositing and animation software on what was basiclly SGI's very lowest end machine. Many even had very low end CPUs that wern't upgraded until long into the project. Meanwhile over at Pixar, they were doing much simpler animation on Octane/Octane2 systems, machines that were an easy 2x faster for CPU (4x faster if you did dual proc) and an easy 4x faster for 3D. A far better fit.
BZFlag is a very fun game. It's not pretty, but then, it's about 13 years old.I have a great deal of fun with bolo, nethack, and xpilot as well and their graphics are even more dated!
The original release of Xandros was Corel Linux. They spruced up a Debian-based distro with a custom configuration of KDE with a few extra utilitites and a nice set of custom icons. They even ported a few of the Windows apps. Corel Draw suite and Corel Word Perfect were both available for Corel Linux. WP was a native port, while Draw used WINE libraries, but was still an impressive piece of work.
Though it was unrelated, anyone remember the Corel Netwider?
I have personal biases, and I think we all do. But the fact remains, the best choice is to use the right tool for the job.
:) The only Mac I use these days is my PowerBook G4 I use for most of my personal web surfing, personal email, digital photography, digital home videos, and other light duty experimenting. It works like a charm and I love it. But I also love my other toys, and it's a PC and a PS2 that I play my games on. I don't bother running anything other than Enigmo, SimCity 4, and UT2004 on the PowerBook.
Mac diehards will flame me, but a Mac is not always the best tool.
If you live or work in a MS world, then no amount of utilities and features will ever make your Mac fit perfectly. However, if you only have to grab mail and swap calendar data from an exchange server and open/edit files from Office, then perhaps a Mac will be OK. If you need to access SMB shares and printers, Mac OS X should work fine out of the box. If your office has a screwy setup then you might need to use Dave. Then again, it might also be configured in such a way that your Mac may never be able to connect. A Mac isn't a Windows PC, so there's no guarantee.
If you need to work on a "pure" Unix system, you should be OK. It's POSIX compliant and has pretty much every OS feature found on FreeBSD (with some bits from OpenBSD) though implementation may be different.
If you need to use a Windows-only application, you could probably use Virtual PC, but performance will suck. Badly. Even on an overclocked 9 GHz G7.
You need to figure out what exactly your needs are and pick the right hardware and software for the job. I have friends and coworkers that use 8-year-old Silicon Graphics (SGI) workstations for both work and home use with great success!! I have friends who are highly talented software engineers that use nothing but Wnidows. I've had a college professor who does GIS R&D on a second generation iMac. There are all types.
As for me, these days I develop software for internal company use. My platform of choice is Linux, though I still run a few Solaris servers for various projects. I also have a company-issue ThinkPad running Win2K for the corp stuff, but it gets little use. At home I have a variety of machines... new, old, and ancient. Win, Mac, Linux, Unix, Prodos.
This has been addressed in several interviews with Win Game -> Mac porting houses. It seems that while raw OpenGL performance is about the same between a Mac and a similar PC, the differences come in the game design / optimization and other libraries used. Many coding decisions made in the original game were done with the Windows / x86 crowd in mind. Even many PC "OpenGL" games still use other DirectX libraries for sound, user I/O, and other tasks. Even some very basic considerations, such as data file read-in, is done in little endian. Differences like this are almost nill in a small academic test case, but are significant in a complex game. Virtual machines, such as those used for AI in games like Quake3 are also optimized for x86.
For an example of a "pretty good" port, try the latest Mac build of UT2004. It has a few more tweaks and makes uses of a newer version of OpenAL. It's still slower than the native Windows version, but it's not too bad.
That's good, they aren't making any more unix-based apps from what I've heard. What kind of silly statement is that?
How about this site:
http://sourceforge.net/softwaremap/trove_list.php
How does Firefox prevent them from patching Windows software?
It doesn't. It's just an excuse for lazy MCSE admins who don't want to add an additional step to their daily advisory-reading / patch-installing cycle.
My point is this: in an established MS shop, it's often very hard to get the admins to approve usage of non-MS software. At my previous job we had many people using MS Publisher and that MS photo suite when InDesign and Photoshop would have been far better for their needs.
I'm not agreeing with the original poster's admin, I'm just saying that MS shops are often set in their ways.
Just install it anyway. There's no way that they can tell you're using it, unless they're looking over your shoulder.
That kind of attitude will get you fired. Management is edgy these days and support/admin money is tight. There just isn't room for someone who doesn't want to go along with the flow. It's not 1998 anymore. The Aeron chairs and the foosball table have been auctioned off and there are many other people just waiting to take your job. Seriously. I've seen several people canned in 2004 by doing things "their own way" despite being told not to.
I worked in an all-Windows shop for awhile. It wasn't too bad and the network and server admins were *very* tuned into the security notices from Microsoft. They would have every machine patched within one business day of the announcement. Maybe your company is the same way, and introducting non-Microsoft software may upset that cycle.
Yes indeed, Tubes Rock!
Does Cray still have its Museum in Chippewa Falls, Wisconsin? I know these days a lot of the Cray buildings are now used by Silicon Graphics and Celestia, but I think Cray still does a fair amount of work in CF.
There are recycling methods for the spent fuel rod assemblies. There's still going to be a significant amount of waste left over, but at least it will reduce the amount of waste that needs to be processed and burried. The French have been recycling portions of their nuclear waste for years. That practice has been outlawed in the USA since the Carter administration to keep the waste away from "the terrorists".
Yucca Mountain is not the end-all, save-all solution. There is already such a huge backlog of waste to be send down there that the current massive tunnel system will be filled up all the waste generated by the year 2013. As soon as work finishes on Yucca Mtn, they're going to have to start drilling another complex elsewhere.
Some of the "newer", safer reactors, such as the Pebble Bed type the Chinese are starting to use, actually produce far less waste than the 1960s style reactors used in the USA. If we update our reactors, we will produce less waste. But then we're still going to have to find a resting place for the old reactors.
Nuclear might be our only hope for gigawatt-scale power production by the end of the century. At the very least it's going to employ many engineers to work on solutions to these problems.
I'm no Mac guy, but isn't that G4 choked by a 133Mhz FSB while the P4 is running 800Mhz(200x4)?
Yep.
But the 1.5 Ghz P4 cited in the parent post would only have a 400 Mhz FSB. I don't think even 533 Mhz FSB came into play until the 2.0 Ghz P4.
This is one of the reasons why many G4 systems had L3 cache, sort of a stop-gap measure between Apple and Motorola to improve performance a tad. Thankfully IBM has much higher FSB on the G5. The dual 2.5 Ghz G5 system has a dedicated 1.25 Ghz bus for *each* CPU. Compare this to a dual or quad Xeon, where all of the CPUs share the same bus. This is one of the reasons why you'll the G5 so far ahead on RAW image conversion and BLAST bioinfometric applications.
Those were the days! I remember my old 21" trinitron had BNC in, but I've never seen a PC equipped with BNC out.
I've never seen BNC on a PC graphics card, or even that many worstations. More common were DB15 and HD15 "VGA" connectors with three (sync on green), four (composite sync), or five (seperate H and V sync) coax cables, each with a BNC connector.
Even better, and much more handy is 13w3, a single D shell with 3 mini-coax connectors and 10 data lines. Most unix machines, including Silicon Graphics worstations used this connector for years. Even some PC cards used it. SGI only recently dropped 13w3 when they moved to DVI.
http://www.foxns.com/cables/1302.jpg
The guys at Best Buy are worse. They'll just spout off nonsense
The best example of this was when Best Buy was selling the original blue iMac. I thought I had heard it all until I overheard a sales goon tell a potential customer, "Bill Gates had a virus on his network, the only way he could remove it was by adding an iMac".
Wow.
I don't like how I cannot use other resolutions than its native resolution. The image looks weird especially with pixels. I also don't want those black sides. I use resolutions from 800x600 to 1280x1024 depending on games.
That's why Barco still sells CRT projectors, some of their clients demand that sort of contrast ratio. Even some home theater buffs shell out the money for the HT version from Runco. (A modified Barco projector and a monster image processor box, unmatched quality, but you need a fairly dark room).
I don't like how I cannot use other resolutions than its native resolution. The image looks weird especially with pixels. I also don't want those black sides. I use resolutions from 800x600 to 1280x1024 depending on games.
One solution is to get a 19" LCD with a native 1600x1200 resolution and a graphics card with plenty of fillrate. For the simpler games, you can run at native resolution. For more demanding games, you can run pixel-doubled at 800x600.
Of course if you're going to get a beefy card, you may as well just get a 17" LCD and run the game at the display's native 1280x1024...
I've played 3D shoot-em-up games on active matrix TFT-LCDs since the high end models hit 40ms (13ms rise, 27ms fall). I noticed some "ghosting" on those LCDs, just as I notice "ghosting" on the latest wiz-bang LCDs. The Apple Studio Display and Silicon Graphics 1600SW were both excellent LCD monitors for DTP, photo, and games back in 1998.
I think there are other issues that make LCDs a turn-off to some:
Not used to LCDs. If you've been a CRT user for more than a solid year, an LCD will seem strange at first. It's hard to describe, but the image just looks a bit different, a bit, "strange".
Pixel density & fixed pixels vs multisync CRT phosphors. While there is really no perfect display for using a variety of resolutions, CRTs are still better than LCDs in this respect. A high quality 19" CRT will display 1024x768 just as nice as it will display 1280x1024 and 1600x1200. Most modern LCDs have very good interpolation circuitry to display non-native resolutions, but you can still spot the fuzzyness.
Adapting to a poor video signal. If you're using an LCD, use DVI, period. LCDs tend to be much less tolerant of a poor analog video signal (typically from a cheap gamer card and/or a poor vga cable). What looks nearly perfect on a CRT might be blurry or even flicker little "sparks" of dropped pixels on a LCD, even at native resolution.
Pixel density. I've noticed this most often on 18.1" LCDs. At 1280x1024, the native resolution for most 18" LCDs, I can usually notice the "screendoor" grid that outlines the pixels. Even with subpixel rendering (which to my eyes reminds me of the awful Apple II "almost-white" text on a color monitor days) and/or various amounts of anti-aliasing, the onscreen text just doesn't look right at first.
That said, I think monitor preference is generally based on one's time with a certain tech, be it LCD or CRT. Both can produce good quality images, both can be adjusted for color correctness. Buy whatever fits best on your desk (and/or budget) and spend a month with it.
Desktop LCD displays may never be fast enough for shutter glasses stereo (3D). However there are many LCD displays with glasses-free stereo features made by companies like DDD, ThreeDeep, and Deep Video Imaging. The biggest problem with most of these is that they tend to be designed for one user. Plus the models that actually work (as opposed to just being a gimmick) are usually very expensive.
From what I gather, OLED monitors will be fast enough for traditional shutter glasses and will also support Z Screen bi-polarization filters.
USB is so 2002.
Nah, USB is so 1998. Ahh, the days when you could count the number of popular USB devices on one hand... Umax 1220U scanner, Epson 740 inkjet printer, Iomega Zip250...
It seemed like every USB "early adopter" had the exact same suite of USB peripherals back in late 1998 / early 1999.
Not to rain on your parade, but...
Just how simple is your device? Will you be able to use existing chipsets or microcontrollers? Even a very simple ASIC can quickly cost you $175,000 in engineering and die costs. Low-gate FPGAs may be an option, but the larger models will drain too much current to run off USB power alone, especially if connected to a bus-powered hub.
Even the plastic enclosure for the device will cost you significant money. Expect to pay for CAD work, molds, and factory setup.
Then there's the USB vendor ID registration, device driver, packaging, sales, support...
I'm not surprised Apple chose #3. Now, why Apple didn't design the iPod photo to download photos directly from a digital camera via a USB 2.0 cable, that's another question entirely...
USB Host support is much more complex than just being a USB device. Plus they would have to deal with multiple protocols and even some device drivers. Just look at the size of the code that makes up libgphoto!
That being said, there is the hardware issue. Although I cannot confirm with OS 9 or previous, but OS X was built on Unix (correct me if I am wrong).
"Classic" Mac OS was originally based on UCSD p-System, but has been overhauled numerous times since 1982. System 6 was a huge overhaul, as was 7. The last *massive* under-the-hood updates were Mac OS 7.6, 8.5, and 8.6. Updates leading up to 9.2.2 were mostly to the various support frameworks, hardware support/optimizations, GUI look-and-feel, and other features of the OS, not so much kernel work.
And yes, Mac OS X (10.x) are based on Unix. As with NeXT OPENSTEP, it's a Mach microkernel with a BSD userland and lots of custom frameworks and libraries. GUI is based on Quartz, which is heavily PDF based (but is not Display Postscript nor "Display PDF").
Or maybe I should ask, does anyone still use AFP in *new* installations? We have a mix of Win/Unix/Mac(OSX). The Unix/Linux workstations and Macs automount several servers via NFS when a user logs in. The Win PCs use a Samba server (ugh). More "important" data is sent via scp. Telnet and ftp have been pretty much abandoned.
I thought that AFP was only used to support old legacy Macs running 9.2.2 or older.
Granted there are NFS clients for Windows and for "Classic" Mac OS 9.2.2 and earlier, but most are pretty ugly.