Now get out of here before I whip ya with this here cable with BNC connectors.
For 1337-speakers that may have never seen those... they were big pieces of METAL on the ends of network cables.
none of those sissy plastic phone-jack "snagless" wires in the olds days. These things were physically keyed. If you tugged on the cable hard enough, the thing you were most likely to do was pull the wire out of the connector. If that didn't happen, then you're probably dragging your computer along the floor.
While I'm being silly about network cables... where the fuck did snagless connectors come from and why are they a good thing? As my arthritis gets progressively worse, I find myself loathing those things more and more.
I'm a technology consultant with over 5 years of experience, this experience is split between work for companies owned by the same umbrella corporation, and companies that are entirely outside of our corporate infrastructure.
Don't assume that your boss, or the people giving you the work, know your situation. Never say no; that's not what you're paid for. But point out technology and resource limitations that make requests unreasonable. Phrase it in terms of impact. If someone asks you to do something, "Ok, I can do that... but requests x, y, and z have priority over that, so it'll be n [days|weeks|months] before I can get to it" Or suggest an alternative like "Sure, I could take that word document from HR and convert it to HTML... but did you know that Office can automatically save to HTML format now?"
The short of it is, it's their money and their time; they can choose to do what they want with you while you're on the clock. Just make sure that the decision makers are aware of your situation, and how their decisions impact the work that you do.
This is the great thing about concerts. I for one, hope this device never sees widespread use. It could ruin the whole concert experience.
RTA, man. An audio engineer in the article estimates that about half the bands out there are already using this technology. It's already in widespread use.
The only change between the first and second cases was that X was more preferred by 2 voters. Because of the additional support, X lost.
Ahh, but that is not the point of an instant runoff. The point of an instant runoff is to eliminate matters of plurality. The question to be answered is not "Do you prefer X, Y, or Z?" The question to be answered is "Out of two candidates, which do you prefer more?"
So in the first election, because more people liked either x or y than like z, the race ended up being "Out of X or Y, which do you prefer?" In the second election, because more people liked either x or z than like y, the race ended up being "Out of X or Z, which do you prefer?"
And in each case, the vote cast by each ballot is what that ballot's caster would have preferred.
However, I can't think to too many switchable x-ray sources other than a fission reaction which off course will cause its own problems...
A Z-pinch is pretty good. You couldn't make a pistol with one, but you could make a rifle...
Basically you take a ring of fine wires, put enough amperage through them to vaporize them, and when the magnetic field pulse the plasma streamers together, they emit in higher and higher wavelengths.
http://plasma-gate.weizmann.ac.il/ZP/
The only problem is repeatability... you basically need a new z-pinch for every shot. But that's no big deal... it is very switchable...
>Note: The information transmitted in this Notice is intended only for the >person or entity to which it is addressed and may contain confidential >and/or privileged material. Any review, reproduction, retransmission, >dissemination or other use of, or taking of any action in reliance upon, >this information by persons or entities other than the intended recipient >is prohibited.
What is this all about? They are trying to hide the fact that they are sending out these letters?
Standard boilerplate legalise. Every email I send out gets this automatically tacked on by our e-mail system. This exact phrasing, IIRC...
It may surprise you Mr. Physics Dude but not only do people actually know something about physics, but there are actually hobbyists doing this kind of thing.
I've made holograms. And I've built interferometers, beam tables, and several other pieces of equipment. I've even measured the coherency fringe length of several lasers.
coherency fringe length directly effects sharpness of a hologram with regard to beam path difference. That is, if your beam path difference is equal to your coherency fringe, then you will have a hologram with 0 sharpness, no information in it. If your difference is 0, you will have a hologram as sharp as your recording medium allows.
When you cut a hologram in half you reduce that sharpness, the amount of information in the hologram. You can measure this loss with an interferometer, with a similar setup that is used to measure coherency fringe length...
Oh and posting anonymously is for cowards, choads, and morons that call other slashdotters idiots without providing factual arguments
Does it help if the USB drive in question was given to you free by Microsoft?
Microsoft sent me a couple of these to me a few months ago, unasked and unexpected. A ton of my co workers got them too. 32MB. They came preloaded with the Microsoft Partner Readiness toolkit.
no kidding.
I'm scared to plug it into my Linux box. I think the first time I do it'll just explode or something.
Well, from http://nths.newtrier.k12.il.us/academics/math/Conn ections/light/hologrsn.htm
Amazingly, a hologram may be cut in half and you will still see the entire image. And you can cut one of the pieces in half again and again and see the entire image. Every part of the hologram has received and recorded light from the entire object!
Now how does it work? Basically, a film hologram is more than just a fancy trick you are playing with light; the entire piece of film is involved in recreating the image. Every piece of film holds all of the information about the object and they all contribute to the whole. By reducing the size of the film; you simply reduce the size of the object.
A better explanation: http://www.emergentmind.org/miller-webbI3b.htm
oh and as for other neat treaks with film holograms? Let's say you have a complicated, fussy optics array. You spend days tuning it and getting it perfectly in focus. Now let's say this optics array is going to be used to focus the aiming laser on an abrams tank. Or as the projection lens for the periscope in a submarine.
You COULD ruggedize the whole setup, and field tune it occasionally.
But you could also use holography; you take a hologram of the entire setup, and the hologram acts exactly like the original optics, up to the resolution of the hologram, at the wavelength you used. You can make what's called a whitelight or broadband hologram.
I don't know why the eyeglasses companies haven't latched onto this. Cheap, light eyeglasses that don't need to be ground or anything. Just cut out the shape to fit the eyepiece, patch it in, and go.
So can we extrapolate how big our Visa cards would be if they didn't chop the dove down to giblet size?
Were it a film holograph, with the right equipment you could; in that laser's have a characteristic coherency, and holograms can only be produced by one of about half a dozen different types of lasers, simply because you have to manufacture the film specifically for the process, and the film only gets cheap in massive quantities. So you could peel the hologram off, or setup an optical front surface mirror arrangement allowing you to shine different types of laser light through and measure the change in coherency.
How does that help? Slicing the image down changed the coherency of the interference pattern. If you shine the original type of laser through it and measure the change in coherency with an interferometer, you can measure the amount of information lost; that corresponds exactly to the change in area
However, the dove on the VISA card is almost certainly a mechanical hologram, not a film hologram. Basically, because atoms are significantly smaller than photons, you can take a hologram and create a metal stamp based on it, and then use that stamp to create a holographic impression on a plastic or metal foil. It's a cheap way to mass produce holograms. But you lose a lot of the coherency information, which is why mass produced holograms are always virtual images (they appear to be inside of, rather than outside of, the hologram), because virtual images are more forgiving of slight errors in the process.
Yeah, but these holograms are projected from a reasonably small screen, so they probably use stereoscopic separation to produce the holographic keypad.
Actually, the patent as written doesn't work. The patent indicates they are using traditional film holography, which cannot be projected from a small screen. For a discussion about the subject, see this post.
I'm not saying this display would be extremely useable for a one-eyed person; I know from experience it would not. While I am still blessed with two working eyes, they have significantly different attributes and my vision without corrective optics is almost entirely two dimensional. All the same, I can see the image, which the original poster averred would not be the case.
Poppy cock. Of course you can see a hologram with only one eye!
You just can't see a stereoscopic vision allowing you to definitely position the object in three dimensional space. But the eyes use other cues than stereoscopic vision to determine position, cues like parallax and brightness, as well as ocular focus.
HoloTouch, images of keypads can be any size, entirely independent of the size of the hardware.
(emphasis my own)
From the patent:
When a hologram is illuminated by a reconstruction beam, it produces a real image (which appears to be between the plane of the hologram and the viewer) and a virtual image (which appears to be behind the plane of the hologram). [snip] Thus, it is preferred that the holographic image 207 be a real image.
Quick review of holography: an extremely high resolution film takes pictures of the interference pattern generated when a coherent light beam strikes an object.
When coherent light of a similar wavelength later shines through this film, the interference patterns cause it to be shined through in exactly the same manner as the original coherent light, up to about half the resolution of the film. Most holographic film is 3000 lines per inch, so the hologram has a "resolution" of about 1500 lines per inch.
You see an image because the light reaching your eyes through the film is exactly as it would be had the object been in front of your eyes and illuminated by the original beam.
The light reaching your eyes is coming through the film and then traveling in a straight line from the film to your eyes. You can only see such light if the holograph is directly behind it, because the path of the photons cannot change after it passes through the hologram (disregarding minor lensing effects due to the atmosphere, that is)
What does this mean? Well if the hologram appears to be one half meter in front of you and the holographic film is one meter in front of you, and the holographic image appears to be 10 cm x 10 cm, then the minimum possible size for the holographic film is 20 cm x 20 cm.
I don't call that entirely independent; as a matter of fact, it's a pretty simple relationship governed by a version of the inverse square law.
Oh, an interesting fact about it is if you take a holographic film and cut it in half, because all the information about the image is stored throughout the film, you don't have half a hologram; you have a hologram of the entire object that is half the size of the original. Pretty cool stuff actually.
I've been wanting to have a Linux firewall that boots from CD (with no HD) for security reasons... script it to reboot every night a 3am, and you could be pretty confident in it not being cracked.
look at any of the excellent Linux Firewall on a floppy projects. Any one of them could be burned to CD.
There is nothing special for Lindows. There's maybe half a dozen steps required to modify a typical distro to boot from CD. Stuff like setting up the startup scripts not to remount the root partition as read write, setting up a ramdisk, and putting the home directories on the ramdisk.
dyne:bolic (www.dynebolic.org) is another distro that does this, and they are fully open source, if you want to check it out.
Just an aside: The POS and MP3 systems would both be better served via PXE boot...
Since in each case you require a central server, you don't even need an optical drive in the POS/MP3 system. Make a suitably small boot iamge (10-50 MB is good) and PXE is decently fast. Plus you don't have to make coasters every time you update the OS
If the mirror receives heat energy from the Sun and converts some of this into free energy, namely the kinetic energy of its motion, it falls into the strict definition of a heat engine, and Carnot's rule defining the maximum efficiency for this energy conversion must apply.
Good thing the mirror doesn't convert heat energy into kinetic energy, or we'd be in trouble!
The mirror converts the momentum of electromagnetic particles into it's own momentum. A Carnot style heat engine is one that derives it's energy from the movement of heat from one portion of a system to another. Steam turbines are an excellent example of a carnot heat engine. A solar sail does not work on that principle at all.
I suppose next he'll publish a paper claiming that these http://www.lonezone.com/2000/catalog/lz888.html will never work either...
did it occur to you to try a, you know, contemporary linux distribution?
Did it occur to you to try a, you know, actual helpful response?
I mean seriously give me some pointers? WTF do you mean by contemporary? What's a contemporary linux distro? All the distros I tried were packaged after the board was released, if that's what you mean. And there's only so much fucking around that I'm willing to do before I just give up and install windows...
Keep in mind that CF cards are only rated for 100,000 writes, usually.
They are not a good replacement for a hard disk. Especially if your operating system is gonna put a pagefile on it. In one test case, the MTBF was 1 month.
That's about 3,000 page swaps a day. Not unrealistic considering that these mini computers are usually underpowered on RAM...
E-mail me for details or questions or pictures (James {dot} McCracken {at} stratapult {dot} com), but here you go:
VIA M10000 - 1GHz Nehemiah processor (good enough) - $150 512 MB RAM (hell 256 MB is probably enough) - $???80??? HDD (size and type doesn't really matter) - $70 Video- Two options here:
1. Get a PCI video card. I know it seems archaic but they still make them and these have decent enough performance compared to having to buy a shuttle.
2. Get a PCI-AGP converter (www.mini-itx.com has them) and a half-height AGP card. Again non-ideal.
Either option is gonna cost you about the same but option 2 actual gives you more options for performance than option 1. You'll get good enough performance for most non directx-9 games with this configuration. If you splurge and get the ATI A-I-W (there's a 9000 radeon PCI version for under $100), then you can turn this machine into a TiVo for free (no additional parts required) when you're not gaming.
Morex Cubid 3677 ($60) is a REAL nice looking case and you could mod a handle into it... or for the same price, get the power supply from the Morex line of cases, and get some acrylic to custom make your own case.
Totla cost: about $460. If you can scrounge some of the parts you'll get off cheaper, of course. I went without a video card, which still gives you good enough performance... 2000 3dmark2001 score... I can play AVP2 with a lot of the eye candy turned off and the res set down to 640x480... and I ended up building it all for $250.
Slashdot Mirror
I thought BNC were positive keying coax cables, not "F-Type Coaxial Connector"
IIRC Cable TV cables are threaded; BNC are keyed...
Have you never seen a TV cable????
:D
Unless you mean the cable that my internet comes on, I don't think I know what you're talking about
Now get out of here before I whip ya with this here cable with BNC connectors.
For 1337-speakers that may have never seen those... they were big pieces of METAL on the ends of network cables.
none of those sissy plastic phone-jack "snagless" wires in the olds days. These things were physically keyed. If you tugged on the cable hard enough, the thing you were most likely to do was pull the wire out of the connector. If that didn't happen, then you're probably dragging your computer along the floor.
While I'm being silly about network cables... where the fuck did snagless connectors come from and why are they a good thing? As my arthritis gets progressively worse, I find myself loathing those things more and more.
I'm a technology consultant with over 5 years of experience, this experience is split between work for companies owned by the same umbrella corporation, and companies that are entirely outside of our corporate infrastructure.
Don't assume that your boss, or the people giving you the work, know your situation. Never say no; that's not what you're paid for. But point out technology and resource limitations that make requests unreasonable. Phrase it in terms of impact. If someone asks you to do something, "Ok, I can do that... but requests x, y, and z have priority over that, so it'll be n [days|weeks|months] before I can get to it" Or suggest an alternative like "Sure, I could take that word document from HR and convert it to HTML... but did you know that Office can automatically save to HTML format now?"
The short of it is, it's their money and their time; they can choose to do what they want with you while you're on the clock. Just make sure that the decision makers are aware of your situation, and how their decisions impact the work that you do.
This is the great thing about concerts. I for one, hope this device never sees widespread use. It could ruin the whole concert experience.
RTA, man. An audio engineer in the article estimates that about half the bands out there are already using this technology. It's already in widespread use.
The only change between the first and second cases was that X was more preferred by 2 voters. Because of the additional support, X lost.
Ahh, but that is not the point of an instant runoff. The point of an instant runoff is to eliminate matters of plurality. The question to be answered is not "Do you prefer X, Y, or Z?" The question to be answered is "Out of two candidates, which do you prefer more?"
So in the first election, because more people liked either x or y than like z, the race ended up being "Out of X or Y, which do you prefer?" In the second election, because more people liked either x or z than like y, the race ended up being "Out of X or Z, which do you prefer?"
And in each case, the vote cast by each ballot is what that ballot's caster would have preferred.
However, I can't think to too many switchable x-ray sources other than a fission reaction which off course will cause its own problems...
A Z-pinch is pretty good. You couldn't make a pistol with one, but you could make a rifle...
Basically you take a ring of fine wires, put enough amperage through them to vaporize them, and when the magnetic field pulse the plasma streamers together, they emit in higher and higher wavelengths.
http://plasma-gate.weizmann.ac.il/ZP/
The only problem is repeatability... you basically need a new z-pinch for every shot. But that's no big deal... it is very switchable...
>Note: The information transmitted in this Notice is intended only for the >person or entity to which it is addressed and may contain confidential >and/or privileged material. Any review, reproduction, retransmission, >dissemination or other use of, or taking of any action in reliance upon, >this information by persons or entities other than the intended recipient >is prohibited.
What is this all about? They are trying to hide the fact that they are sending out these letters?
Standard boilerplate legalise. Every email I send out gets this automatically tacked on by our e-mail system. This exact phrasing, IIRC...
It may surprise you Mr. Physics Dude but not only do people actually know something about physics, but there are actually hobbyists doing this kind of thing.
I've made holograms. And I've built interferometers, beam tables, and several other pieces of equipment. I've even measured the coherency fringe length of several lasers.
coherency fringe length directly effects sharpness of a hologram with regard to beam path difference. That is, if your beam path difference is equal to your coherency fringe, then you will have a hologram with 0 sharpness, no information in it. If your difference is 0, you will have a hologram as sharp as your recording medium allows.
When you cut a hologram in half you reduce that sharpness, the amount of information in the hologram. You can measure this loss with an interferometer, with a similar setup that is used to measure coherency fringe length...
Oh and posting anonymously is for cowards, choads, and morons that call other slashdotters idiots without providing factual arguments
Does it help if the USB drive in question was given to you free by Microsoft?
Microsoft sent me a couple of these to me a few months ago, unasked and unexpected. A ton of my co workers got them too. 32MB. They came preloaded with the Microsoft Partner Readiness toolkit.
no kidding.
I'm scared to plug it into my Linux box. I think the first time I do it'll just explode or something.
Well, from http://nths.newtrier.k12.il.us/academics/math/Conn ections/light/hologrsn.htm
Amazingly, a hologram may be cut in half and you will still see the entire image. And you can cut one of the pieces in half again and again and see the entire image. Every part of the hologram has received and recorded light from the entire object!
Now how does it work? Basically, a film hologram is more than just a fancy trick you are playing with light; the entire piece of film is involved in recreating the image. Every piece of film holds all of the information about the object and they all contribute to the whole. By reducing the size of the film; you simply reduce the size of the object.
A better explanation: http://www.emergentmind.org/miller-webbI3b.htm
oh and as for other neat treaks with film holograms? Let's say you have a complicated, fussy optics array. You spend days tuning it and getting it perfectly in focus. Now let's say this optics array is going to be used to focus the aiming laser on an abrams tank. Or as the projection lens for the periscope in a submarine.
You COULD ruggedize the whole setup, and field tune it occasionally.
But you could also use holography; you take a hologram of the entire setup, and the hologram acts exactly like the original optics, up to the resolution of the hologram, at the wavelength you used. You can make what's called a whitelight or broadband hologram.
I don't know why the eyeglasses companies haven't latched onto this. Cheap, light eyeglasses that don't need to be ground or anything. Just cut out the shape to fit the eyepiece, patch it in, and go.
So can we extrapolate how big our Visa cards would be if they didn't chop the dove down to giblet size?
Were it a film holograph, with the right equipment you could; in that laser's have a characteristic coherency, and holograms can only be produced by one of about half a dozen different types of lasers, simply because you have to manufacture the film specifically for the process, and the film only gets cheap in massive quantities. So you could peel the hologram off, or setup an optical front surface mirror arrangement allowing you to shine different types of laser light through and measure the change in coherency.
How does that help? Slicing the image down changed the coherency of the interference pattern. If you shine the original type of laser through it and measure the change in coherency with an interferometer, you can measure the amount of information lost; that corresponds exactly to the change in area
However, the dove on the VISA card is almost certainly a mechanical hologram, not a film hologram. Basically, because atoms are significantly smaller than photons, you can take a hologram and create a metal stamp based on it, and then use that stamp to create a holographic impression on a plastic or metal foil. It's a cheap way to mass produce holograms. But you lose a lot of the coherency information, which is why mass produced holograms are always virtual images (they appear to be inside of, rather than outside of, the hologram), because virtual images are more forgiving of slight errors in the process.
Yeah, but these holograms are projected from a reasonably small screen, so they probably use stereoscopic separation to produce the holographic keypad.
Actually, the patent as written doesn't work. The patent indicates they are using traditional film holography, which cannot be projected from a small screen. For a discussion about the subject, see this post.
I'm not saying this display would be extremely useable for a one-eyed person; I know from experience it would not. While I am still blessed with two working eyes, they have significantly different attributes and my vision without corrective optics is almost entirely two dimensional. All the same, I can see the image, which the original poster averred would not be the case.
Can you read the patent and figure out how the "The holographic image generator 200" works?
It's traditional film holography. Which (as I've posted elsewhere in this discussion) is woefully inadequate to meet their needs.
If you're interested in how traditional film holography works, this site is pretty decent at a guide, including a home howto.
You can't see a hologram with only one eye
Poppy cock. Of course you can see a hologram with only one eye!
You just can't see a stereoscopic vision allowing you to definitely position the object in three dimensional space. But the eyes use other cues than stereoscopic vision to determine position, cues like parallax and brightness, as well as ocular focus.
Unfortunately they didn't do their homework :(
From the website:
HoloTouch, images of keypads can be any size, entirely independent of the size of the hardware.
(emphasis my own)
From the patent:
When a hologram is illuminated by a reconstruction beam, it produces a real image (which appears to be between the plane of the hologram and the viewer) and a virtual image (which appears to be behind the plane of the hologram). [snip] Thus, it is preferred that the holographic image 207 be a real image.
Quick review of holography: an extremely high resolution film takes pictures of the interference pattern generated when a coherent light beam strikes an object.
When coherent light of a similar wavelength later shines through this film, the interference patterns cause it to be shined through in exactly the same manner as the original coherent light, up to about half the resolution of the film. Most holographic film is 3000 lines per inch, so the hologram has a "resolution" of about 1500 lines per inch.
You see an image because the light reaching your eyes through the film is exactly as it would be had the object been in front of your eyes and illuminated by the original beam.
The light reaching your eyes is coming through the film and then traveling in a straight line from the film to your eyes. You can only see such light if the holograph is directly behind it, because the path of the photons cannot change after it passes through the hologram (disregarding minor lensing effects due to the atmosphere, that is)
What does this mean? Well if the hologram appears to be one half meter in front of you and the holographic film is one meter in front of you, and the holographic image appears to be 10 cm x 10 cm, then the minimum possible size for the holographic film is 20 cm x 20 cm.
I don't call that entirely independent; as a matter of fact, it's a pretty simple relationship governed by a version of the inverse square law.
Oh, an interesting fact about it is if you take a holographic film and cut it in half, because all the information about the image is stored throughout the film, you don't have half a hologram; you have a hologram of the entire object that is half the size of the original. Pretty cool stuff actually.
Because you can't get drunk off of water, cereal, and yeast.
Well you can; you just have to wait a while before consuming...
... 7.1 channel Audio
Why?
It's a HANDHELD!!!
I can just imagine the proprietary headphone now... it's a wrap around band for your whole head, with a subwoofer on top.
Oh and the name... PSP?
I hate to say that this guy is already using the name....
Oh and in case I haven't karma whored enough?
Here's pics of the presentation
I've been wanting to have a Linux firewall that boots from CD (with no HD) for security reasons... script it to reboot every night a 3am, and you could be pretty confident in it not being cracked.
look at any of the excellent Linux Firewall on a floppy projects. Any one of them could be burned to CD.
There is nothing special for Lindows. There's maybe half a dozen steps required to modify a typical distro to boot from CD. Stuff like setting up the startup scripts not to remount the root partition as read write, setting up a ramdisk, and putting the home directories on the ramdisk.
dyne:bolic (www.dynebolic.org) is another distro that does this, and they are fully open source, if you want to check it out.
Just an aside:
The POS and MP3 systems would both be better served via PXE boot...
Since in each case you require a central server, you don't even need an optical drive in the POS/MP3 system. Make a suitably small boot iamge (10-50 MB is good) and PXE is decently fast. Plus you don't have to make coasters every time you update the OS
Good thing the mirror doesn't convert heat energy into kinetic energy, or we'd be in trouble!
The mirror converts the momentum of electromagnetic particles into it's own momentum. A Carnot style heat engine is one that derives it's energy from the movement of heat from one portion of a system to another. Steam turbines are an excellent example of a carnot heat engine. A solar sail does not work on that principle at all.
I suppose next he'll publish a paper claiming that these http://www.lonezone.com/2000/catalog/lz888.html will never work either...
Use a VIA EDEN 5000 and a 60 Watt DC-DC power supply... get that with attractive case for $160/each... or if you leave off the case, $130/each...
Make a PXE boot server and roll your own operating system and application set from the linux distros out there.
End result? You've got a stack of cheap, general purpose, configurable, completely solid state computers.
did it occur to you to try a, you know, contemporary linux distribution?
Did it occur to you to try a, you know, actual helpful response?
I mean seriously give me some pointers? WTF do you mean by contemporary? What's a contemporary linux distro? All the distros I tried were packaged after the board was released, if that's what you mean. And there's only so much fucking around that I'm willing to do before I just give up and install windows...
Keep in mind that CF cards are only rated for 100,000 writes, usually.
They are not a good replacement for a hard disk. Especially if your operating system is gonna put a pagefile on it. In one test case, the MTBF was 1 month.
That's about 3,000 page swaps a day. Not unrealistic considering that these mini computers are usually underpowered on RAM...
E-mail me for details or questions or pictures (James {dot} McCracken {at} stratapult {dot} com), but here you go:
VIA M10000 - 1GHz Nehemiah processor (good enough) - $150
512 MB RAM (hell 256 MB is probably enough) - $???80???
HDD (size and type doesn't really matter) - $70
Video- Two options here:
1. Get a PCI video card. I know it seems archaic but they still make them and these have decent enough performance compared to having to buy a shuttle.
2. Get a PCI-AGP converter (www.mini-itx.com has them) and a half-height AGP card. Again non-ideal.
Either option is gonna cost you about the same but option 2 actual gives you more options for performance than option 1. You'll get good enough performance for most non directx-9 games with this configuration. If you splurge and get the ATI A-I-W (there's a 9000 radeon PCI version for under $100), then you can turn this machine into a TiVo for free (no additional parts required) when you're not gaming.
Morex Cubid 3677 ($60) is a REAL nice looking case and you could mod a handle into it... or for the same price, get the power supply from the Morex line of cases, and get some acrylic to custom make your own case.
Totla cost: about $460. If you can scrounge some of the parts you'll get off cheaper, of course. I went without a video card, which still gives you good enough performance... 2000 3dmark2001 score... I can play AVP2 with a lot of the eye candy turned off and the res set down to 640x480... and I ended up building it all for $250.