Have you heard of Stanley Miller's classic experiment? He mixed several simple molecules (hydrogen, water, methane, and ammonia) and exposed the mixture to an electrical spark discharge. This had the result of producing complex molecules including some of the amino acids which are part of living systems. The only "agent" required was the energy supplying the electrical discharge. More details here. There are many other examples of physical and chemical systems exhibiting self-organizing behaviour, with no violations of the thermodynamic laws.
I'm not going to speculate on andover's hidden agenda, but I would like to add my vote for postgresql. I've used it at work for about 18 months (database is now ~120M in size, with >100000 rows in a few tables). One of the main reasons I chose it was its opensource license. Obviously it doesn't have the capabilities of Oracle, but it's a good product and it meets my needs.
I don't remember anything about ST:TNG being rendered on Amigas. Are you sure you're not thinking of Babylon 5 (which started on Amigas, though they later switched to x86)?
> such as the blitter chip - A much more sophisticted one than the simple rectangle movement chips in a lot of SVGA cards too. This could take 3 sources and AND and OR them together.
I once programmed my A500 to run the cellular-automata Life simulation using only the blitter. I think it used 4 memory-blocks (1 displayed, 3 for calculation) and about a dozen blitter operations per timestep. At the time, it was a lot faster than using the CPU (though this advantage went away in later models like the 4000).
In terms of booting, the Amiga is much faster than any x86 box. My A4000 gets to a graphical desktop before my x86 box finishes its BIOS startup stuff.
You cannot emulate an Amiga in real-time with a 386DX-25. I haven't tried a recent version of UAE, but the last time I did it was slower than the real thing (25MHz Amiga, 233MHz K6).
I am not disputing that for CPU-bound applications like rendering, any modern x86 box will blow a classic (680x0) Amiga out of the water. But when it comes to responsiveness of the user interface (including non-CPU-bound applications like word processors), the old Amigae still have an edge. I'm sorry, but GNU/Linux with the X Window System is a bloated hog. Less bloated than Win98, but I think it's more useful to compare yourself to the best than to the worst.
The Amiga will be dead when everything it does is done better by something else, and not before.
Some models (A1000, maybe the 3000?) had "WOM", which had to be loaded from the "Kickstart" disk when you first started your machine. These machines had a small bit of code in ROM, just enough to copy the contents of a floppy into the special area of RAM (which was then write-protected by flipping a bit in some configuration register, and which survived warm-boots). This was the 1985 equivalent to flashing the BIOS in an x86 box.
Robert Forward's "Rocheworld" dealt with this issue, though it was a one-way trip to Barnard's star rather than to Mars.
Re: Date/Time (of topic (but more interesting...))
on
Happy Pi Day!
·
· Score: 1
I also use the YYYY-MM-DD format whenever I can (though I didn't know it was an ISO standard until a few months ago). It makes programming a lot easier when you can sort dates chronologically just by comparing their numeric value. Also it turns out that you can fit a code of the format YYYYMMDDnn (where 'nn' is an arbitrary 2-digit number) into a 32-bit integer. One place this is used is in the serial number for DNS records, as in:
slashdot.org origin = slashdot.org mail addr = malda.slashdot.org serial = 2000022300
It's a pity, but I agree with you. RC5-64 is incredibly boring and pointless; it proves nothing that wasn't shown by the previous RC5-56 contest. OGR interested me because it was new and was potentially, slightly useful to some people. I only run the d.net client on one box now (plus any others I've forgotten about at work), and the rest of them are in the "APM Contest", a global network of computers attempting to conserve whatever small bits of electricity they can in their spare time. A more worthwhile cause IMHO, and it's open-source too.
what do source RPM's expand to? usually a gzipped tarball.
No, they expand to 1 or more gzipped tarballs, patches, and a "spec" file with build options for various platforms plus other descriptive information. There's a difference.
As for binary stuff, a gzipped tarball is not a package. It is a compressed archive of files, but it lacks the metadata contained in an RPM package (install scripts, dependencies, author info, installation-directory relocation, etc). Yes, it is possible to put this metadata into a tarball by adding "magic" files that don't actually get installed, but I personally feel the RPM approach is cleaner.
It's convenient. It's reasonably well designed. It's free, open, and documented (www.rpm.org). I happen to like RPM.
Memory protection was actually one of the big things the Amiga did NOT have (A1000 "write-once" kickstart memory excluded). It was quite common for a user process to step on the memory space of another process or of the operating system. Still, the Amiga did a decent job given the lack of hardware memory management on the 68000.
Booting from a ramdisk sure was a nice feature. On a 1M A500, I often copied my system floppy into a recoverable ramdisk; after all, why let all that memory go to waste?:)
If you're serious about this, I wouldn't mind joining a project to work on it. I've considered buying either a Tivo or a ReplayTV, but the main reason I haven't yet is this - control. I do not want Yet Another Sealed Box with a proprietary and cumbersome remote-control-navigated interface (such as my recently-purchased StarChoice DTH satellite dish). I do not want boxes that have to phone the central office at random hours (the StarChoice phones in to order pay-per-view movies, but is otherwise a stand-alone unit. It gets its programming information from the satellite signals). I want the ability to integrate my various units so I can (for example) program a week's worth of TV viewing from my desktop PC's web browser (or my Palm IIIx).
I see a few problems with the DIY approach:
- Noise, size, heat, and power consumption. A typical desktop PC is an offensive beast compared to an embedded-style unit. A Netwinder would be a nice base platform (it even has IR), except I don't believe its video hardware can handle full-motion video.
- Hard drive contention - you may not be able to do the simultaneous record/playback without a special hard drive (but you could do it with two drives).
- Compression ratio / compression speed. You need some serious CPU power unless your video card can do hardware compression. From their web page it looks like the ATI card doesn't ("Pentium III recommended for full resolution MPEG-2 capture"). The Matrox Marvel G400-TV has hardware MJPEG (Motion JPEG) compression, so it might be a better choice.
- Hardware drivers and chipset programming information. The situation is improving, but I don't know if there's enough available for Linux yet.
In the end, I don't know that a DIY solution would end up any cheaper than a Tivo/ReplayTV. It would certainly be larger and louder. However, it could easily be a superior product for the technically-inclined folks who want to take back control of their appliances.
Now that we're dealing with digital technology, it should be possible to create a signature of known commercials (i.e. you see a commercial, then press a button telling the unit to remember it and skip over that sequence in future). All you'd need is to store a few frames of each commercial in a database, then scan the video stream looking for matches. This could also be done in the background whenever the unit was not recording a program.
Advertisers wouldn't like this of course, so it would be hard for a manufacturer to sell a unit with this feature. However, the advertisers could still use 'Truman Show'-style product placements to replace their beloved commercials.
Ever try Corel Office for Java? It was, shall we say, less than useful. I like Java for little things, but for big apps like office suites I think you really *want* to recompile for each platform just to have a chance of using some optimization rather than writing for a lowest-common-denominator virtual machine.
Portable code is your friend - "Write once, {./configure ; make ; make install } anywhere" (tm & patent pending).
As I read this article, all I can think of is the 'B' Ark from the Hitchhiker's Guide trilogy, whereby a planet rid itself of the useless third of its society.
But I work for a company called Sonigistix, so who am I to criticize? (It's the 5th name we've had since I've worked here; I don't know what we paid for this one).
TCP/IP is going to have difficulty communicating over huge-latency links, because its error-correcting model is based on ACKs and re-transmission requests. A better model is forward error-correction such as the digital TV satellites use. Here there is no possibility for a receiver to ask for a re-transmission of a particular chunk of data, so instead they build enough redundancy and error-correcting codes into the data stream that you only need to get some percent of it to get a perfect reconstruction of the data. This would also apply (and is probably already in use somewhere) to Internet multimedia transmissions like RealAudio streams or videoconferencing, so that a small number of lost UDP packets wouldn't cause a video/audio dropout.
Click #1: "topics" on the left side of the page, http://slashdot.org/topics.shtml Clich #2: "News" option, http://slashdot.org/search.pl?topic=news Click#3: "Australian Government Cracks Down on Net Users" from Nov. 26 (Currently #8 on the menu]), http://slashdot.org/article.pl?sid=99/11/26/105123 6&mode=nested Click#4, Right-hand side of the page, "Related Links" box, you will find... CIA (http://www.cia.gov)
Berlin is trying to replace X with something more modern (using bits of OpenGL, Corba,...). I'm not involved in the project at all, but I think it's worth a look if you're interested in alternatives to X. As a former Amiga user, I certainly agree about the 'big, slow, bloated, and old' comments. However, X has its good points (e.g. network-transparency).
Actually, from what I remember, you can't say what happens. There are a family of possible solutions, but no way of telling which one of them will be observed. In other words, mechanics with time-travel is non-deterministic sort of like quantum stuff is.
The family of possible solutions involve the 'original' ball coming in, a 'second' ball coming out of the wormhole and hitting the 'original' one, the 'original' one entering the wormhole, then the 'second' one moving off. The point at which the balls collide is uncertain, as is the final angle at which the 'second' ball leaves. It's even possible to have a case where the ball may or may not even go through the wormhole.
Consider a ball heading toward a pair of wormhole mouths, exactly perpendicular to the line between them. One solution is that the ball simply passes between them and continues off into space. However, it's also possible for the ball to collide with itself in between the mouths, deflecting itself at right angles into one wormhole mouth, out the other one at an earlier time, then finally colliding with its earlier self and being deflected through another right-angle to continue on its original course.
I tried to do an ASCII-art sketch, but after seeing what 'Preview' did to it I decided to give up.:( Anyway, read the book as it does a much better job of explaining this.
Kip Thorne's "Black Holes and Time Warps", which simplifies the discussion by removing the issue of free will. He describes situations where a billiard ball is sent back in time through a wormhole, coming out on a trajectory where it deflects itself from entering the wormhole in the first place. This is a simple enough problem that it can be solved mathematically.
Robert L. Forward's "Timemaster ", which (without discussion of its literary merits) contains an interesting bit where the protagonist figures out a way to save his family from harm by filtering the information going back in time.
The key seems to be that whatever happens, must be self-consistent. So, Forward's version of your scenario might go like:
1. I know how to travel through time 2. A third person says "your friend and a car were involved in an incident at [place] and [time]" 3. I travel back to that place and time, and pull my friend back just before the car hits him, but it still runs over his shopping bag. 4. The third person reports that "your friend and a car were involved in an incident at [place] and [time]". 5. I travel back [....]
In this scenario, I do not change any known [to me] facts about the universe. However, I do ensure that any ambiguous situations are resolved in my favour.
Finally, the CANDU reactors run on natural uranium.
This is also nice because it doesn't give the country an excuse to develop uranium-enrichment plants, which makes it a bit harder to build weapons[1], and also removes a potential accident source (like the recent Japanese mishap where they poured too much uranium into a mixing tank).
Interestingly, natural nuclear reactions in the ground are not unheard of.
The Oklo mine, in West Africa. Some details are here. It's pretty neat - water-moderated, and regulated because the water would boil away when too much power was produced. Sort of like a CANDU, actually.
[1] Bombs, plus the 'depleted uranium' bullets that the US seems to like spraying all over the landscape.
Slashdot Mirror
Have you heard of Stanley Miller's classic experiment? He mixed several simple molecules (hydrogen, water, methane, and ammonia) and exposed the mixture to an electrical spark discharge. This had the result of producing complex molecules including some of the amino acids which are part of living systems. The only "agent" required was the energy supplying the electrical discharge. More details here. There are many other examples of physical and chemical systems exhibiting self-organizing behaviour, with no violations of the thermodynamic laws.
I'm not going to speculate on andover's hidden agenda, but I would like to add my vote for postgresql. I've used it at work for about 18 months (database is now ~120M in size, with >100000 rows in a few tables). One of the main reasons I chose it was its opensource license. Obviously it doesn't have the capabilities of Oracle, but it's a good product and it meets my needs.
I don't remember anything about ST:TNG being rendered on Amigas. Are you sure you're not thinking of Babylon 5 (which started on Amigas, though they later switched to x86)?
> such as the blitter chip - A much more sophisticted one than the simple rectangle movement chips in a lot of SVGA cards too. This could take 3 sources and AND and OR them together.
I once programmed my A500 to run the cellular-automata Life simulation using only the blitter. I think it used 4 memory-blocks (1 displayed, 3 for calculation) and about a dozen blitter operations per timestep. At the time, it was a lot faster than using the CPU (though this advantage went away in later models like the 4000).
In terms of booting, the Amiga is much faster than any x86 box. My A4000 gets to a graphical desktop before my x86 box finishes its BIOS startup stuff.
You cannot emulate an Amiga in real-time with a 386DX-25. I haven't tried a recent version of UAE, but the last time I did it was slower than the real thing (25MHz Amiga, 233MHz K6).
I am not disputing that for CPU-bound applications like rendering, any modern x86 box will blow a classic (680x0) Amiga out of the water. But when it comes to responsiveness of the user interface (including non-CPU-bound applications like word processors), the old Amigae still have an edge. I'm sorry, but GNU/Linux with the X Window System is a bloated hog. Less bloated than Win98, but I think it's more useful to compare yourself to the best than to the worst.
The Amiga will be dead when everything it does is done better by something else, and not before.
Some models (A1000, maybe the 3000?) had "WOM", which had to be loaded from the "Kickstart" disk when you first started your machine. These machines had a small bit of code in ROM, just enough to copy the contents of a floppy into the special area of RAM (which was then write-protected by flipping a bit in some configuration register, and which survived warm-boots). This was the 1985 equivalent to flashing the BIOS in an x86 box.
>You know, film an entirely bogus Mars Mission in some warehouse in the desert SW and show it on the nightly news and news specials as 'real'.
Like the 1978 movie Capricorn One?
Robert Forward's "Rocheworld" dealt with this issue, though it was a one-way trip to Barnard's star rather than to Mars.
I also use the YYYY-MM-DD format whenever I can (though I didn't know it was an ISO standard until a few months ago). It makes programming a lot easier when you can sort dates chronologically just by comparing their numeric value. Also it turns out that you can fit a code of the format YYYYMMDDnn (where 'nn' is an arbitrary 2-digit number) into a 32-bit integer. One place this is used is in the serial number for DNS records, as in:
slashdot.org
origin = slashdot.org
mail addr = malda.slashdot.org
serial = 2000022300
Commercial tripwire is available for WinNT, though I haven't used it. www.tripwiresecurity.com
It's a pity, but I agree with you. RC5-64 is incredibly boring and pointless; it proves nothing that wasn't shown by the previous RC5-56 contest. OGR interested me because it was new and was potentially, slightly useful to some people. I only run the d.net client on one box now (plus any others I've forgotten about at work), and the rest of them are in the "APM Contest", a global network of computers attempting to conserve whatever small bits of electricity they can in their spare time. A more worthwhile cause IMHO, and it's open-source too.
what do source RPM's expand to? usually a gzipped tarball.
No, they expand to 1 or more gzipped tarballs, patches, and a "spec" file with build options for various platforms plus other descriptive information. There's a difference.
As for binary stuff, a gzipped tarball is not a package. It is a compressed archive of files, but it lacks the metadata contained in an RPM package (install scripts, dependencies, author info, installation-directory relocation, etc). Yes, it is possible to put this metadata into a tarball by adding "magic" files that don't actually get installed, but I personally feel the RPM approach is cleaner.
It's convenient. It's reasonably well designed. It's free, open, and documented (www.rpm.org). I happen to like RPM.
Memory protection was actually one of the big things the Amiga did NOT have (A1000 "write-once" kickstart memory excluded). It was quite common for a user process to step on the memory space of another process or of the operating system. Still, the Amiga did a decent job given the lack of hardware memory management on the 68000.
:)
Booting from a ramdisk sure was a nice feature. On a 1M A500, I often copied my system floppy into a recoverable ramdisk; after all, why let all that memory go to waste?
If you're serious about this, I wouldn't mind joining a project to work on it. I've considered buying either a Tivo or a ReplayTV, but the main reason I haven't yet is this - control. I do not want Yet Another Sealed Box with a proprietary and cumbersome remote-control-navigated interface (such as my recently-purchased StarChoice DTH satellite dish). I do not want boxes that have to phone the central office at random hours (the StarChoice phones in to order pay-per-view movies, but is otherwise a stand-alone unit. It gets its programming information from the satellite signals). I want the ability to integrate my various units so I can (for example) program a week's worth of TV viewing from my desktop PC's web browser (or my Palm IIIx).
I see a few problems with the DIY approach:
- Noise, size, heat, and power consumption. A typical desktop PC is an offensive beast compared to an embedded-style unit. A Netwinder would be a nice base platform (it even has IR), except I don't believe its video hardware can handle full-motion video.
- Hard drive contention - you may not be able to do the simultaneous record/playback without a special hard drive (but you could do it with two drives).
- Compression ratio / compression speed. You need some serious CPU power unless your video card can do hardware compression. From their web page it looks like the ATI card doesn't ("Pentium III recommended for full resolution MPEG-2 capture"). The Matrox Marvel G400-TV has hardware MJPEG (Motion JPEG) compression, so it might be a better choice.
- Hardware drivers and chipset programming information. The situation is improving, but I don't know if there's enough available for Linux yet.
In the end, I don't know that a DIY solution would end up any cheaper than a Tivo/ReplayTV. It would certainly be larger and louder. However, it could easily be a superior product for the technically-inclined folks who want to take back control of their appliances.
Now that we're dealing with digital technology, it should be possible to create a signature of known commercials (i.e. you see a commercial, then press a button telling the unit to remember it and skip over that sequence in future). All you'd need is to store a few frames of each commercial in a database, then scan the video stream looking for matches. This could also be done in the background whenever the unit was not recording a program.
Advertisers wouldn't like this of course, so it would be hard for a manufacturer to sell a unit with this feature. However, the advertisers could still use 'Truman Show'-style product placements to replace their beloved commercials.
Ever try Corel Office for Java? It was, shall we say, less than useful. I like Java for little things, but for big apps like office suites I think you really *want* to recompile for each platform just to have a chance of using some optimization rather than writing for a lowest-common-denominator virtual machine.
./configure ; make ; make install } anywhere" (tm & patent pending).
Portable code is your friend - "Write once, {
But I still prefer SGI's.
The old cube one, I hope. The new logo just plain sucks.
As I read this article, all I can think of is the 'B' Ark from the Hitchhiker's Guide trilogy, whereby a planet rid itself of the useless third of its society.
But I work for a company called Sonigistix, so who am I to criticize? (It's the 5th name we've had since I've worked here; I don't know what we paid for this one).
TCP/IP is going to have difficulty communicating over huge-latency links, because its error-correcting model is based on ACKs and re-transmission requests. A better model is forward error-correction such as the digital TV satellites use. Here there is no possibility for a receiver to ask for a re-transmission of a particular chunk of data, so instead they build enough redundancy and error-correcting codes into the data stream that you only need to get some percent of it to get a perfect reconstruction of the data. This would also apply (and is probably already in use somewhere) to Internet multimedia transmissions like RealAudio streams or videoconferencing, so that a small number of lost UDP packets wouldn't cause a video/audio dropout.
Click #1: "topics" on the left side of the page, http://slashdot.org/topics.shtml3 6&mode=nested
:-)
Clich #2: "News" option, http://slashdot.org/search.pl?topic=news
Click#3: "Australian Government Cracks Down on Net Users" from Nov. 26 (Currently #8 on the menu]), http://slashdot.org/article.pl?sid=99/11/26/10512
Click#4, Right-hand side of the page, "Related Links" box, you will find... CIA (http://www.cia.gov)
What do I win? I hope it's a bowl of grits.
Berlin is trying to replace X with something more modern (using bits of OpenGL, Corba, ...). I'm not involved in the project at all, but I think it's worth a look if you're interested in alternatives to X. As a former Amiga user, I certainly agree about the 'big, slow, bloated, and old' comments. However, X has its good points (e.g. network-transparency).
Actually, from what I remember, you can't say what happens. There are a family of possible solutions, but no way of telling which one of them will be observed. In other words, mechanics with time-travel is non-deterministic sort of like quantum stuff is.
:(
The family of possible solutions involve the 'original' ball coming in, a 'second' ball coming out of the wormhole and hitting the 'original' one, the 'original' one entering the wormhole, then the 'second' one moving off. The point at which the balls collide is uncertain, as is the final angle at which the 'second' ball leaves.
It's even possible to have a case where the ball may or may not even go through the wormhole.
Consider a ball heading toward a pair of wormhole mouths, exactly perpendicular to the line between them. One solution is that the ball simply passes between them and continues off into space. However, it's also possible for the ball to collide with itself in between the mouths, deflecting itself at right angles into one wormhole mouth, out the other one at an earlier time, then finally colliding with its earlier self and being deflected through another right-angle to continue on its original course.
I tried to do an ASCII-art sketch, but after seeing what 'Preview' did to it I decided to give up.
Anyway, read the book as it does a much better job of explaining this.
A couple of books to read:
Kip Thorne's "Black Holes and Time Warps", which simplifies the discussion by removing the issue of free will. He describes situations where a billiard ball is sent back in time through a wormhole, coming out on a trajectory where it deflects itself from entering the wormhole in the first place. This is a simple enough problem that it can be solved mathematically.
Robert L. Forward's "Timemaster ", which (without discussion of its literary merits) contains an interesting bit where the protagonist figures out a way to save his family from harm by filtering the information going back in time.
The key seems to be that whatever happens, must be self-consistent. So, Forward's version of your scenario might go like:
1. I know how to travel through time
2. A third person says "your friend and a car were involved in an incident at [place] and [time]"
3. I travel back to that place and time, and pull my friend back just before the car hits him, but it still runs over his shopping bag.
4. The third person reports that "your friend and a car were involved in an incident at [place] and [time]".
5. I travel back [....]
In this scenario, I do not change any known [to me] facts about the universe. However, I do ensure that any ambiguous situations are resolved in my favour.
Does this mean that [...] I could hop into one of these reverse-time zones, hang around for a while, and come out earlier than when I went in?
:-)
Sure, provided you had a sufficient supply of carbon dioxide, urine, and feces to consume while you were there...
Red Dwarf's "Backwards" episode covers this in more detail.
Finally, the CANDU reactors run on natural uranium.
This is also nice because it doesn't give the country an excuse to develop uranium-enrichment plants, which makes it a bit harder to build weapons[1], and also removes a potential accident source (like the recent Japanese mishap where they poured too much uranium into a mixing tank).
Interestingly, natural nuclear reactions in the ground are not unheard of.
The Oklo mine, in West Africa. Some details are here. It's pretty neat - water-moderated, and regulated because the water would boil away when too much power was produced. Sort of like a CANDU, actually.
[1] Bombs, plus the 'depleted uranium' bullets that the US seems to like spraying all over the landscape.