College used to have a cluster of Suns, but they gradually became unmaintained and were removed after they were all hacked. As a result, we ended up with a bunch of Windows machines and no UNIX provision. What we ended up doing was designing a net-booting Linux system that required no access to the local hard drive (documentation here) and just used that until the COs finally gave up and made it official. At around the same time, people finally gave up with the university's policy regarding undergraduate access to UNIX systems (ie, the only general provision would be access to the mail server running a heavily limited shell designed for the express purpose of reading email and carrying out various mail-related tasks) and set up a university-wide service with some support from the student union. The SRCF was the result. Of course, both these could probably have gone very differently if the authorities had taken a different view of things (the SRCF was set up after consultation with the university computing service, and our Linux system happened to coincide with a time when the college COs were too busy fighting with each other to give a damn what we did), but even so if you're unhappy with the computing facilities available to you it is worth attempting to do something about it.
Progeny already have a Debian based system with a very nice hardware autodetection system. The beta versions have been able to pick up everything on the various pieces of hardware I've tested it on. They also have a very cool tool for installing multiple systems with the same setup without having to do each one by hand - install on one, set up a DHCP server or a file containing MAC addresses and networking information, create floppies for every other machine, boot them all, come back and find that they've installed everything and configured themselves in the same way as the first machine.
Pretty much. For evolution, it's usually easier to reuse something that's already there than it is to come up with a novel solution. The vast majority of proteins fall into a relatively small number of families, and the interactions between all of them are complex. Many genes can produce several different (related) products depending on how the RNA intermediate is spliced before translation occurs. The way in which a gene is regulated can depend upon the folding of the DNA surrounding it, which itself can depend on whether another gene nearby is switched on or not. Simulating this thing is going to be a nightmare, but it also looks fun. Look out for more computer simulations of protein and gene systems in the future - this is likely to be one of the next big fields.
Re:Can we establish a society to honor these peopl
on
Genetic Stone Soup
·
· Score: 1
People that give their talents for humanity instead of profit should be honored accordingly. Do the nobel prizes in sciences require this?
What, you mean like pretty much everyone engaged in academic research?:) I'm looking at about £6500 for the next 3 years while I do my PhD (assuming I do an 8 hour day 5 days a week, that works out at £3.13 an hour. That's under minimum wage, and in reality I'll be working longer hours than that), and assuming I don't run off into industry or produce work of such brilliance that someone grabs me to set up a lab somewhere I won't be looking at an even semi-decent wage until I'm 40 or so. Alternatively I could chuck it all in now, go into computing and earn substantially more now.
If I turn down the money and go the research route, do I get anything out of your society?:)
(What I'm pretty much trying to say is that very few scientists are in it for the money, because there isn't much. People know this before they start. If you want to honour people for putting humanity before profit, you're going to have a very long list of people to honour)
On another, slightly more disturbing note, I am somewhat concerned about the use of academic funding to compete with commercial enterprises. Just because RMS does it doesn't make it right.
Celera have released their sequence under a license that restricts commercial usage (something vaguely like the Sun open source license thing, whatever it's called), whereas the public effort has released their work into the public domain (pretty BSDish, really). If Celera were the only group releasing this data, academic research into the human genome would not be able to attract the same sort of investment and would proceed significantly more slowly than it otherwise would. Using academic funding in this case secures a future for academic research in a very important field.
The problem is that companies aren't just patenting genes once they've developed a specific treatment - they're finding a gene, coming up with several hundred potential uses for it, and then patenting the lot. This isn't that hard - if you know that a gene is involved in cell-division checkpointing, you can immediatly assume that it'll be involved in some tumours and so come up with a bunch of hypothetical cancer treatments based on it. You'll get your patent. Whether or not it'll stand up is probably another matter, but most companies will be unwilling to fight you on it if you're bigger than them, and if you're smaller than them they'll just buy you out so you win anyway.
Bandwidth costs. Hard drive space costs. Sysadmin time costs. Employing people to deal with spam costs. Upgrading your mail servers to deal with the increased volume of mail you receive costs. Dial-up charges in those countries which still have per-minute billing cost. Downtime caused by mail servers falling over after someone tries to relay several million messages through them costs. I've no idea whether that reaches 10 billion euros a year, but it wouldn't surprise me.
If spam were directly analagous to unsolicited snail mail, telephone solicitation, door-to-door sales and so on, I wouldn't have to pay for the spam I receive and I wouldn't get as pissed with it as I currently do. At the point where spam no longer costs anyone other than the sender money and either carries a tag that can be filtered or can be opted out of in one easy step then I'll be happy. Up until that point it's still theft of my resources in a way that none of the other examples you cited are.
SGI are likely to promote Linux on their hardware. I don't see Irix being that long for the world in the lower-end workstation market based on a couple of conversations I've had with SGI people. If they port most of the useful features to Linux they get the advantage of a kernel that people will work on for free, letting them concentrate their efforts on the high-end multi-processor machines and media systems they make more money on.
I can't speak for performance (it feels fast, but I haven't used ext2 on the same hardware), but it feels pretty nice here. Install is currently a dump/restore affair, but I was building a new machine so just installed a base Debian system on a 150MB partition, copied over an XFS kernel, partitioned and then dumped stuff into the new root.
I managed to nuke the primary superblock while attempting to install grub (I got my partition numbering confused), but thankfully there are recovery tools on SGI's website. It took about a minute to recover things from the secondary superblock. My only other problem was that 2.95 miscompiles the XFS code (you get long hangs during even moderately heavy file i/o), and I couldn't find any EGCS packages for Debian that didn't conflict with more recent ones. I ended up installing a gcc 2.97 snapshot and dropping back to 2.95 for the couple of files that gave me compiler errors. Still, it gives me a damn cool version string:)
Linux version 2.4.0-xfs (root@cavan) (gcc driver version 2.95.3 20010125 (prerelease) executing gcc version 2.97) #11 Wed Jan 31 11:41:20 GMT 2001
Having just upgraded my system with the latest Debian, the installation leaves a lot to be desired when compared with Windows
I installed Beta 2 of Progeny today. It's a Debian-based dist written by Debian developers, and it has a whizzy graphical install. I'd have no qualms about giving it to someone who's never installed Linux before.
Hotmail addresses are recycled a short time after accounts are closed. Somebody else may have had the same account name before you and been less careful. Of course, it's also possible that one of your friends accidently put it on a web page somewhere, or mentioned it in a news posting, or half a dozen other things where address scrapers can pick it up.
MKLinux supports a large number of PCI powermacs, most of which also support LinuxPPC. In addition, there are now patches to let you run the standard kernel on nubus machines - see http://nubus-pmac.sourceforge.net/.
MKLinux runs around 10-15% slower than normal Linux on the same hardware. Citing it as a counterexample to a claim that basing MacOSX on top of Mach isn't going to be a performance issue is probably not a great move.
If you have PCMCIA modules, you'll have to get the latest PCMCIA modules package and use make-kpkg to create a.deb and install that as well. It is just as easy as making a kernel.deb.
2.4 includes PCMCIA support, so you probably don't even need to do that.
This is all documented in the Changes file, but the main ones are:
GCC 2.91 or higher. 2.7.2 won't do.
modutils 2.3.21 or higher. The layout of/lib/modules/`uname -r` has changed drastically.
PPP 2.4.0 or higher if you're on a dialup connection.
If you use PCMCIA or ISDN you'll need to update the utilities for those as well. Other than that, things should pretty much work, though you won't get LFS (larger than 2GB files on 32 bit systems) unless you recompile glibc against the 2.4 kernel headers. Oh, and don't have devfs set to automatically mount unless you have devfsd installed and setup beforehand if you want to have device nodes where you expect them to be.
They didn't need to. Infocom produced a virtual machine for all their adventure games, with interpreters for different platforms. The same data file can be used on a PC, a Spectrum and an Osbourne providing that you have an interpreter that supports that version of the machine. Someone wrote an interpreter in Java, but the data file used is still the same as the one included on the original HHGTG disks.
Debian has file-rc, which converts those SysV init scripts into more BSD type ones. It also supports the standard Debian rc.d management tools, so it automatically gets updated when you install new packages.
Thats the analogy my biology teacher used I personally think that its a crappy analogy, but i cant really think of anything better.
It's also not terribly accurate. Some of the non-coding DNA is involved in controlling gene activity, but large chunks really do seem to be, well, junk. Removing small chunks of non-coding DNA is unlikely to be enough of an advantage to give you a significantly better survival chance, and removing large chunks is difficult without also removing coding genes.
What would be really cool would be to work in some industrial strength error correction code into the copying mechanism (grey codes for example).
DNA replication is already good to 1 error in 10**9 bases duplicated. Assuming I haven't screwed my numbers up completely, that's about 3 single base pair errors per complete duplication of the genome. That's pretty damn good. What's more of a problem is DNA repair, which is somewhat less efficient (not just because it makes mistakes, but also because sometimes information has been completely lost). Some sort of checksumming would be nice, but it's not something you can retrofit without changing the entire biology of the cell.
Note that this wouldn't necessarily be a good thing - reducing mutation is probably good for the survival of individuals, but not necessarily for the long-term good of the species. Mind you, by the time we've reached sufficiently advanced technology to be able to cope with a project that size we'd probably be able to "enhance" genes somewhat more accurately than random changes.
does anyone know whether representation plays into gene interpretation? Ie the physical folding of the protein interacting with how the data it holds is interpreted (visualise paper tape which holds not only printed data, but is also knotted in such a way to expose some letters and obscure others).
I assume you mean DNA rather than protein? If so, yup. In its default state, DNA is tightly wound around a protein scaffold and can't be transcribed into RNA. Whether or not the DNA is in this inaccessible form depends to some extent on the sequence of the non-coding DNA around it amongst a variety of other things.
DNA polymerase, the enzyme which replicates DNA, isn't capable of getting all the way to the end of the chromosome. As a result, when you duplicate the DNA during cell division the copy you end up with is shorter than the original. This has obvious problems if you replicate enough. The solution is to code for an enzyme called telomerase which is capable of adding extra DNA to the end of a chromosome to get it back to the original length. This DNA doesn't code for anything (in most organisms it's made up of highly repetative sequences), and is only there to make sure you don't end up with shorter chromosomes than you started with.
Interestingly, Drosophila Melanogaster doesn't do this - rather than have an enzyme that adds DNA, transposable elements (short sections of DNA that are capable of moving themselves from place to place in the genome. Very cool.) jump in from further down the chromosome and replace the removed bits directly. AFAIK, it's the only organism known to do this - even yeast has a setup similar to the one we use.
There's no terribly strong link between genome complexity and evolutionary complexity, let alone genome size and evolutionary complexity. Aribidopsis (a small flowering plant of no great significance other than somebody having sequenced its entire genome) has about 10,000 more genes than Drosophila (fruit flies) despite being somewhat more primative. One suggestion has been that different organisms have different rates of gene duplication, leading to alterations in the size of the genome over time. Other things that tend to influence genome size and complexity are occasional doubling events where an organism ends up with two copies of its genome. The resulting lack of selection pressure against one copy results in divergence. To be honest, it's just one more thing about genomes that we don't understand yet.
Re:EULA enforcability is a fiction...
on
EULA In Games
·
· Score: 1
The GPL allows you to do things with software that you wouldn't otherwise be allowed to. Presumably if it's unenforceable you just end up limited to what you're allowed to do under copyright law, which doesn't include distributing the binaries.
X is the protocol that X applications talk in order to get themselves displayed on your screen.
QT is a graphical toolkit used for producing GUIs. The UNIX version of QT talks X, allowing QT apps to display on your X server (There's also a Windows version of QT which talks to the Windows GUI rather than talking to the X server).
QT/Embedded is a version of QT optimised for embedded applications. IIRC, it's able to talk to the framebuffer device directly. This allows you to avoid having to run an X server or using all of the X libraries, reducing memory overhead considerably.
KDE is a desktop environment written using the QT libraries.
Slashdot Mirror
College used to have a cluster of Suns, but they gradually became unmaintained and were removed after they were all hacked. As a result, we ended up with a bunch of Windows machines and no UNIX provision. What we ended up doing was designing a net-booting Linux system that required no access to the local hard drive (documentation here) and just used that until the COs finally gave up and made it official. At around the same time, people finally gave up with the university's policy regarding undergraduate access to UNIX systems (ie, the only general provision would be access to the mail server running a heavily limited shell designed for the express purpose of reading email and carrying out various mail-related tasks) and set up a university-wide service with some support from the student union. The SRCF was the result. Of course, both these could probably have gone very differently if the authorities had taken a different view of things (the SRCF was set up after consultation with the university computing service, and our Linux system happened to coincide with a time when the college COs were too busy fighting with each other to give a damn what we did), but even so if you're unhappy with the computing facilities available to you it is worth attempting to do something about it.
Progeny already have a Debian based system with a very nice hardware autodetection system. The beta versions have been able to pick up everything on the various pieces of hardware I've tested it on. They also have a very cool tool for installing multiple systems with the same setup without having to do each one by hand - install on one, set up a DHCP server or a file containing MAC addresses and networking information, create floppies for every other machine, boot them all, come back and find that they've installed everything and configured themselves in the same way as the first machine.
Pretty much. For evolution, it's usually easier to reuse something that's already there than it is to come up with a novel solution. The vast majority of proteins fall into a relatively small number of families, and the interactions between all of them are complex. Many genes can produce several different (related) products depending on how the RNA intermediate is spliced before translation occurs. The way in which a gene is regulated can depend upon the folding of the DNA surrounding it, which itself can depend on whether another gene nearby is switched on or not. Simulating this thing is going to be a nightmare, but it also looks fun. Look out for more computer simulations of protein and gene systems in the future - this is likely to be one of the next big fields.
People that give their talents for humanity instead of profit should be honored accordingly. Do the nobel prizes in sciences require this?
:) I'm looking at about £6500 for the next 3 years while I do my PhD (assuming I do an 8 hour day 5 days a week, that works out at £3.13 an hour. That's under minimum wage, and in reality I'll be working longer hours than that), and assuming I don't run off into industry or produce work of such brilliance that someone grabs me to set up a lab somewhere I won't be looking at an even semi-decent wage until I'm 40 or so. Alternatively I could chuck it all in now, go into computing and earn substantially more now.
:)
What, you mean like pretty much everyone engaged in academic research?
If I turn down the money and go the research route, do I get anything out of your society?
(What I'm pretty much trying to say is that very few scientists are in it for the money, because there isn't much. People know this before they start. If you want to honour people for putting humanity before profit, you're going to have a very long list of people to honour)
On another, slightly more disturbing note, I am somewhat concerned about the use of academic funding to compete with commercial enterprises. Just because RMS does it doesn't make it right.
Celera have released their sequence under a license that restricts commercial usage (something vaguely like the Sun open source license thing, whatever it's called), whereas the public effort has released their work into the public domain (pretty BSDish, really). If Celera were the only group releasing this data, academic research into the human genome would not be able to attract the same sort of investment and would proceed significantly more slowly than it otherwise would. Using academic funding in this case secures a future for academic research in a very important field.
The problem is that companies aren't just patenting genes once they've developed a specific treatment - they're finding a gene, coming up with several hundred potential uses for it, and then patenting the lot. This isn't that hard - if you know that a gene is involved in cell-division checkpointing, you can immediatly assume that it'll be involved in some tumours and so come up with a bunch of hypothetical cancer treatments based on it. You'll get your patent. Whether or not it'll stand up is probably another matter, but most companies will be unwilling to fight you on it if you're bigger than them, and if you're smaller than them they'll just buy you out so you win anyway.
Bandwidth costs. Hard drive space costs. Sysadmin time costs. Employing people to deal with spam costs. Upgrading your mail servers to deal with the increased volume of mail you receive costs. Dial-up charges in those countries which still have per-minute billing cost. Downtime caused by mail servers falling over after someone tries to relay several million messages through them costs. I've no idea whether that reaches 10 billion euros a year, but it wouldn't surprise me.
If spam were directly analagous to unsolicited snail mail, telephone solicitation, door-to-door sales and so on, I wouldn't have to pay for the spam I receive and I wouldn't get as pissed with it as I currently do. At the point where spam no longer costs anyone other than the sender money and either carries a tag that can be filtered or can be opted out of in one easy step then I'll be happy. Up until that point it's still theft of my resources in a way that none of the other examples you cited are.
SGI are likely to promote Linux on their hardware. I don't see Irix being that long for the world in the lower-end workstation market based on a couple of conversations I've had with SGI people. If they port most of the useful features to Linux they get the advantage of a kernel that people will work on for free, letting them concentrate their efforts on the high-end multi-processor machines and media systems they make more money on.
I can't speak for performance (it feels fast, but I haven't used ext2 on the same hardware), but it feels pretty nice here. Install is currently a dump/restore affair, but I was building a new machine so just installed a base Debian system on a 150MB partition, copied over an XFS kernel, partitioned and then dumped stuff into the new root.
:)
I managed to nuke the primary superblock while attempting to install grub (I got my partition numbering confused), but thankfully there are recovery tools on SGI's website. It took about a minute to recover things from the secondary superblock. My only other problem was that 2.95 miscompiles the XFS code (you get long hangs during even moderately heavy file i/o), and I couldn't find any EGCS packages for Debian that didn't conflict with more recent ones. I ended up installing a gcc 2.97 snapshot and dropping back to 2.95 for the couple of files that gave me compiler errors. Still, it gives me a damn cool version string
Linux version 2.4.0-xfs (root@cavan) (gcc driver version 2.95.3 20010125 (prerelease) executing gcc version 2.97) #11 Wed Jan 31 11:41:20 GMT 2001
specifically the mess that is their init scripts, but then I think that SysV-style init is just beginning to annoy me in general
apt-get install file-rc
It converts your init scripts into a more BSD style while still playing nicely with the Debian package system.
Having just upgraded my system with the latest Debian, the installation leaves a lot to be desired when compared with Windows
I installed Beta 2 of Progeny today. It's a Debian-based dist written by Debian developers, and it has a whizzy graphical install. I'd have no qualms about giving it to someone who's never installed Linux before.
Hotmail addresses are recycled a short time after accounts are closed. Somebody else may have had the same account name before you and been less careful. Of course, it's also possible that one of your friends accidently put it on a web page somewhere, or mentioned it in a news posting, or half a dozen other things where address scrapers can pick it up.
MKLinux will not run on PCI powermacs.
MKLinux supports a large number of PCI powermacs, most of which also support LinuxPPC. In addition, there are now patches to let you run the standard kernel on nubus machines - see http://nubus-pmac.sourceforge.net/.
MKLinux runs around 10-15% slower than normal Linux on the same hardware. Citing it as a counterexample to a claim that basing MacOSX on top of Mach isn't going to be a performance issue is probably not a great move.
If you have PCMCIA modules, you'll have to get the latest PCMCIA modules package and use make-kpkg to create a .deb and install that as well. It is just as easy as making a kernel .deb.
2.4 includes PCMCIA support, so you probably don't even need to do that.
This is all documented in the Changes file, but the main ones are:
/lib/modules/`uname -r` has changed drastically.
GCC 2.91 or higher. 2.7.2 won't do.
modutils 2.3.21 or higher. The layout of
PPP 2.4.0 or higher if you're on a dialup connection.
If you use PCMCIA or ISDN you'll need to update the utilities for those as well. Other than that, things should pretty much work, though you won't get LFS (larger than 2GB files on 32 bit systems) unless you recompile glibc against the 2.4 kernel headers. Oh, and don't have devfs set to automatically mount unless you have devfsd installed and setup beforehand if you want to have device nodes where you expect them to be.
They didn't need to. Infocom produced a virtual machine for all their adventure games, with interpreters for different platforms. The same data file can be used on a PC, a Spectrum and an Osbourne providing that you have an interpreter that supports that version of the machine. Someone wrote an interpreter in Java, but the data file used is still the same as the one included on the original HHGTG disks.
Debian has file-rc, which converts those SysV init scripts into more BSD type ones. It also supports the standard Debian rc.d management tools, so it automatically gets updated when you install new packages.
Thats the analogy my biology teacher used I personally think that its a crappy analogy, but i cant really think of anything better.
It's also not terribly accurate. Some of the non-coding DNA is involved in controlling gene activity, but large chunks really do seem to be, well, junk. Removing small chunks of non-coding DNA is unlikely to be enough of an advantage to give you a significantly better survival chance, and removing large chunks is difficult without also removing coding genes.
What would be really cool would be to work in some industrial strength error correction code into the copying mechanism (grey codes for example).
DNA replication is already good to 1 error in 10**9 bases duplicated. Assuming I haven't screwed my numbers up completely, that's about 3 single base pair errors per complete duplication of the genome. That's pretty damn good. What's more of a problem is DNA repair, which is somewhat less efficient (not just because it makes mistakes, but also because sometimes information has been completely lost). Some sort of checksumming would be nice, but it's not something you can retrofit without changing the entire biology of the cell.
Note that this wouldn't necessarily be a good thing - reducing mutation is probably good for the survival of individuals, but not necessarily for the long-term good of the species. Mind you, by the time we've reached sufficiently advanced technology to be able to cope with a project that size we'd probably be able to "enhance" genes somewhat more accurately than random changes.
does anyone know whether representation plays into gene interpretation? Ie the physical folding of the protein interacting with how the data it holds is interpreted (visualise paper tape which holds not only printed data, but is also knotted in such a way to expose some letters and obscure others).
I assume you mean DNA rather than protein? If so, yup. In its default state, DNA is tightly wound around a protein scaffold and can't be transcribed into RNA. Whether or not the DNA is in this inaccessible form depends to some extent on the sequence of the non-coding DNA around it amongst a variety of other things.
DNA polymerase, the enzyme which replicates DNA, isn't capable of getting all the way to the end of the chromosome. As a result, when you duplicate the DNA during cell division the copy you end up with is shorter than the original. This has obvious problems if you replicate enough. The solution is to code for an enzyme called telomerase which is capable of adding extra DNA to the end of a chromosome to get it back to the original length. This DNA doesn't code for anything (in most organisms it's made up of highly repetative sequences), and is only there to make sure you don't end up with shorter chromosomes than you started with.
Interestingly, Drosophila Melanogaster doesn't do this - rather than have an enzyme that adds DNA, transposable elements (short sections of DNA that are capable of moving themselves from place to place in the genome. Very cool.) jump in from further down the chromosome and replace the removed bits directly. AFAIK, it's the only organism known to do this - even yeast has a setup similar to the one we use.
There's no terribly strong link between genome complexity and evolutionary complexity, let alone genome size and evolutionary complexity. Aribidopsis (a small flowering plant of no great significance other than somebody having sequenced its entire genome) has about 10,000 more genes than Drosophila (fruit flies) despite being somewhat more primative. One suggestion has been that different organisms have different rates of gene duplication, leading to alterations in the size of the genome over time. Other things that tend to influence genome size and complexity are occasional doubling events where an organism ends up with two copies of its genome. The resulting lack of selection pressure against one copy results in divergence. To be honest, it's just one more thing about genomes that we don't understand yet.
The GPL allows you to do things with software that you wouldn't otherwise be allowed to. Presumably if it's unenforceable you just end up limited to what you're allowed to do under copyright law, which doesn't include distributing the binaries.
X is the protocol that X applications talk in order to get themselves displayed on your screen.
QT is a graphical toolkit used for producing GUIs. The UNIX version of QT talks X, allowing QT apps to display on your X server (There's also a Windows version of QT which talks to the Windows GUI rather than talking to the X server).
QT/Embedded is a version of QT optimised for embedded applications. IIRC, it's able to talk to the framebuffer device directly. This allows you to avoid having to run an X server or using all of the X libraries, reducing memory overhead considerably.
KDE is a desktop environment written using the QT libraries.