Warning: maybe I'm overstepping my bounds by trying to speak for a large segment of the hacker community...
I think that I and many others in the hacker/OSS/Linux/whatever community are not at all pleased with the current business and legal climate that has produced Microsoft Corporation and all it's attendent shenanigans (like the Java suit and the anti-trust trial). Mind you, I don't think Microsoft itself is the source of the problem, merely a symptom of the society and business climate that produced it.
I'm tired of business-as-usual. I'm tired of bogus patents (which includes most software ones). I'm tired of lawsuits and lawyers. I definitely don't want to see the GPL tested in court.
What I want is for everyone involved to talk it over and work something out. I want everyone to get back to work (or whatever they enjoy doing). I want everyone involved with GPL'ed software to have respect for the GPL and the ideals behind it. If there is a conflict, I want people to create a solution that satisfies all parties. The spirit of hackerdom is sharing information and working/playing on cool stuff, not settling things in court.
I suppose lawsuits are necessary when things get too far out of hand and there is no other recourse. But often they are an incredible waste of time and money; not accomplishing a damn thing except to make some lawyers richer.
There were some good solutions posted above, I hope someone at Corel checks them out.
Linux kernel 2.0.32 came out around November 1997, so that comment is almost 2 years old. Things seemed to have stabilized a lot by 2.0.33 which came out in December.
As I recall, the 1.2 networking was OK. Anyone remember that far back? I know I ran 1.2.8 for quite a while, and didn't have any kernel-related problems.
Yeah, that's kinda funny. Why don't we ever see a 6502 or a Z80 that runs at 100MHz or more? I would think it'd be so much easier to design a high speed 8-bit processor than a 32 or 64 bit one.
Since we're reminiscing about old processors... I cut my teeth on assembly language programming with my trusty 6809 processor in my Tandy Color Computer. It didn't have as many features as the Commodore 64 which was twice the price, but it was still pretty good.
Programming in assembly gave me a greater appreciation for how computers work, and served me well at college. At Purdue, I took EE 362 (Microprocessors and Interfacing) which also used the 6809. I was the only guy in that semester's class who had done assembly language programming, so I aced it easily. I hadn't done interrupt-driven programming before, so I still learned a lot.
Yep, those were the good ol' days. I bought the Tech Ref manual for my CoCo. It had complete schematics, as well as a description of all major functions. Compare that to today, when I was searching around Intel's site to see how big a L1 cache the PIII has. Sheesh, found some half-useless performance figures and a lot of marketing crap, but no spec sheet. What has the world come to?
Of course, these days aren't so bad. I remember when I was struggling to get an 80-column display for my home system (finally got it with the CoCoIII but it wasn't that good). Then the struggle was to get something close to vi because I had been bitten by the Unix bug. I had started writing my own editor (not nearly as much OSS back then). Now I work and play on a pair of SGI 1600SWs in 24-bit color, with several 32-bit multi-tasking, multi-user OSs at my beck and call. With megabytes (instead of kilobytes) of RAM. I also have my choice of OO languages and compilers, instead of being faced with the decision of Basic vs. Assembly. So it's the good new days too.
Does continuous voice recognition software use some sort of neural net technique to interpret the sounds?
No. Most common speech recognition systems use something like Hidden Markov Modeling. Basically it means that some heavy mathmatical analysis is done of the speech data to extract interesting features. Then the features are matched to known patterns.
The trainning stuff for current speech recognition systems is to tune the mathmatical models of speech components to a particular user.
Speech recognition using neural networks might work well, if you can do the processing fast enough. I can't imagine any software-simulated neural net being run fast enough for real-time recognition. But a hardware solution could.
Caveat: I am not a speech expert. Ask your local speech expert for more details.
And of course, if you wanted to develop a driver with a BSD license, being able to read the source for an existing one is an enormous help.
Take me as an example, I've been working on some IR driver code for Windows NT (yes, I know, yuck). Being able to look at the IR drivers under Linux has made my job considerably easier. I certainly can't just copy the code (it's going into a propriatary product) but I'm glad for the existance of a free driver anyway, even if I can't use it directly (as I could with a BSD licensed module). I'm sure that the original writer of the driver won't mind if I learn the interface off his code, or use an algorithm or two.
I try to give back to the community too. There's a little OSS code out there that I've written, and there'll be more in the future. Any kind of true OSS is cool.
Most (all?) 3com cards come with a limited lifetime warranty. Just send the card back, and they'll send you another. You have to pay for shipping of course.
The amount of common DNA varies substantially depending on the species you are comparing.
For instance, a chimp and a human have from 95% to 99% of their DNA in common (I'm not sure of a more precise figure). Any two mammal species are supposed to share about 80%.
As you start getting away from vertebrates, the figure drops even lower. I've heard 40% between us and insects, but don't quote me on that one.
Then there's those funky bacteria that live near sulfur plumes at the bottom of the ocean. Their biological processes are so different from ours because of their environment. We probably share only a small percentage of DNA with them.
If God did create all the species of the Earth all at once (like the Creationists would have us believe) then he seems to have been in a lazy mood that week He created the Earth. Looks like He spent just barely enough time to tweak the DNA of similar designs to fit particular ecosystems, without bothering to come up with new and original designs for each ecological niche.
Flipped around this way, Creationism makes God look small, not like the all-powerful creator of the Universe. I don't think that they (the Creationists) have really thought through the implications of their ideology. It's foolish of them, but what can you expect from that bunch anyway. There are much better Christians out there.
Smart Dust in Vernor Vinge novel
on
Smart Dust
·
· Score: 1
Vernor Vinge also also has been thinking about smart dust. Check out his latest novel "A Deepness in the Sky". About midway through, he introduces 'locators', microscopic microwave-powered sensors/computer systems. There are some interesting shenanigans the main characters get into with them. The novel is also good because of an interesting alien race.
The civilizations in the novel don't develop full nanotech nor smart AI because they're stuck in the "Slow Zone" which It should be noted, that Vinge invented to keep the action at a human-comprehensible level. Mr. Vinge was one of the first SF writers to realize that with AI and IA (intelligence amplification) that our civilization will quickly develop individuals who are many orders of magnitude smarter than we are now (literally Marvin the Android's "brain the size of a planet" stuff).
This creates a problem for him as a SF writer, because a realistic, hard SF novel that takes place 100 years from now would be incomprehensible to us now, and of course impossible for him to write. So he's created a significant crutch in the "Slow Zone" to prevent people from getting too smart, and nanotech from getting too good.
An earlier novel "A Fire Upon the Deep" gives us a peek at the "Beyond" where FTL and limited nanotech is possible, and the "Transcend" where super-AIs and unlimited nanotech is possible. A very good read, and I highly recommend it to anyone interested in the future.
1) they have a random chance of a moderation succeeding, say a 1 out of 20 chance. Don't let them know if they succeed or not.
They'd be able to figure it out... just refresh the page. Particularly obsessive people would just keep trying to moderate a post until it suceeded. I suppose the software could track to see if you have tried and failed to moderate a post, and then further deny you the ability to moderate it, but that would be a lot of overhead.
Me personnally, I'd only want to see the controls when I have some use for them. Otherwise it's just screen clutter.
Moderation is a special duty (should you choose to accept it Mr. Phelps), and it should be taken more seriously than ordinary reading, as some of the other posters have indicated.
This system seems to be OK, let's just try it out for a while and then maybe in a couple weeks, have another discussion article about it.
Gahhhh! Segment registers, yuck. You'd think we'd be done with them by now, but they're creeping back.
I did a little 8088 programming back in the Eighties, and all I can say is "Never Again". I don't mind at all on a machine with a flat memory model, but I don't want to touch segments ever again if I can help it. It didn't help that those segments on the 8088 were only 64K in size.
Netscape's browser was not completely free when first released. It was free for educational/non-commercial use only. If you were a commercial user, you were allowed to download it for evaluation purposes, but you couldn't legally keep using it without paying.
And it did work for them. There were many companies, including a former employer of mine that actually bought licenses for the browser. Of course, we got a volume discount, so it was only USD $10 or so per seat.
Check your rsize and wsize options that you use on Linux to mount the NFS stuff. I've read that increasing these to 4096 can help with Solaris interoperability.
The body's immune system would reject the head, necessitating those unpleasant anti-rejection drugs. Immune system cells (like white blood cells) are produced in the bone marrow of large bones but not the skull.
Interestingly, there's been some recent work on transplanting bone marrow at the same time as an organ transplant to try to eliminate the need for those anti-rejection drugs. Results are promising, but still a ways from being completely sucessful.
The point I was trying to make is that SCO had domminance in the low-end Unix business for point-of-sale and other business applications. Most SCO machines run a small database, some custom apps, supporting maybe a dozen users.
64 processor shared memory machines in the Intel architechture don't exist, AFAIK. But they certainly do from Sun, SGI and IBM (maybe HP too). When I think "high-end", SCO defintely doesn't come to mind. SCO never had that market for Unix, and they never will. Now they are losing the low-end, so in my opinion they are screwed.
Most of the existing apps out there were written for Openserver. I'm sure the transition to Unixware will require non-trivial porting effort.
Now is the time for them to dump Openserver, dump Unixware, and help their ISVs to port their applications to Linux (instead of Windows NT).
SCO may not have that much money in the future either. Their business model assumes that they can sell OSs for $1000+, compilers for extra, and support for way extra. The new business model (like with RedHat) is that the OS is free, and support is reasonably priced. If you read the Innovator's Dilemma, you see why companies can have a hard time switching business models.
Re:SCO Tech support? No thanks!
on
SCO does Linux
·
· Score: 1
SCO is actually on much firmer ground than Redhat. They've been in existance for many many years, and have carved out a niche for themselves. Hell, they even generated a profit last quarter!
Financially, yes, SCO is solid as a rock. But good finances alone do not make a good company (unless you are a bank).
SCO got to their position by being just about the only company that sold (relatively) low-cost Unix. But now, their prices are way out of line with what's freely available. And now that database apps like Oracle are coming to Linux, there's not much reason to be sticking with SCO.
I never used Unixware, and some people say it's pretty good. But no one's writing applications for it anymore. Heck, when's the last time you saw it as a target in a Makefile?
However, I have used Openserver 5, and I thought it stunk. I'm glad to be done with it.
So I'd be worried if I was SCO. They're like that guy falling out of a 100 story building: you ask him how he's doing as he passes the 50th floor and he says "So far, everything's OK."
I think the article writer meant that SCO is new to the Linux services market, not the services market in general.
At any rate, the article is good for a laugh or two. SCO will still be recommending it's Unix for "high-end" applications. I'm sorry, but if I'm worried about scalibility, I'll run Solaris AIX, HP-UX or Irix, definitely not SCO Unix. The Intel hardware out there that SCO can run on just can't scale like Solaris on a 64 processor Starfire. No one makes those kind of boxes yet, but it looks like SGI is going to.:-)
I've had some ideas banging around in my head (no snide jokes about the sound) about a completely object-based OS. It would in the style of the Smalltalk/Java virtual machine, but with multi-user access and completely network transparent.
It would be a very cool environment to program in, making it easy to extend to other uses. I'm also thinking about ways to automatically distribute and replicate objects across the network, so you don't have to worry about reliability. I'd also include Eiffel's Design by Contract (if not Eiffel itself) to enhance correctness.
The point of it all would be to make the system do all the grunge work, allowing the developer to think about system design.
Too bad there's no running code, or even a written specification. I'd need at least a year of free time to work some stuff out.
I don't think RSA ever contributed to the development of any version of PGP, but they did grudingly allow the use of one of their libraries in it.
I have yet to evaluate GNUPG vs. PGP, but if the GNU version is even half as good, I'll probably use it. I don't like dealing with sometimes free software like PGP. I don't want to have to think about if it's for corporate vs. private use.
Yeah, when it doesn't hang your system. Mandrake is a good example of this.
Mandrake 6.0 has a hdparm line in/etc/rc.d/rc.sysinit, which optimizes HD performance on most systems. However, on some systems, it causes the HD to hang.
By booting without starting up init (linux -b at the LILO prompt), I was able to find and comment out the hdparm line. However, since the system hung during the first boot after install, the RPM database got trashed. I ended up having to completely re-install it again. That was a pain.
Mandrake should not have put in an unnecessary optimization in as a default. Looking at the discussion about this on their mailing list, they didn't seem to care, though.
If that's all there was it wouldn't be very interesting.
However, some research summaries I've read (sorry that I can't give references) imply that there is a way to see (measure) a photon in the polarization of your choice.
To extend your analogy, this would be where person 1 chooses to see his marble as white, which would then force person 2's marble to be black. So person 1 has just transmitted a bit of information to person 2 instantaniously.
AFAIK, no one's proven that this can actually happen, so it may be in the same category as travelling back in time. It's just some interesting speculation.
I have some experience with high end NetApps, and I think you should give them serious consideration for this kind of application.
My recommendation is to go with a cluster / failover pair of F760s, which can have 1TB+ of storage between the two of them.
In normal operation, each filer of the pair serves 1/2 the files. If one dies, the other does file serving for both by taking possesion of the other's disks. This pretty much happens seamlessly.
As a test, we started copying a large directory tree to one filer, and then turned it off mid-way through. After about a minute, the other filer of the pair assumed control, and the copy continued without interruption. After the copy finished, I did a file-by-file comparison of the original and the copy (GNU diff rocks) and they were exacly the same. Very cool.
In addition to cluster/failover, the NetApps have quite a number of other high-availibility features: multiple fans, good RAID subsystem, snapshots and checkpoints, journaling filesystem, etc.
I love Linux (set up the 4th box for my home this weekend), but it's not ready for this kind of task. It's much better to go with an established vendor, instead of spending a lot of time trying to build it yourself.
Please feel free to contact me if someone has more questions about Netapps. There are also a lot of good technology papers on their web site http://www.netapp.com.
Slashdot Mirror
Warning: maybe I'm overstepping my bounds by trying to speak for a large segment of the hacker community...
I think that I and many others in the hacker/OSS/Linux/whatever community are not at all pleased with the current business and legal climate that has produced Microsoft Corporation and all it's attendent shenanigans (like the Java suit and the anti-trust trial). Mind you, I don't think Microsoft itself is the source of the problem, merely a symptom of the society and business climate that produced it.
I'm tired of business-as-usual. I'm tired of bogus patents (which includes most software ones). I'm tired of lawsuits and lawyers. I definitely don't want to see the GPL tested in court.
What I want is for everyone involved to talk it over and work something out. I want everyone to get back to work (or whatever they enjoy doing). I want everyone involved with GPL'ed software to have respect for the GPL and the ideals behind it. If there is a conflict, I want people to create a solution that satisfies all parties. The spirit of hackerdom is sharing information and working/playing on cool stuff, not settling things in court.
I suppose lawsuits are necessary when things get too far out of hand and there is no other recourse. But often they are an incredible waste of time and money; not accomplishing a damn thing except to make some lawyers richer.
There were some good solutions posted above, I hope someone at Corel checks them out.
Linux kernel 2.0.32 came out around November 1997, so that comment is almost 2 years old. Things seemed to have stabilized a lot by 2.0.33 which came out in December.
Check out Linux Kernel 2.0 patches page at LinuxHQ for more details.
As I recall, the 1.2 networking was OK. Anyone remember that far back? I know I ran 1.2.8 for quite a while, and didn't have any kernel-related problems.
Yeah, that's kinda funny. Why don't we ever see a 6502 or a Z80 that runs at 100MHz or more? I would think it'd be so much easier to design a high speed 8-bit processor than a 32 or 64 bit one.
Since we're reminiscing about old processors... I cut my teeth on assembly language programming with my trusty 6809 processor in my Tandy Color Computer. It didn't have as many features as the Commodore 64 which was twice the price, but it was still pretty good.
Programming in assembly gave me a greater appreciation for how computers work, and served me well at college. At Purdue, I took EE 362 (Microprocessors and Interfacing) which also used the 6809. I was the only guy in that semester's class who had done assembly language programming, so I aced it easily. I hadn't done interrupt-driven programming before, so I still learned a lot.
Yep, those were the good ol' days. I bought the Tech Ref manual for my CoCo. It had complete schematics, as well as a description of all major functions. Compare that to today, when I was searching around Intel's site to see how big a L1 cache the PIII has. Sheesh, found some half-useless performance figures and a lot of marketing crap, but no spec sheet. What has the world come to?
Of course, these days aren't so bad. I remember when I was struggling to get an 80-column display for my home system (finally got it with the CoCoIII but it wasn't that good). Then the struggle was to get something close to vi because I had been bitten by the Unix bug. I had started writing my own editor (not nearly as much OSS back then). Now I work and play on a pair of SGI 1600SWs in 24-bit color, with several 32-bit multi-tasking, multi-user OSs at my beck and call. With megabytes (instead of kilobytes) of RAM. I also have my choice of OO languages and compilers, instead of being faced with the decision of Basic vs. Assembly. So it's the good new days too.
Does continuous voice recognition software use some sort of neural net technique to interpret the sounds?
No. Most common speech recognition systems use something like Hidden Markov Modeling. Basically it means that some heavy mathmatical analysis is done of the speech data to extract interesting features. Then the features are matched to known patterns.
The trainning stuff for current speech recognition systems is to tune the mathmatical models of speech components to a particular user.
Speech recognition using neural networks might work well, if you can do the processing fast enough. I can't imagine any software-simulated neural net being run fast enough for real-time recognition. But a hardware solution could.
Caveat: I am not a speech expert. Ask your local speech expert for more details.
And of course, if you wanted to develop a driver with a BSD license, being able to read the source for an existing one is an enormous help.
Take me as an example, I've been working on some IR driver code for Windows NT (yes, I know, yuck). Being able to look at the IR drivers under Linux has made my job considerably easier. I certainly can't just copy the code (it's going into a propriatary product) but I'm glad for the existance of a free driver anyway, even if I can't use it directly (as I could with a BSD licensed module). I'm sure that the original writer of the driver won't mind if I learn the interface off his code, or use an algorithm or two.
I try to give back to the community too. There's a little OSS code out there that I've written, and there'll be more in the future. Any kind of true OSS is cool.
Most (all?) 3com cards come with a limited lifetime warranty. Just send the card back, and they'll send you another. You have to pay for shipping of course.
The amount of common DNA varies substantially depending on the species you are comparing.
For instance, a chimp and a human have from 95% to 99% of their DNA in common (I'm not sure of a more precise figure). Any two mammal species are supposed to share about 80%.
As you start getting away from vertebrates, the figure drops even lower. I've heard 40% between us and insects, but don't quote me on that one.
Then there's those funky bacteria that live near sulfur plumes at the bottom of the ocean. Their biological processes are so different from ours because of their environment. We probably share only a small percentage of DNA with them.
If God did create all the species of the Earth all at once (like the Creationists would have us believe) then he seems to have been in a lazy mood that week He created the Earth. Looks like He spent just barely enough time to tweak the DNA of similar designs to fit particular ecosystems, without bothering to come up with new and original designs for each ecological niche.
Flipped around this way, Creationism makes God look small, not like the all-powerful creator of the Universe. I don't think that they (the Creationists) have really thought through the implications of their ideology. It's foolish of them, but what can you expect from that bunch anyway. There are much better Christians out there.
Vernor Vinge also also has been thinking about smart dust. Check out his latest novel "A Deepness in the Sky". About midway through, he introduces 'locators', microscopic microwave-powered sensors/computer systems. There are some interesting shenanigans the main characters get into with them. The novel is also good because of an interesting alien race.
The civilizations in the novel don't develop full nanotech nor smart AI because they're stuck in the "Slow Zone" which It should be noted, that Vinge invented to keep the action at a human-comprehensible level. Mr. Vinge was one of the first SF writers to realize that with AI and IA (intelligence amplification) that our civilization will quickly develop individuals who are many orders of magnitude smarter than we are now (literally Marvin the Android's "brain the size of a planet" stuff).
This creates a problem for him as a SF writer, because a realistic, hard SF novel that takes place 100 years from now would be incomprehensible to us now, and of course impossible for him to write. So he's created a significant crutch in the "Slow Zone" to prevent people from getting too smart, and nanotech from getting too good.
An earlier novel "A Fire Upon the Deep" gives us a peek at the "Beyond" where FTL and limited nanotech is possible, and the "Transcend" where super-AIs and unlimited nanotech is possible. A very good read, and I highly recommend it to anyone interested in the future.
1) they have a random chance of a moderation succeeding, say a 1 out of 20 chance. Don't let them know if they succeed or not.
They'd be able to figure it out... just refresh the page. Particularly obsessive people would just keep trying to moderate a post until it suceeded. I suppose the software could track to see if you have tried and failed to moderate a post, and then further deny you the ability to moderate it, but that would be a lot of overhead.
Me personnally, I'd only want to see the controls when I have some use for them. Otherwise it's just screen clutter.
Moderation is a special duty (should you choose to accept it Mr. Phelps), and it should be taken more seriously than ordinary reading, as some of the other posters have indicated.
This system seems to be OK, let's just try it out for a while and then maybe in a couple weeks, have another discussion article about it.
Gahhhh! Segment registers, yuck. You'd think we'd be done with them by now, but they're creeping back.
I did a little 8088 programming back in the Eighties, and all I can say is "Never Again". I don't mind at all on a machine with a flat memory model, but I don't want to touch segments ever again if I can help it. It didn't help that those segments on the 8088 were only 64K in size.
Let's just move to 64-bit processors, shall we?
Netscape's browser was not completely free when first released. It was free for educational/non-commercial use only. If you were a commercial user, you were allowed to download it for evaluation purposes, but you couldn't legally keep using it without paying.
And it did work for them. There were many companies, including a former employer of mine that actually bought licenses for the browser. Of course, we got a volume discount, so it was only USD $10 or so per seat.
Check your rsize and wsize options that you use on Linux to mount the NFS stuff. I've read that increasing these to 4096 can help with Solaris interoperability.
Check the NFS-HOWTO for Linux for details.
James
The body's immune system would reject the head, necessitating those unpleasant anti-rejection drugs. Immune system cells (like white blood cells) are produced in the bone marrow of large bones but not the skull.
Interestingly, there's been some recent work on transplanting bone marrow at the same time as an organ transplant to try to eliminate the need for those anti-rejection drugs. Results are promising, but still a ways from being completely sucessful.
Hypwerwave also had some other interesting features such as 3-D navigation of the information space.
The versions I used back in 1997 were kind of buggy, hopefully this has improved since then.
Also back in '97, it was quite expensive. No mention of prices on their current web site.
The point I was trying to make is that SCO had domminance in the low-end Unix business for point-of-sale and other business applications. Most SCO machines run a small database, some custom apps, supporting maybe a dozen users.
64 processor shared memory machines in the Intel architechture don't exist, AFAIK. But they certainly do from Sun, SGI and IBM (maybe HP too). When I think "high-end", SCO defintely doesn't come to mind. SCO never had that market for Unix, and they never will. Now they are losing the low-end, so in my opinion they are screwed.
Most of the existing apps out there were written for Openserver. I'm sure the transition to Unixware will require non-trivial porting effort.
Now is the time for them to dump Openserver, dump Unixware, and help their ISVs to port their applications to Linux (instead of Windows NT).
SCO may not have that much money in the future either. Their business model assumes that they can sell OSs for $1000+, compilers for extra, and support for way extra. The new business model (like with RedHat) is that the OS is free, and support is reasonably priced. If you read the Innovator's Dilemma, you see why companies can have a hard time switching business models.
SCO is actually on much firmer ground than Redhat. They've been in existance for many many years, and have carved out a niche for themselves. Hell, they even generated a profit last quarter!
Financially, yes, SCO is solid as a rock. But good finances alone do not make a good company (unless you are a bank).
SCO got to their position by being just about the only company that sold (relatively) low-cost Unix. But now, their prices are way out of line with what's freely available. And now that database apps like Oracle are coming to Linux, there's not much reason to be sticking with SCO.
I never used Unixware, and some people say it's pretty good. But no one's writing applications for it anymore. Heck, when's the last time you saw it as a target in a Makefile?
However, I have used Openserver 5, and I thought it stunk. I'm glad to be done with it.
So I'd be worried if I was SCO. They're like that guy falling out of a 100 story building: you ask him how he's doing as he passes the 50th floor and he says "So far, everything's OK."
I think the article writer meant that SCO is new to the Linux services market, not the services market in general.
At any rate, the article is good for a laugh or two. SCO will still be recommending it's Unix for "high-end" applications. I'm sorry, but if I'm worried about scalibility, I'll run Solaris AIX, HP-UX or Irix, definitely not SCO Unix. The Intel hardware out there that SCO can run on just can't scale like Solaris on a 64 processor Starfire. No one makes those kind of boxes yet, but it looks like SGI is going to. :-)
James
I've had some ideas banging around in my head (no snide jokes about the sound) about a completely object-based OS. It would in the style of the Smalltalk/Java virtual machine, but with multi-user access and completely network transparent.
It would be a very cool environment to program in, making it easy to extend to other uses. I'm also thinking about ways to automatically distribute and replicate objects across the network, so you don't have to worry about reliability. I'd also include Eiffel's Design by Contract (if not Eiffel itself) to enhance correctness.
The point of it all would be to make the system do all the grunge work, allowing the developer to think about system design.
Too bad there's no running code, or even a written specification. I'd need at least a year of free time to work some stuff out.
One of these days...
EarthLink uses Unix systems and NetApp filers for mail service. They have a lot of customers.
They've written up a a paper on their mail system architechture: http://www.earthlink.com/about/ papers/mailarch.html. It's definitely worth a read.
Advertisments for advertisments. (Groan) What is the world coming to?
James
I don't think RSA ever contributed to the development of any version of PGP, but they did grudingly allow the use of one of their libraries in it.
I have yet to evaluate GNUPG vs. PGP, but if the GNU version is even half as good, I'll probably use it. I don't like dealing with sometimes free software like PGP. I don't want to have to think about if it's for corporate vs. private use.
James
They probably read the novel and then designed that animated background to induce Snowcrash on their web site visitors!
Yeah, when it doesn't hang your system. Mandrake is a good example of this.
Mandrake 6.0 has a hdparm line in /etc/rc.d/rc.sysinit, which optimizes HD performance on most systems. However, on some systems, it causes the HD to hang.
By booting without starting up init (linux -b at the LILO prompt), I was able to find and comment out the hdparm line. However, since the system hung during the first boot after install, the RPM database got trashed. I ended up having to completely re-install it again. That was a pain.
Mandrake should not have put in an unnecessary optimization in as a default. Looking at the discussion about this on their mailing list, they didn't seem to care, though.
If that's all there was it wouldn't be very interesting.
However, some research summaries I've read (sorry that I can't give references) imply that there is a way to see (measure) a photon in the polarization of your choice.
To extend your analogy, this would be where person 1 chooses to see his marble as white, which would then force person 2's marble to be black. So person 1 has just transmitted a bit of information to person 2 instantaniously.
AFAIK, no one's proven that this can actually happen, so it may be in the same category as travelling back in time. It's just some interesting speculation.
I have some experience with high end NetApps, and I think you should give them serious consideration for this kind of application.
My recommendation is to go with a cluster / failover pair of F760s, which can have 1TB+ of storage between the two of them.
In normal operation, each filer of the pair serves 1/2 the files. If one dies, the other does file serving for both by taking possesion of the other's disks. This pretty much happens seamlessly.
As a test, we started copying a large directory tree to one filer, and then turned it off mid-way through. After about a minute, the other filer of the pair assumed control, and the copy continued without interruption. After the copy finished, I did a file-by-file comparison of the original and the copy (GNU diff rocks) and they were exacly the same. Very cool.
In addition to cluster/failover, the NetApps have quite a number of other high-availibility features: multiple fans, good RAID subsystem, snapshots and checkpoints, journaling filesystem, etc.
I love Linux (set up the 4th box for my home this weekend), but it's not ready for this kind of task. It's much better to go with an established vendor, instead of spending a lot of time trying to build it yourself.
Please feel free to contact me if someone has more questions about Netapps. There are also a lot of good technology papers on their web site http://www.netapp.com.
James Graves