Google'd for Robert Ursin (mentioned in the article), quickly realized they meant Rupert Ursin, went to his site and looked for the most recent article, then found the same article on Arxiv (so his site wouldn't get/.'ed). It'd be altogether easier if newspaper articles had bibliographies....
There was a text-based browser before Mosaic, written at CERN and called www. That's the earliest web browser. I even remember using on a shell account in 1992 or so, though an early version of Lynx was available as well.
In the interests of Internet history, I'd like to see www. It should be able to run fine on a Linux system, as it's a simple line-based program. However, I haven't been able to find a copy, as browsers.evolt.org doesn't go back that far. Does anyone have the source?
I'm a former student of Robert G. Valiant, whom I believe works/worked for CTA a while back. Say hi to him for me.
As other posters have said, use KDE 3. You'll need to write some scripts to set up the accounts properly, since you really can't set up multiple accounts in KDE by copying the.kde directory (lots of programs need a directory to store data in, they get it from a.kde config file, but the config file says/home/username/data rather than ~/data, so copying.kde directories leads to weird hard-to-reproduce errors).
KDE3 has a nifty kiosk mode, which I don't think anyone has mentioned. It allows you to restrict access to programs on the application menu only - people don't get a terminal, and they don't get any filesystem access through the file manager. It's great for Web browsing and e-mail, though it can lead to trouble when you want to, say, rename a file.
Use KDE, NIS, and NFS so home dirs are shared across the system, of course. That's easy to set up. Using rdist for the KDE distribution itself is a good plan too.
If you spend the time to set up Linux properly, it's a very competitive alternative to Win2K for public labs.
Instead of investing several hundred billion dollars in this project, why don't we instead pass a law that every country will stop transmitting all high-power, low-frequency (under 500MHz) radio signals for 12 hours a year? That means no non-cable TV, no radio, and no HAM, but we'd survive. Airplanes use higher-frequency transmissions to talk to air-traffic control, so they wouldn't be affected.
The ban would leave the skies clear for a relatively cheap geosynchronous sattelite to take a snapshot of the 30MHz band - and even though the protection wouldn't be perfect, it probably would be much better than what we have now. Keep in mind the radio transmission ban would only have to be enforced over half the Earth at once, and it could be at night, when hardly anyone would notice anyways. Put the ban on the countries that are asleep when they're facing the direction you want to look, and time things right so that the majority of the radio ban is over a major ocean, and there wouldn't be any real hardship.
The article mentions that you can use a mag-lev system to vastly increase the velocity of an aircraft. But on a carrier, it's also necessary to slow the aircraft down very quickly for a landing. Mag-lev is suited to this task as well - by turning the magnets "backwards," it is possible to reverse the direction of the track.
By using mag-lev for both takeoffs and landings, the Navy could presumably have takeoffs and landings on the same boat very close to each other, without the complexity of the current mechanical system. But, of course, mag-levs are useless for landings from spacee, since spacecraft usually don't have wings - and those that do can just use parachutes for losing speed.
Last time they tried this, the major roadblock was that no one could figure out how to build a server fast enough to stream multiple, unique video streams. Even assuming you're using conventional televisions and the stream size is limited to 500kB/s, you've maxed out Fibre Channel bus at 40 users under ideal conditions - and for each such group of 40 users, you need a complete copy of all the video material available, at perhaps a terabyte. There's just no way, using today's technology, to get more data on to the network - so the cable company will be stuck with tens of thousands of VoD servers, all reading information off their hard drives at the maximum rate for 24 hours a day.
I just can't see them making that kind of investment.
It sounds like Nortek developed some special software to unlock these drives. Here's how I'd do it:
Write a Linux driver that can unlock a drive given a password. This would require modifying the IDE code in the kernel. Yuck.
Crack the password. It's only 8*256 possibilities, which shouldn't take more than a few minutes since the drive is inside the machine.
Alternatively, it's possible that IBM just included a default, fallback password. I read somewhere that nearly every BIOS has a "cheat" password. Of course, motherboards aren't used to store your company's most valuable data....
At Seattle Central Community College, learning QBASIC is the first requirement for an AA degree in computer analysis. I know because a friend of mine saw me messing around with some code, and he said, "Did you know I can program too?" He then took an 800-page textbook out of his book bag and showed me what he was studying - the first introduction of control structures, somewhere around page 300.
I may have misunderstood, but I gather if he got an A in that class and the next several classes, he'd progress to Visual Basic.
Cybertown is a goal-free universe. It has quite a few nifty features, but didn't actually work all that well last time I checked (2 years ago on a PII-233:). www.cybertown.com
TAF is truly amazing. I'm surprised you volunteer their, though, since they use almost all paid staff....
Anyway, the TAF people are very very good at what they do. The students who graduate TAF programs are really smart, they not only know individual programs but quite a bit of hard CS as well. These are inner-city kids, many from very poor families....
Awhile ago, I read some stuff about a new broadband startup deciding on Grant County as a test market. There was some kind of funny technology that made doing this very cheap, and they wanted to test the technology out of the way of competition. Anyone else see it? (Was this in MIT Tech Magazine?)
It will be very interesting to see how the coming of broadband will affect one of the most conservative places outside the South. This is the palouse mentioned at the end of Cryptonomicon, where technology is still not widely understood. The biggest city in Eastern Washington still doesn't have fluoridated water!
Now Eastern Washington is competing directly with agriculture in China. China will probably win, due to the lower cost of labor, better irrigation water supply and higher government subsidies (the Chinese government funds farms at about 10% of the farm's output, while the US, a WTO member, is limited to 5%). But perhaps an influx of broadband can make Eastern Washington a viable place for tech companies to locate. One or two startups would pobably revitalize the entire region's economy.
Today I was looking through some Microsoft Knowledge Base articles about something unrelated, and I found a bunch of stuff about how Microsoft's Terminal Services accessibility API has some bugs when working with Citrix. Clearly, if it has bugs, it must exist (though the inverse may only be true with Microsoft products).
The KB article I was reading had to do with Microsoft's voice synthesizer for the deaf speaking your password out loud, even though it isn't echoed to the screen. Apparently RDP sends meta-information about the text on the screen just so accessibility apps will work but isn't smart enough to realize some information shouldn't be sent.
So it sounds like you could restrict yourself to using Microsoft's accessibility products, which are mostly developed as proofs-of-concept in research labs and aren't very polished. Or you could just give people who need accessibility apps access over SMB (hope you're not doing something fancy with DCOM or ODBC....).
By the way, this is precisely why terminal apps should send more than just a bitmap of the screen. You can't have the server preemptively deciding what data the client might possibly need and sending it off - what if the client needs to print a file locally, or save a screenscrape to disk, or change the app's resolution to fit the local screen? The ideal thin client architecture would transfer the GUI to the client at runtime, then use an RPC system to send input and events back and forth. HTML comes close and DPS comes closer, but neither are really solid. I'm contemplating doing a demonstration of these semithin clients in Python, but it's a lot of work.
Use Python and the PIDDLE graphics API. It can write very easily to PDF files, and it has nifty compression so an average page is around 1k - that makes 100-page documents reasonable. Furthermore, if you want to print to a network printer, you just change one parameter in PIDDLE and it will output PostScript instead of PDF, which you can then send to the nearest network printer. You can get the raw data using any of Python's numerous database APIs. I've done some very impressive stuff using this combination, and Python is so easy to learn it hardly feels like programming.
People are continually complaining about Open Source GUIs. Now, I hate to mention the dreaded K, but I'm running KDE2.1 on FreeBSd and it is really, really smooth. I like the design much better than Mac OS X (although it doesn't have all the nifty features), and it puts Windows to shame.
So I'd encourage you to look at modern open-source GUIs before you assume they need a lot of help in that department.
Erm, actually FreeBSD supports ATA100 fine. In fact, Debian had some problems with my weird ATA100 controller while FreeBSD works fine. I think the real problem is there is little DOCUMENTATION about drivers, so it always takes a little bit of guesswork.
The best docs available are the hardware compatibility guides included with each release's Release Notes. Check this out before posting to the mailing lists.
Have you checked out infrared lasers? I've heard of people trying to link cities with lasers so it might be able to do that distance. The market is still very new...does anyone know how much that would cost?
Write a script to convert the "graphing calculator" representation to LaTeX format. This shouldn't be too hard. Then have your script render the LaTeX source to DVI, and convert that to GIF and save in a temporary file. Then your CGI on the server side sends the name of this temporary file in HTML.
But you wanted to be able to edit your equations after they're written. So store in a database the one-line representation of the equation along with the LaTeX code. That way, when you push "Edit," you get a form with the one-line representation; when you submit the change, the LaTeX and temporary image are modified as well.
Of course, what exactly you are doing with these equations has some impact on how you implement the above....
It runs on every platform you can think of, and is accelerated in Windows and Linux. Supports 2D, 3D, console, network, physics, Windows-ish GUIs, and even Python scripting.
Also think about Allegro. It's much more low-level than CrystalSpace, but it's easier to learn (C instead of C++). It exists for Windows and X, though I haven't heard of a Mac version.
Anyone have actual experience with these two libraries? I've never used either in anything serious.
Oh yeah - check out PyGame too. Write the whole game in Python, and it will run on any platform with SDL! Samesource code for every platform. (I don't think SDL alone is what you're after, though.)
Use the same OS as your clients. In all the benchmarks I've seen, FreeBSD serving to Linux is slightly faster than Linux to FreeBSD, but FreeBSD to FreeBSD and Linux to Linux are easily twice as fast. I'd surmise that the same situation arises between Solaris and Linux.
I've heard this is caused by different tuning values in the NFS stack. I suppose you could try to change these values to match across your network, but it'd be far easier to standardize on one OS.
Why don't you use MULTIPLE Linux NFS servers? You might be able to use IDE disks that way, which would save you a bundle. Just an idea.
Does anyone know how SAN devices would work in this kind of setting? 200 Gb is just about the smallest SAN device you coulld get, but presumably it would be more optimized than a real computer. But I'm not sure.
that blocks the ads from Bess. Presuumably this would make surf-habit detection a bit more difficult (no using DoubleClick cookies). And with a few more lines it could send out spurious requests to random sites and make their logfiles useless. I did set the script up on the servers at Garfield High in Seattle, but no one uses it because it is verrryyyy slow to filter lots of HTTP traffic on our already overworked server.
Oh well, there's no such thing as a technological fix for a social problem.
In the comment I just posted I said connecting the line and Japan would be difficult. Well, it sounds like that's already being taken care of. This is very important - while not too many people want to go to Siberia these days, and western Russia is just too far away, Japan might be a great place for Americans to go for a holiday.
I still think it's unlikely such a thing would be built, especially by Russia, but with Japan involved it almost makes sense.
Slashdot Mirror
Google'd for Robert Ursin (mentioned in the article), quickly realized they meant Rupert Ursin, went to his site and looked for the most recent article, then found the same article on Arxiv (so his site wouldn't get /.'ed). It'd be altogether easier if newspaper articles had bibliographies....
Read the paper here!
There was a text-based browser before Mosaic, written at CERN and called www. That's the earliest web browser. I even remember using on a shell account in 1992 or so, though an early version of Lynx was available as well.
In the interests of Internet history, I'd like to see www. It should be able to run fine on a Linux system, as it's a simple line-based program. However, I haven't been able to find a copy, as browsers.evolt.org doesn't go back that far. Does anyone have the source?
I'm a former student of Robert G. Valiant, whom I believe works/worked for CTA a while back. Say hi to him for me.
.kde directory (lots of programs need a directory to store data in, they get it from a .kde config file, but the config file says /home/username/data rather than ~/data, so copying .kde directories leads to weird hard-to-reproduce errors).
As other posters have said, use KDE 3. You'll need to write some scripts to set up the accounts properly, since you really can't set up multiple accounts in KDE by copying the
KDE3 has a nifty kiosk mode, which I don't think anyone has mentioned. It allows you to restrict access to programs on the application menu only - people don't get a terminal, and they don't get any filesystem access through the file manager. It's great for Web browsing and e-mail, though it can lead to trouble when you want to, say, rename a file.
Use KDE, NIS, and NFS so home dirs are shared across the system, of course. That's easy to set up. Using rdist for the KDE distribution itself is a good plan too.
If you spend the time to set up Linux properly, it's a very competitive alternative to Win2K for public labs.
The ban would leave the skies clear for a relatively cheap geosynchronous sattelite to take a snapshot of the 30MHz band - and even though the protection wouldn't be perfect, it probably would be much better than what we have now. Keep in mind the radio transmission ban would only have to be enforced over half the Earth at once, and it could be at night, when hardly anyone would notice anyways. Put the ban on the countries that are asleep when they're facing the direction you want to look, and time things right so that the majority of the radio ban is over a major ocean, and there wouldn't be any real hardship.
By using mag-lev for both takeoffs and landings, the Navy could presumably have takeoffs and landings on the same boat very close to each other, without the complexity of the current mechanical system. But, of course, mag-levs are useless for landings from spacee, since spacecraft usually don't have wings - and those that do can just use parachutes for losing speed.
Last time they tried this, the major roadblock was that no one could figure out how to build a server fast enough to stream multiple, unique video streams. Even assuming you're using conventional televisions and the stream size is limited to 500kB/s, you've maxed out Fibre Channel bus at 40 users under ideal conditions - and for each such group of 40 users, you need a complete copy of all the video material available, at perhaps a terabyte. There's just no way, using today's technology, to get more data on to the network - so the cable company will be stuck with tens of thousands of VoD servers, all reading information off their hard drives at the maximum rate for 24 hours a day.
I just can't see them making that kind of investment.
Alternatively, it's possible that IBM just included a default, fallback password. I read somewhere that nearly every BIOS has a "cheat" password. Of course, motherboards aren't used to store your company's most valuable data....
I may have misunderstood, but I gather if he got an A in that class and the next several classes, he'd progress to Visual Basic.
Cybertown is a goal-free universe. It has quite a few nifty features, but didn't actually work all that well last time I checked (2 years ago on a PII-233 :). www.cybertown.com
Anyway, the TAF people are very very good at what they do. The students who graduate TAF programs are really smart, they not only know individual programs but quite a bit of hard CS as well. These are inner-city kids, many from very poor families....
It will be very interesting to see how the coming of broadband will affect one of the most conservative places outside the South. This is the palouse mentioned at the end of Cryptonomicon, where technology is still not widely understood. The biggest city in Eastern Washington still doesn't have fluoridated water!
Now Eastern Washington is competing directly with agriculture in China. China will probably win, due to the lower cost of labor, better irrigation water supply and higher government subsidies (the Chinese government funds farms at about 10% of the farm's output, while the US, a WTO member, is limited to 5%). But perhaps an influx of broadband can make Eastern Washington a viable place for tech companies to locate. One or two startups would pobably revitalize the entire region's economy.
you know what I mean...
:)
The KB article I was reading had to do with Microsoft's voice synthesizer for the deaf speaking your password out loud, even though it isn't echoed to the screen. Apparently RDP sends meta-information about the text on the screen just so accessibility apps will work but isn't smart enough to realize some information shouldn't be sent.
So it sounds like you could restrict yourself to using Microsoft's accessibility products, which are mostly developed as proofs-of-concept in research labs and aren't very polished. Or you could just give people who need accessibility apps access over SMB (hope you're not doing something fancy with DCOM or ODBC....).
By the way, this is precisely why terminal apps should send more than just a bitmap of the screen. You can't have the server preemptively deciding what data the client might possibly need and sending it off - what if the client needs to print a file locally, or save a screenscrape to disk, or change the app's resolution to fit the local screen? The ideal thin client architecture would transfer the GUI to the client at runtime, then use an RPC system to send input and events back and forth. HTML comes close and DPS comes closer, but neither are really solid. I'm contemplating doing a demonstration of these semithin clients in Python, but it's a lot of work.
Good luck!
Use Python and the PIDDLE graphics API. It can write very easily to PDF files, and it has nifty compression so an average page is around 1k - that makes 100-page documents reasonable. Furthermore, if you want to print to a network printer, you just change one parameter in PIDDLE and it will output PostScript instead of PDF, which you can then send to the nearest network printer. You can get the raw data using any of Python's numerous database APIs.
I've done some very impressive stuff using this combination, and Python is so easy to learn it hardly feels like programming.
People are continually complaining about Open Source GUIs. Now, I hate to mention the dreaded K, but I'm running KDE2.1 on FreeBSd and it is really, really smooth. I like the design much better than Mac OS X (although it doesn't have all the nifty features), and it puts Windows to shame.
So I'd encourage you to look at modern open-source GUIs before you assume they need a lot of help in that department.
The best docs available are the hardware compatibility guides included with each release's Release Notes. Check this out before posting to the mailing lists.
Have you checked out infrared lasers? I've heard of people trying to link cities with lasers so it might be able to do that distance. The market is still very new...does anyone know how much that would cost?
But you wanted to be able to edit your equations after they're written. So store in a database the one-line representation of the equation along with the LaTeX code. That way, when you push "Edit," you get a form with the one-line representation; when you submit the change, the LaTeX and temporary image are modified as well.
Of course, what exactly you are doing with these equations has some impact on how you implement the above....
crystal.linuxgames.com
It runs on every platform you can think of, and is accelerated in Windows and Linux. Supports 2D, 3D, console, network, physics, Windows-ish GUIs, and even Python scripting.
Also think about Allegro. It's much more low-level than CrystalSpace, but it's easier to learn (C instead of C++). It exists for Windows and X, though I haven't heard of a Mac version.
Anyone have actual experience with these two libraries? I've never used either in anything serious.
Oh yeah - check out PyGame too. Write the whole game in Python, and it will run on any platform with SDL! Samesource code for every platform. (I don't think SDL alone is what you're after, though.)
Use the same OS as your clients. In all the benchmarks I've seen, FreeBSD serving to Linux is slightly faster than Linux to FreeBSD, but FreeBSD to FreeBSD and Linux to Linux are easily twice as fast. I'd surmise that the same situation arises between Solaris and Linux.
I've heard this is caused by different tuning values in the NFS stack. I suppose you could try to change these values to match across your network, but it'd be far easier to standardize on one OS.
Why don't you use MULTIPLE Linux NFS servers? You might be able to use IDE disks that way, which would save you a bundle. Just an idea.
Does anyone know how SAN devices would work in this kind of setting? 200 Gb is just about the smallest SAN device you coulld get, but presumably it would be more optimized than a real computer. But I'm not sure.
Good luck!
that blocks the ads from Bess. Presuumably this would make surf-habit detection a bit more difficult (no using DoubleClick cookies). And with a few more lines it could send out spurious requests to random sites and make their logfiles useless. I did set the script up on the servers at Garfield High in Seattle, but no one uses it because it is verrryyyy slow to filter lots of HTTP traffic on our already overworked server.
Oh well, there's no such thing as a technological fix for a social problem.
Here are the results of grep "s.*s.*h$" /usr/share/dict/web2 | tr hs HS ....
abySSolitH
AddreSSograpH
arguSfiSH
aSpiSH
aSSiSH
aStoniSH
BakSHaiSH
bakSHeeSH
BaSkiSH
beaStiSH
beSlaSH
beSplaSH
blueStockingiSH
bookSelleriSH
BoSniScH
braSSiSH
briSkiSH
bruSHbuSH
buSybodyiSH
buSyiSH
cloSiSH
clotHeSbruSH
coarSiSH
contradiStinguiSH
counterdiStinguiSH
croSSfiSH
croSSHatcH
croSSpatcH
croSSpatH
daiSybuSH
damSelfiSH
diSembelliSH
diSenmeSH
diSeStabliSH
diSfleSH
diSfurniSH
diSgarniSH
diSpleniSH
diSreliSH
diSSoconcH
diStinguiSH
duSkiSH
eSSayiSH
eStabliSH
faStiSH
fleSHbruSH
foreStiSH
freSHiSH
FrieSiSH
froStfiSH
fuStianiSH
gaStroSopH
geySeriSH
gHoStfiSH
gHoStiSH
glaSSfiSH
gloSSograpH
gooSefiSH
gooSiSH
graSSHopperiSH
greaSebuSH
griSettiSH
guStoiSH
gutterSnipiSH
gypSyiSH
HadaSSaH
HandSomeiSH
HarSHiSH
HaSHiSH
HaStiSH
HorSefiSH
HorSefleSH
HouSeSmitH
HouSewifiSH
IbSeniSH
Hmm, I don't seem to have a full dict file! It doesn't matter though, I don't think the OpenSSH folks will find a name better than "Horseflesh."
Everybody always says it's my-ess-queue-ell, but I don't believe them. Try this:
:)
mysql> select soundex("mysql");
mysql> select soundex("mysequel");
You see, even the MySQL developers think it's pronounced "sequel"!
In the comment I just posted I said connecting the line and Japan would be difficult. Well, it sounds like that's already being taken care of. This is very important - while not too many people want to go to Siberia these days, and western Russia is just too far away, Japan might be a great place for Americans to go for a holiday.
I still think it's unlikely such a thing would be built, especially by Russia, but with Japan involved it almost makes sense.