Why use IT staff? Schools are full of eager nerds who would love to hold an install fest...
There are many schools in Finland for example, which have student-maintained computer classrooms, school servers, web pages, etc. However, the higher level education a school offers, the less likely it is to find students maintaining the IT infrastructure...
This is indeed good news, as free software always is. But eRServer can only operate in single-master mode, which makes it unsuitable for high-availability kind of work. Single-master systems are good for load-balancing on installations where most of the queries to the DB are SELECTs.
eRServer comes a bit late. We already have PostgreSQL Replicator, which is multi-master. Unfortunately PG Replicator is not supported anymore. The latest version it can work with is 7.1, and the project's latest news are timestamped nearly two years ago.
Your mathematical utilities would be more useful if you had programmed them in C. Your choice of language will limit their adoption. Basically because using Perl scripts is not as fast as calling compiled C programs. This fact alone will make people reductant of using your utils in their code.
Because FreeBSD doesn't ship Perl as standard part of their distribution anymore, it'll be likely that your utils will not get included in any BSD software because it would pull in Perl. It may be a reason for Linux distributions too for not using your num-utils. Debian may be the only distribution which relies on Perl.
windowsupdate.microsoft.com has links for browsing the updates if your browser doesn't support activex. The activex-less pages have updates for 2000/NT and 98, but not XP, so you're left with the service packs.
But if you're not using IE or OE, you're probably safe from 99% of all the exploits Windows Update would patch. With a firewall and an anti-virus software you're safe from other kinds of exploits, or most of them at least.
BTW: Why did you remove IE? You could just make Mozilla your default browser and remove shortcuts for IE. Then there'd be no problems with Windows Update.
When the update contract expires, you'll have to get a newer version or switch to some other distribution. Redhat is on only offering the much-talked-about one-year upgrade period.
It's long time since RH5 and Debian 2.1. I suggest you take a look at Debian 3.0 (woody); It has improved much since 1998.
If you don't want to leave RPM-based systems, then you'd find SuSE an easy choice. I've used their distribution before I switched to Debian, and I don't have anything against them (save for RPMs). The YAST2 setup tool is really helpful, and you'd benefit from their participation in UnitedLinux and commitment to LSB and FHS.
If your RedHat installation doesn't give you what you need from a server operating system, it's a good time to think about switching. I would not rush it, since there is always the chance something is screwed and restoring an entire system back is not easily done even if you keep full backups. You should at least familiarize yourself with Debian before you start running servers with it.
The Debian install is not much different from the rest of the distros. If you know how Linux works, the text-based installation progress is quite simple. You can set it up to get packages from the net, so there's no need to burn all the CD's. The Debian web site has links to netinstall ISO images which are only 10-30 MB.
For me Debian install takes about half an hour. You might need considerably more on your first run. But heed this warning: Don't run task-sel nor dselect to pick the apps. I've yet to hear of a successful use of those two utilities. Especially for a server environment, you probably don't want to have all kinds of software lying around (both task-sel and dselect install tons of software you didn't want).
After you have your Debian system configured, it's a simple matter of apt-cache search'ing and apt-get install'ing the software you need. The dependency system will take care of the libraries and softwares to which the software you want depends on. After two hours you should be set up (depending on your network connection speed) with the software you need.
To keep up with security, choose only stable Debian packages. Then it's a simple matter of scripting to set up a cron job to do atp-get update and apt-get dist-upgrade periodically.
You might also want to take a look at Gentoo Linux, which offers similar packaging system to Debian. Gentoo philosophy is that you're provided with the package information and source codes which are compiled on your system for optimal setup; there's no binary distribution. I don't know of their security update model.
And if you're open-minded, there are always the BSDs.
I think signing with PGP/GPG is enough proof of autheticy. If the public key has been available on a public keyserver prior to the writing of the original story.
Besides, you can phone the company/person to verify that the response is authentic. If they withold their contact information, you cannot verify them, and thus cannot be made to publish their response on grounds it's origin unconfirmable. The key issue here is wether the writer of the original story is responsible for the costs of authenticating the response.
No central server, just a chain of peers, like peer2peer2peer2peer. Each peer would act as a server, serving connections and forwarding them, if needed.
Peer chains are easy to create. At first, you connect to a know peer from which you ask a list of peers known to it (and then you could poll other peers for the peers they know of). Then you'll have a pool of candidate peers through which you can connect to any peer on the network.
It's not "direct connect" when there are intermediate peers (like those ultrapeers in gnutella). However, the intermediates would be random and they would not know of each other.
Not quite so. Using a scheme akin of anonymous remailer that problem could be avoided altogether:
Clients establish encryped connections with masters. The connections are either direct (client to master) or forwarded (client to master to another master). For forwarded connections, the client does not tell from which machine the connection came from (it could be the machine or some of it's clients). So the server at the end of the request chain sees only the server immediately before it. That server sees only one before it and so on. None but the originating client knows which machine is the client.
Of course that would be slow, and quite un-interactive, but it would work well enough if each master would randomize the path it tries to follow to the final target. With a TTL-like requests, the number of intermediate servers could be controlled by the client. And the masters would never make up the route to the client, but use the route the client connected with them.
Not in small scale like single marriage. But psychohistory could be used to predict behaviour in nationwide or global scale.
Let's say: A nation attacks another nation, and leaves some valuable historical relics unguarded. What does psychohistory predict will happen? A number of people (note: not all) will loot the unguarded relics with the intend to get rich.
All you need for psychohistory to work, is to model historical events, and the variables that most affected them.
Can you point me to some place where tuning them is discussed. I don't like to hack them myself. I've installed the FireGL drivers from ATI. They were the only drivers that I found from their site. Work fine with X, only OpenGL applications are slow, like 140 FPS with Radeon versus 250 with TNT2. My guess is OpenGL acceleration is not working, or something. But I don't know how to turn it on. nVidia's drivers were accompanied with nVidia's OpenGL implementation, but I didn't see anything like it with the FireGL drivers.
Genetic engineers have yet a hard task ahead
on
The Rights of GM Humans
·
· Score: 2, Interesting
Genetic manipulation from a capitalistic world must support other capitalistic ventures. That's a sad fact we sometimes forget when we're thinking about advancements in science. If we lived in socialist world, then there would be a fair chance we'll get physically and mentally superior GM'd people.
But the first task of GM is not to create more intelligent people, no. The first task they'll have to undertake is to enable people to eat more unhealthy food without getting fat. Yes, beauty comes before brains in alphabet.
Besides, altering intelligence might not be so good an idea: it would generate more resistance to stupid laws, stupid politics, and stupid corporations. The people in charge of GM-corporations do not generally fall into the category of free-thinking liberalists. They're after money, not freedom.
I want speed in addition to quality. nVisia's inexpensive gForce4-based cards do not offer the same 2D image quality ATI does. It's just that I need to boot to Windows if I want to play games...
How do you know they're bogus? I know some users fake their User-Agent stings to match Win/MSIE to fool witty JavaScript writers. But I doubt Google would lie on purpose. Do you have any links that might shed some light to it? Like statistics which show there are more Linux users...
Googles Zeitgeist tells me 4% of people are using Macs whereas only 1% are using Linux. It would be more profitable to make a Mac game than a Linux game.
Besides, I don't think Linux's graphics acceleration support is so mature a casual Linux-user would get the best performance from their machine. At least my Radeon 8500 gives me less FPS than TNT2 I previously had.
And that goes for OO as well. Not every database (or a collection of data) needs to be accessed in Object-Oriented way. Most (or should I say all) data I store to small tables would not benefit from being objects.
And how does this differ from storing non-object-oriented data structures in RAM? You'd still need to implement searches, and how do you search an collection of objects without placing them on the relational line.
Re:I've found Gnome 1.4 to be the best of Gnome ye
on
Has GNOME Become LAME?
·
· Score: 1
I'd like it if programs were desktop environment agnostics, or generally would not care of your library infrastructure.
It would be nice to have usefull applications which you could use with any widget library, and the application would conform to the environment; use Gnome if it's available and running, use KDE if it's the one the user likes, or use anything that's there.
Now that we have choice of desktop, the OSS developers need to do the same applications for multiple widget libraries. All the basic applications exist for all the libraries. Developers' time is wasted to reinvent the application using some different set of libraries. But it doesn't need to be that way.
Don't take me wrong: I'm not asking anyone to invent a library which would guess which other library to use. This needs to go deeper than that if it's to succeed.
Since most spammers use open email relays to send spam, why not attack the open relays? There's no reason I can think of for leaving a relay open, so they must be open because of either lack of knowledge or because the admin wants people to be able to use it without authentication.
So why not create bots that scan for open relays and start DDOSing them once found. I doubt open relays have enough hardware to avert angry spam-recipients counter-attack.
And if the relays are open because of lack of knowledge, the DDOS attack could send some message like "Your relay is open and I like spam THIS much."
That's because most spam messages have faked envelope senders. The bounced messages never reach the spammers. And if they did, you'd have that much easier to spam the spammers. They don't want that, and they can avoid it.
This idea could be turned into a distributed project, much like distributed.net. The idea would be to create a new mail filter, or filtering plugin which would pass all URLs found in any mail message stored to spam folder to the distributed network. Each participant of the network would automatically start crawling the URLs submitted in an easy pace so as not to have an effect for the user. But the combined effect would be dramatic for the targetted websites. Anyone stupid enough to go seeking for their stuff would time out due to server load.
And of course, the User-Agent would be faked to be MSIE.
Instead of returning NOT_FOUND, timeout the request. But you propably can't do that with Apache without damaging your own capability to serve pages. But if you can, you could also send OK response after a long delay. But not deliver any mail, of course. With OK, the spammer would presume your formmail.cgi honeypot relay was functional and continue to use it. However, due to the "lag" they can not "send" mail in great speed, damaging their capability to spam. And if you want to be really evil, the script could try to figure if the spammer tries to send mail to himself, and deliver such messages to keep the spammer fooled.
And once you spot a spammer IP, introduce a lag for all requests made from that domain. So there be no questions that your formmail.cgi is pissing them in the eye.
Gnome 2 lacks easy configuration options. I don't understand the Gnome developers who think I'll install an desktop environment which my wife could not configure herself. And if I do the configuration for her, I'll hear pretty rough words if I have to use some mystical registry editor to change settings.
I'm not using Gnome anymore. Haven't checked out 2.2, but propably will some day. The main reason for not using Gnome lies not in Gnome, but in application support. I've found KMail to be the best email agent out there. It's features surpass Balsa and Evolution. And Konqueror is a superior browser compared to Mozilla. It just seems LAME to me to run KDE applications from Gnome desktop. So that's why I use KDE.
But if anyone prefers Gnome to KDE, that's OK with me. I just hope Sun Microsystems will make an effort to better Gnome like Apple is doing for KDE.
Slashdot Mirror
Why use IT staff? Schools are full of eager nerds who would love to hold an install fest...
There are many schools in Finland for example, which have student-maintained computer classrooms, school servers, web pages, etc. However, the higher level education a school offers, the less likely it is to find students maintaining the IT infrastructure...
This is indeed good news, as free software always is. But eRServer can only operate in single-master mode, which makes it unsuitable for high-availability kind of work. Single-master systems are good for load-balancing on installations where most of the queries to the DB are SELECTs.
eRServer comes a bit late. We already have PostgreSQL Replicator, which is multi-master. Unfortunately PG Replicator is not supported anymore. The latest version it can work with is 7.1, and the project's latest news are timestamped nearly two years ago.
Your mathematical utilities would be more useful if you had programmed them in C. Your choice of language will limit their adoption. Basically because using Perl scripts is not as fast as calling compiled C programs. This fact alone will make people reductant of using your utils in their code.
Because FreeBSD doesn't ship Perl as standard part of their distribution anymore, it'll be likely that your utils will not get included in any BSD software because it would pull in Perl. It may be a reason for Linux distributions too for not using your num-utils. Debian may be the only distribution which relies on Perl.
windowsupdate.microsoft.com has links for browsing the updates if your browser doesn't support activex. The activex-less pages have updates for 2000/NT and 98, but not XP, so you're left with the service packs.
But if you're not using IE or OE, you're probably safe from 99% of all the exploits Windows Update would patch. With a firewall and an anti-virus software you're safe from other kinds of exploits, or most of them at least.
BTW: Why did you remove IE? You could just make Mozilla your default browser and remove shortcuts for IE. Then there'd be no problems with Windows Update.
When the update contract expires, you'll have to get a newer version or switch to some other distribution. Redhat is on only offering the much-talked-about one-year upgrade period.
It's long time since RH5 and Debian 2.1. I suggest you take a look at Debian 3.0 (woody); It has improved much since 1998.
If you don't want to leave RPM-based systems, then you'd find SuSE an easy choice. I've used their distribution before I switched to Debian, and I don't have anything against them (save for RPMs). The YAST2 setup tool is really helpful, and you'd benefit from their participation in UnitedLinux and commitment to LSB and FHS.
If your RedHat installation doesn't give you what you need from a server operating system, it's a good time to think about switching. I would not rush it, since there is always the chance something is screwed and restoring an entire system back is not easily done even if you keep full backups. You should at least familiarize yourself with Debian before you start running servers with it.
The Debian install is not much different from the rest of the distros. If you know how Linux works, the text-based installation progress is quite simple. You can set it up to get packages from the net, so there's no need to burn all the CD's. The Debian web site has links to netinstall ISO images which are only 10-30 MB.
For me Debian install takes about half an hour. You might need considerably more on your first run. But heed this warning: Don't run task-sel nor dselect to pick the apps. I've yet to hear of a successful use of those two utilities. Especially for a server environment, you probably don't want to have all kinds of software lying around (both task-sel and dselect install tons of software you didn't want).
After you have your Debian system configured, it's a simple matter of apt-cache search'ing and apt-get install'ing the software you need. The dependency system will take care of the libraries and softwares to which the software you want depends on. After two hours you should be set up (depending on your network connection speed) with the software you need.
To keep up with security, choose only stable Debian packages. Then it's a simple matter of scripting to set up a cron job to do atp-get update and apt-get dist-upgrade periodically.
You might also want to take a look at Gentoo Linux, which offers similar packaging system to Debian. Gentoo philosophy is that you're provided with the package information and source codes which are compiled on your system for optimal setup; there's no binary distribution. I don't know of their security update model.
And if you're open-minded, there are always the BSDs.
So does this method allow DOSsing connections, with forged IP addresses?
I think signing with PGP/GPG is enough proof of autheticy. If the public key has been available on a public keyserver prior to the writing of the original story.
Besides, you can phone the company/person to verify that the response is authentic. If they withold their contact information, you cannot verify them, and thus cannot be made to publish their response on grounds it's origin unconfirmable. The key issue here is wether the writer of the original story is responsible for the costs of authenticating the response.
No central server, just a chain of peers, like peer2peer2peer2peer. Each peer would act as a server, serving connections and forwarding them, if needed.
Peer chains are easy to create. At first, you connect to a know peer from which you ask a list of peers known to it (and then you could poll other peers for the peers they know of). Then you'll have a pool of candidate peers through which you can connect to any peer on the network.
It's not "direct connect" when there are intermediate peers (like those ultrapeers in gnutella). However, the intermediates would be random and they would not know of each other.
Not quite so. Using a scheme akin of anonymous remailer that problem could be avoided altogether:
Clients establish encryped connections with masters. The connections are either direct (client to master) or forwarded (client to master to another master). For forwarded connections, the client does not tell from which machine the connection came from (it could be the machine or some of it's clients). So the server at the end of the request chain sees only the server immediately before it. That server sees only one before it and so on. None but the originating client knows which machine is the client.
Of course that would be slow, and quite un-interactive, but it would work well enough if each master would randomize the path it tries to follow to the final target. With a TTL-like requests, the number of intermediate servers could be controlled by the client. And the masters would never make up the route to the client, but use the route the client connected with them.
Not in small scale like single marriage. But psychohistory could be used to predict behaviour in nationwide or global scale.
Let's say: A nation attacks another nation, and leaves some valuable historical relics unguarded. What does psychohistory predict will happen? A number of people (note: not all) will loot the unguarded relics with the intend to get rich.
All you need for psychohistory to work, is to model historical events, and the variables that most affected them.
Sorry if I was not precise enough. I measured the frame rate with glxgears with default window size. That's something like 320x200.
Can you point me to some place where tuning them is discussed. I don't like to hack them myself. I've installed the FireGL drivers from ATI. They were the only drivers that I found from their site. Work fine with X, only OpenGL applications are slow, like 140 FPS with Radeon versus 250 with TNT2. My guess is OpenGL acceleration is not working, or something. But I don't know how to turn it on. nVidia's drivers were accompanied with nVidia's OpenGL implementation, but I didn't see anything like it with the FireGL drivers.
Genetic manipulation from a capitalistic world must support other capitalistic ventures. That's a sad fact we sometimes forget when we're thinking about advancements in science. If we lived in socialist world, then there would be a fair chance we'll get physically and mentally superior GM'd people.
But the first task of GM is not to create more intelligent people, no. The first task they'll have to undertake is to enable people to eat more unhealthy food without getting fat. Yes, beauty comes before brains in alphabet.
Besides, altering intelligence might not be so good an idea: it would generate more resistance to stupid laws, stupid politics, and stupid corporations. The people in charge of GM-corporations do not generally fall into the category of free-thinking liberalists. They're after money, not freedom.
I want speed in addition to quality. nVisia's inexpensive gForce4-based cards do not offer the same 2D image quality ATI does. It's just that I need to boot to Windows if I want to play games...
How do you know they're bogus? I know some users fake their User-Agent stings to match Win/MSIE to fool witty JavaScript writers. But I doubt Google would lie on purpose. Do you have any links that might shed some light to it? Like statistics which show there are more Linux users...
Sounds so much like the X-Men 2 slogans. Mutants, I mean GM'ed people, must stand together...
Googles Zeitgeist tells me 4% of people are using Macs whereas only 1% are using Linux. It would be more profitable to make a Mac game than a Linux game.
Besides, I don't think Linux's graphics acceleration support is so mature a casual Linux-user would get the best performance from their machine. At least my Radeon 8500 gives me less FPS than TNT2 I previously had.
And that goes for OO as well. Not every database (or a collection of data) needs to be accessed in Object-Oriented way. Most (or should I say all) data I store to small tables would not benefit from being objects.
And how does this differ from storing non-object-oriented data structures in RAM? You'd still need to implement searches, and how do you search an collection of objects without placing them on the relational line.
I'd like it if programs were desktop environment agnostics, or generally would not care of your library infrastructure.
It would be nice to have usefull applications which you could use with any widget library, and the application would conform to the environment; use Gnome if it's available and running, use KDE if it's the one the user likes, or use anything that's there.
Now that we have choice of desktop, the OSS developers need to do the same applications for multiple widget libraries. All the basic applications exist for all the libraries. Developers' time is wasted to reinvent the application using some different set of libraries. But it doesn't need to be that way.
Don't take me wrong: I'm not asking anyone to invent a library which would guess which other library to use. This needs to go deeper than that if it's to succeed.
Since most spammers use open email relays to send spam, why not attack the open relays? There's no reason I can think of for leaving a relay open, so they must be open because of either lack of knowledge or because the admin wants people to be able to use it without authentication.
So why not create bots that scan for open relays and start DDOSing them once found. I doubt open relays have enough hardware to avert angry spam-recipients counter-attack.
And if the relays are open because of lack of knowledge, the DDOS attack could send some message like "Your relay is open and I like spam THIS much."
That's because most spam messages have faked envelope senders. The bounced messages never reach the spammers. And if they did, you'd have that much easier to spam the spammers. They don't want that, and they can avoid it.
This idea could be turned into a distributed project, much like distributed.net. The idea would be to create a new mail filter, or filtering plugin which would pass all URLs found in any mail message stored to spam folder to the distributed network. Each participant of the network would automatically start crawling the URLs submitted in an easy pace so as not to have an effect for the user. But the combined effect would be dramatic for the targetted websites. Anyone stupid enough to go seeking for their stuff would time out due to server load.
And of course, the User-Agent would be faked to be MSIE.
Instead of returning NOT_FOUND, timeout the request. But you propably can't do that with Apache without damaging your own capability to serve pages. But if you can, you could also send OK response after a long delay. But not deliver any mail, of course. With OK, the spammer would presume your formmail.cgi honeypot relay was functional and continue to use it. However, due to the "lag" they can not "send" mail in great speed, damaging their capability to spam. And if you want to be really evil, the script could try to figure if the spammer tries to send mail to himself, and deliver such messages to keep the spammer fooled.
And once you spot a spammer IP, introduce a lag for all requests made from that domain. So there be no questions that your formmail.cgi is pissing them in the eye.
Gnome 2 lacks easy configuration options. I don't understand the Gnome developers who think I'll install an desktop environment which my wife could not configure herself. And if I do the configuration for her, I'll hear pretty rough words if I have to use some mystical registry editor to change settings.
I'm not using Gnome anymore. Haven't checked out 2.2, but propably will some day. The main reason for not using Gnome lies not in Gnome, but in application support. I've found KMail to be the best email agent out there. It's features surpass Balsa and Evolution. And Konqueror is a superior browser compared to Mozilla. It just seems LAME to me to run KDE applications from Gnome desktop. So that's why I use KDE.
But if anyone prefers Gnome to KDE, that's OK with me. I just hope Sun Microsystems will make an effort to better Gnome like Apple is doing for KDE.