That's my point. Even with this stupidly simple interface there is still far too much that can go wrong. Every extra piece of functionality increases the potential for things to go wrong. Basically the problem is that computers are open-ended and there's too much that's outside the seller's control. (That doesn't excuse sellers from supporting properly the software that *they included* with the machine.)
The reason computer support sucks compared to sofas is that a computer system (including software) is much more complicated. If you want to offer a product that Just Works, and support that doesn't require either great expertise, guesswork over the phone, or dealing with thousands of trivial problems, you'd need to make the product much simpler.
Imagine a computer with three buttons: Send Mail. Read Mail. Browse Web. And a keyboard, a one-button mouse, and a big 'Go' button for when the message is composed. You could support that easily enough, except when the user goes to a website which itself is broken. You'd need to certify websites to some standard which says they will work with your software, and (trickier) that their user interface works the way the user expects. And of course you can forget all about third-party software.
I don't think I would consider using sapdb until there was evidence that someone out there was maintaining it. SAP need to put up a proper website explaining how their DBMS compares to Postgres, Oracle, DB2 and so on, giving instructions on how to use it, and details of the development team. Otherwise it looks like just a collection of cobwebby old code (mixed Pascal and C? Yuck!) which they GPLed because there was no good reason not to.
What's the state of ACL support with Linux? I heard that they were kinda-supported in 2.2 but not stored in the file system - what's it like with 2.4, and can Samba use them?
Is this a first for TrustedBSD, or can you get the same ACL support with Solaris, Linux or other 'nixes?
There are just too many licences already! Do we need another? Off the top of my head, I can think of Open Content, the LDP's documentation licence, the GNU documentation licence, that group of teachers who wrote a licence for educational stuff, and of course the plain old GPL applied to non-software works.
Of course music may need its own provisions to reflect performing rights and the like, but that does apply to some other forms of art too. Couldn't the Open Content people extend the scope of their legalese to make it suitable for music too? And if the FSF and other documentation writers could join in, that would be great too.
Probably the YAL problem isn't as serious as with software, since you don't have an equivalent of 'linking'. Usually when two copyrighted works are distributed together it would count as 'mere aggregation' (to use the GPL's terms). Even so, it does look a bit silly.
Somebody did have the idea of using Napster for scientific papers - just an idea at present, I don't think it has been implemented. Search for 'docster'.
In this sort of contest, it's traditional that finding ingenious ways to exploit loopholes in the rules is a perfectly valid way to win. Then the rules are tightened up if ever the contest is re-run.
Mike Goldman didn't make it clear what 'total size' meant - I think most programmers if you asked them would assume the sum of file sizes, if you didn't specify otherwise. His fault, I feel. He should pay the $5000 and write the rules more carefully next time.
BTW, if you do assume that 'total size' is just the sum of file sizes, you can compress anything down to zero bytes. Number each possible sequence of binary digits - the empty string is zero, '0' is one, '1' is two, '00' is three, '01' is four, '10' is five, '11' is six, and so on. Then just create the required number of zero-length files.
The OpenSSH people use 3DES as the default cipher and include Blowfish as a faster but less secure alternative. I don't know what you mean by 'Blowfish128' - is that the same as plain Blowfish?
Even if Blowfish is designed to be more secure than triple-DES, that doesn't mean it is more trustworthy. It's newer and hasn't had the same amount of hammering on it over the years as 3DES has had. Triple-DES with three different 56-bit keys is generally considered a good symmetric cipher; its disadvantage is slowness.
But remember that regexps in Perl and other languages are not 'regular'. They have extended features which aren't available in true regular expressions, such as backreferences.
The point is not 'if it is encrypted, it will be secure'. That has never been true.
Rather, what you should remember is that 'if it is not encrypted, it is not secure'.
Personally, I think it is more important to get encryption in there - even with *bad* practices - than to worry about getting the last 1% of security from already-encrypted apps. For example, going from telnet to ssh with password-sending (your password is encrypted in transit) is a huge leap in security. Going from ssh password-sending to public-key authentication is only a small extra step, if anything. Choosing a long passphrase, or going from Blowfish to 3DES, are pretty unimportant for most people. Few crackers are going to see encrypted bits going over the wire and attempt to crack that - even if the passphrase might only be quite short. More important to focus on replacing the existing highly insecure protocols such as NFS.
The obstacle courses were scrapped after the first few series of Robot Wars. That's when I stopped watching - endless robot battles with nothing else are a bit too monotonous.
The house robots are usually much tougher than the competing robots, but they don't have an 'advantage' because they are not part of the competition. Normally they act only when you go around the edge of the arena, or are pushed there. Myself I find the 'perimeter patrol zone' with Killalot and chums much more interesting than Battlebots' rather limp ramps and saws.
Craig Charles _is_ annoying, it's true; but he doesn't talk for very long (unlike the two Battlebots presenters). If you've been watching the first series, with Jeremy Clarkson, be happy that he gets replaced soon.
I don't get it. On the early series of Robot Wars in Britain, various people (mainly some guy with a long beard) were introduced as being 'the champion of Robot Wars in the US', 'the founder of Robot Wars' and so on. This led me to think that the British Robot Wars was just an adaptation of an American show.
Then I heard some mumblings on Slashdot about how Robot Wars had died / sold out, the original creator had been crushed by the soulless TV networks, et cetera.
Now it seems that Robot Wars is being imported to America from the UK. But is it actually a re-import?
Don't we have the same thing here on Slashdot with people inserting lowercase junk to get past the lameness filter?
Another example was the 'readability checker' in use at one large company - it made sure sentences were short enough on average in all electronic mail. People got round it by adding a row of full stops to the bottom of each message.
Any computer-based attempt to filter human-readable content based on its _meaning_ is bound to fail - at least until AI gets to the stage where computers can understand as well as a human. (In other words, not for a long while...) The only kind of filtering that works is crude looking for strings, eliminating 'fuck' but allowing 'fuq'. Some people would be offended by the former but not the latter, so that kind of filtering might be useful.
3. You may copy and distribute the Program (or a work based on it, under Section 2) in object code or executable form under the terms of Sections 1 and 2 above provided that you also do one of the following:
a) Accompany it with the complete corresponding machine-readable source code, which must be distributed under the terms of Sections 1 and 2 above on a medium customarily used for software interchange;
There are a couple of other options, one of which is to provide a written offer to provide source code for the cost of distribution. But if you choose option (a), you are not obliged to give out source code to 'anyone who asks' - you just have to ship the source code together with the binary.
I don't think IBM's customers will like the new name. It contains one of those commie lowercase letters, and fails to include a dull-sounding number after the slash. How do they expect to get enterprise-class reliability with a name like that?
OS/380 might have been more appropriate - though even that is a bit racy.
Slashdot Mirror
Why should I learn Common Lisp rather than Scheme? Or why Scheme rather than Common Lisp?
That's my point. Even with this stupidly simple interface there is still far too much that can go wrong. Every extra piece of functionality increases the potential for things to go wrong. Basically the problem is that computers are open-ended and there's too much that's outside the seller's control. (That doesn't excuse sellers from supporting properly the software that *they included* with the machine.)
The reason computer support sucks compared to sofas is that a computer system (including software) is much more complicated. If you want to offer a product that Just Works, and support that doesn't require either great expertise, guesswork over the phone, or dealing with thousands of trivial problems, you'd need to make the product much simpler.
Imagine a computer with three buttons: Send Mail. Read Mail. Browse Web. And a keyboard, a one-button mouse, and a big 'Go' button for when the message is composed. You could support that easily enough, except when the user goes to a website which itself is broken. You'd need to certify websites to some standard which says they will work with your software, and (trickier) that their user interface works the way the user expects. And of course you can forget all about third-party software.
Does a sofa have any of these problems?
I've just poked around on SAP's site and seen that they _do_ have a reasonable amount about sapdb. Sorry, I should have checked that before posting.
Still, an independent comparison of sapdb with Postgres and Oracle would be very handy.
I don't think I would consider using sapdb until there was evidence that someone out there was maintaining it. SAP need to put up a proper website explaining how their DBMS compares to Postgres, Oracle, DB2 and so on, giving instructions on how to use it, and details of the development team. Otherwise it looks like just a collection of cobwebby old code (mixed Pascal and C? Yuck!) which they GPLed because there was no good reason not to.
What's the state of ACL support with Linux? I heard that they were kinda-supported in 2.2 but not stored in the file system - what's it like with 2.4, and can Samba use them?
Is this a first for TrustedBSD, or can you get the same ACL support with Solaris, Linux or other 'nixes?
There are just too many licences already! Do we need another? Off the top of my head, I can think of Open Content, the LDP's documentation licence, the GNU documentation licence, that group of teachers who wrote a licence for educational stuff, and of course the plain old GPL applied to non-software works.
Of course music may need its own provisions to reflect performing rights and the like, but that does apply to some other forms of art too. Couldn't the Open Content people extend the scope of their legalese to make it suitable for music too? And if the FSF and other documentation writers could join in, that would be great too.
Probably the YAL problem isn't as serious as with software, since you don't have an equivalent of 'linking'. Usually when two copyrighted works are distributed together it would count as 'mere aggregation' (to use the GPL's terms). Even so, it does look a bit silly.
Somebody did have the idea of using Napster for scientific papers - just an idea at present, I don't think it has been implemented. Search for 'docster'.
In this sort of contest, it's traditional that finding ingenious ways to exploit loopholes in the rules is a perfectly valid way to win. Then the rules are tightened up if ever the contest is re-run.
Mike Goldman didn't make it clear what 'total size' meant - I think most programmers if you asked them would assume the sum of file sizes, if you didn't specify otherwise. His fault, I feel. He should pay the $5000 and write the rules more carefully next time.
BTW, if you do assume that 'total size' is just the sum of file sizes, you can compress anything down to zero bytes. Number each possible sequence of binary digits - the empty string is zero, '0' is one, '1' is two, '00' is three, '01' is four, '10' is five, '11' is six, and so on. Then just create the required number of zero-length files.
'FreeLinux has always been at war with OpenLinux.'
We need a new moderation option (+1, Troll).
The OpenSSH people use 3DES as the default cipher and include Blowfish as a faster but less secure alternative. I don't know what you mean by 'Blowfish128' - is that the same as plain Blowfish?
Even if Blowfish is designed to be more secure than triple-DES, that doesn't mean it is more trustworthy. It's newer and hasn't had the same amount of hammering on it over the years as 3DES has had. Triple-DES with three different 56-bit keys is generally considered a good symmetric cipher; its disadvantage is slowness.
But remember that regexps in Perl and other languages are not 'regular'. They have extended features which aren't available in true regular expressions, such as backreferences.
The point is not 'if it is encrypted, it will be secure'. That has never been true.
Rather, what you should remember is that 'if it is not encrypted, it is not secure'.
Personally, I think it is more important to get encryption in there - even with *bad* practices - than to worry about getting the last 1% of security from already-encrypted apps. For example, going from telnet to ssh with password-sending (your password is encrypted in transit) is a huge leap in security. Going from ssh password-sending to public-key authentication is only a small extra step, if anything. Choosing a long passphrase, or going from Blowfish to 3DES, are pretty unimportant for most people. Few crackers are going to see encrypted bits going over the wire and attempt to crack that - even if the passphrase might only be quite short. More important to focus on replacing the existing highly insecure protocols such as NFS.
But Millionaire was based on The $64000 Question.
The obstacle courses were scrapped after the first few series of Robot Wars. That's when I stopped watching - endless robot battles with nothing else are a bit too monotonous.
The house robots are usually much tougher than the competing robots, but they don't have an 'advantage' because they are not part of the competition. Normally they act only when you go around the edge of the arena, or are pushed there. Myself I find the 'perimeter patrol zone' with Killalot and chums much more interesting than Battlebots' rather limp ramps and saws.
Craig Charles _is_ annoying, it's true; but he doesn't talk for very long (unlike the two Battlebots presenters). If you've been watching the first series, with Jeremy Clarkson, be happy that he gets replaced soon.
I don't get it. On the early series of Robot Wars in Britain, various people (mainly some guy with a long beard) were introduced as being 'the champion of Robot Wars in the US', 'the founder of Robot Wars' and so on. This led me to think that the British Robot Wars was just an adaptation of an American show.
Then I heard some mumblings on Slashdot about how Robot Wars had died / sold out, the original creator had been crushed by the soulless TV networks, et cetera.
Now it seems that Robot Wars is being imported to America from the UK. But is it actually a re-import?
Don't we have the same thing here on Slashdot with people inserting lowercase junk to get past the lameness filter?
Another example was the 'readability checker' in use at one large company - it made sure sentences were short enough on average in all electronic mail. People got round it by adding a row of full stops to the bottom of each message.
Any computer-based attempt to filter human-readable content based on its _meaning_ is bound to fail - at least until AI gets to the stage where computers can understand as well as a human. (In other words, not for a long while...) The only kind of filtering that works is crude looking for strings, eliminating 'fuck' but allowing 'fuq'. Some people would be offended by the former but not the latter, so that kind of filtering might be useful.
IANAL, but I've read that if the company sends out a confirmation, they are legally obliged to fulfil the order. Which makes sense.
Which countries are the ones that want internationalized domain names? Probably those which don't use ASCII. Which country was the patent granted in?
Couldn't the DNS servers be run in the countries where they are needed, where they won't be affected by the screwed-up US patent system?
What the GPL says is:
There are a couple of other options, one of which is to provide a written offer to provide source code for the cost of distribution. But if you choose option (a), you are not obliged to give out source code to 'anyone who asks' - you just have to ship the source code together with the binary.
Wasn't the first microprocessor, the Intel 4004, priced at $370 to convey the message that it could do 'anything' one of IBM's big machines could?
Go away and read what the GPL says. Then come back.
(It's at http://www.gnu.org/copyleft/gpl.html.)
I don't think IBM's customers will like the new name. It contains one of those commie lowercase letters, and fails to include a dull-sounding number after the slash. How do they expect to get enterprise-class reliability with a name like that?
OS/380 might have been more appropriate - though even that is a bit racy.
But does it run Linux?