I'm a proprietary developer and I'm out to make good software.
I am working on the theory that if you write good software then people prefer it to competition software that has more bugs and is harder to use and has less features.
Of course this theory breaks down when you have companies with big marketing machines but that isn't everyone:)
Re:a perfect game
on
Awari Solved
·
· Score: 3, Interesting
Not so; in some games the second player wins, here's an example:
you have a pile of 21 matches. players alternate turns. on your turn you may take either 1, 2, or 3 matches. whoever takes the last match LOSES.
Re:depends on what you call perfect
on
Awari Solved
·
· Score: 2
That looks like a paper-scissors-stone competition.. I thought roshambo was where you kick each other in the nuts as hard as you can until someone falls over?
Re:With enough storage, Chess could be solved too.
on
Awari Solved
·
· Score: 2
Not really, because the number of possible positions is bigger than the number of particles in the universe
You would have to come up with a way of encoding trillions of positions in one electron, or something
Maybe the reason Literate Programming never took off is because it wastes the time of good developers. I rarely have trouble reading code written by myself or the other developers at my work, and I can even completely rewrite someone else's software and get it almost all right.
I guess it would be useful for novice programmers who do not know how to write useful comments (that is, 'self-documenting' code, as well as actual comments).
If I spent the time writing document outlines and program plans and crap beforehand it would be just that, spending time for no reason; not to mention the design changes that often go on as you are actually writing code ; the last thing in the world I want would be to have to go back and change all these plans because I decided to change an aspect of what I was doing.
That's the whole point. It IS broken. But it will still run anyway, and you can't make an exact replica because your hardware can't create broken disks.
If you have just been appointed to lead this project, as you say, I would have thought the more important question would be: how should I design the toolkit so that it will be the most effective? (that is, easy to install, easy to get started, flexible, powerful, etc....) What are common problems other people have had with embedded system devkits?
Well you got 'karma' wrong, for a start. If you're going to criticize the editors (unpaid volunteers) you could at least put the effort in yourself to produce error-free material.
I've always found Netscape (and therefore Mozilla I guess) to handle security properly. (The fact that the rest of it is so horrible to use, not withstanding). In fact the article says that Mozilla is not vulnerable.
I'm annoyed that this is reported as 'making SSL insecure' or making a 'joke' of it. It isn't. It is a failure of the browser to verify the certificate authority chain.
With OpenSSL you can generate a certificate request, and then use another certificate to sign it (or, in this case, submit it to Verisign so that they can sign it with their certificate). You can then use this new certificate to sign more, and so on. So the chain might look like:
Obviously in this example www.fakeserver.com only belongs to the group of servers trusted by www.myserver.com and not to the group trusted by Verisign. The bug being reported is that IE and Konq mistakenly assign www.fakeserver.com to the group trusted by Verisign.
Now, what is the upshot of all this? What we have lost here, from the client's point of view, is the assurance that the server is who they say they are. Other aspects of SSL (secure encryption, inability for other parties to intercept connection, client validation) still work. A successful workaround would involve the person operating the client manually inspecting the certificate chain, and checking that *all* the sites on it are ones he/she trusts, not just the top one.
That's because doubles are the same size as ints, and CPUs process the fastest when using their int size.
You could save memory as you say by casting them to floats as you go. Or if you were really serious about only using floats, you could write a new library for float arithmetic and then use that. Can you do that in FORTRAN?
Troll - win98 counts its uptime as milliseconds, so it can't report an uptime of over 2^32 milliseconds (less than 2 months)
Re:Why are people pretending this is not a problem
on
Shattering Windows
·
· Score: 1
Windows regularly process thousands of messages per second. Checking each one for security credentials would slow down everything. I would even go so far as to say that on older hardware, when Windows 3.1 was being developed, it would have made the system completely unusable.
No - the problem still occurs even if VirusScan does check the size of the edit control. It occurs even if VirusScan never looks at the contents of the edit control. The edit control automatically allocates memory when it receives a PASTE message. Even if you then access the edit control and say "give me your first 4 bytes" (or "truncate to 4 bytes"), the entire paste is still in the allocated memory, which is enough for this exploit to work.
To prevent this paste you would have to either filter WM_PASTE messages (and any other messages that have similar functions), and/or filter the message which sets the maximum length of text.
Well, some functions don't provide sanity checking and some do. The advantage of C is that you can select a function that runs as fast as possible (and isn't slowed down by bounds checks).
If you program in C then you MUST understand exactly what the upshot of each of your statements will be, otherwise you are liable to write insecure code.
sprintf() and strcpy() are useful sometimes, for example:
is perfectly safe. No possibility of error. It will run faster than if the third line were
strncpy(buf1, buf2, sizeof(buf1));
and much much faster than
snprintf(buf1, sizeof(buf1), "%s", buf2);
On the other hand, gets() has no possible secure usage that I can think of. Nevertheless it can be useful in learning C (if only as a way of introducing the topic of buffer overflows!)
Oh yes they would; timers can only run callback functions (they can't simply set event flags or anything). Timers are used all over the place; any event that has to happen which isn't in response to some other event (this includes any animation, flashing items, etc.)
NB - It would be possible to implement animation by a thread with Sleeps, but you have to be pretty gay to make a whole thread just for that
Re:Microsoft has had 7 years of warning.
on
Shattering Windows
·
· Score: 2
The only way to fix this, at the OS level, would be to implement some sort of ACLs for windows messages. Apart from making things staggeringly slower, it would break all existing applications. Most Windows applications use timers.
Re:Shoddy work, mistake in the whitepaper!
on
Shattering Windows
·
· Score: 2
High-level IDEs (eg Visual C++, Visual Basic, C++Builder etc.) automatically provide message queue processing for each control. Even with the message you are suggesting, one would have to filter the calls to DispatchMethod. To secure an application against this attack would be a monumental piece of work -- to write your own message handler for every control, and filter out all possible harmful messages (or alternatively, filter in exactly the messages you want). And then for the dangerous messages (WM_TIMER etc). you have to figure out whether it was a real timer in your application, or a spoofed timer message.
I'd hate to see the source code of a program like this.
The source I downloaded (http://developer.kde.org/~sewardj/valgrind-1.0.0. tar.bz2) didn't have a 'vg_schedular.c' ; and the closest-looking filename ('vg_scheduler.c') did not contain any calls to socket or bind, or the strings 11, 0B or 0xB. Of course this is not to say that the file does not do this, because any backdoor-writer worth his salt would obfuscate it under lots of #defines and such. I would like to see the original poster (or anyone else with the same compiler) paste the compiler output, including the line numbers involved, and also say where he got the source from.
I'm afraid they aren't.. the Scouring is gonna be cut. Saruman gets killed on a spiky wheel somewhere. I think the rationale was that the long build-down after the climax wouldn't suit the movie format.
Slashdot Mirror
I'm a proprietary developer and I'm out to make good software.
:)
I am working on the theory that if you write good software then people prefer it to competition software that has more bugs and is harder to use and has less features.
Of course this theory breaks down when you have companies with big marketing machines but that isn't everyone
Not so; in some games the second player wins, here's an example:
you have a pile of 21 matches. players alternate turns. on your turn you may take either 1, 2, or 3 matches. whoever takes the last match LOSES.
That looks like a paper-scissors-stone competition.. I thought roshambo was where you kick each other in the nuts as hard as you can until someone falls over?
Not really, because the number of possible positions is bigger than the number of particles in the universe
You would have to come up with a way of encoding trillions of positions in one electron, or something
Maybe the reason Literate Programming never took off is because it wastes the time of good developers. I rarely have trouble reading code written by myself or the other developers at my work, and I can even completely rewrite someone else's software and get it almost all right.
I guess it would be useful for novice programmers who do not know how to write useful comments (that is, 'self-documenting' code, as well as actual comments).
If I spent the time writing document outlines and program plans and crap beforehand it would be just that, spending time for no reason; not to mention the design changes that often go on as you are actually writing code ; the last thing in the world I want would be to have to go back and change all these plans because I decided to change an aspect of what I was doing.
That's the whole point. It IS broken. But it will still run anyway, and you can't make an exact replica because your hardware can't create broken disks.
If you have just been appointed to lead this project, as you say, I would have thought the more important question would be: how should I design the toolkit so that it will be the most effective? (that is, easy to install, easy to get started, flexible, powerful, etc....) What are common problems other people have had with embedded system devkits?
Well you got 'karma' wrong, for a start. If you're going to criticize the editors (unpaid volunteers) you could at least put the effort in yourself to produce error-free material.
I've always found Netscape (and therefore Mozilla I guess) to handle security properly. (The fact that the rest of it is so horrible to use, not withstanding). In fact the article says that Mozilla is not vulnerable.
w .fakeserver.com]
I'm annoyed that this is reported as 'making SSL insecure' or making a 'joke' of it. It isn't. It is a failure of the browser to verify the certificate authority chain.
With OpenSSL you can generate a certificate request, and then use another certificate to sign it (or, in this case, submit it to Verisign so that they can sign it with their certificate). You can then use this new certificate to sign more, and so on. So the chain might look like:
[Verisign root certificate]
--->
[www.myserver.com]
--->
[ww
Obviously in this example www.fakeserver.com only belongs to the group of servers trusted by www.myserver.com and not to the group trusted by Verisign. The bug being reported is that IE and Konq mistakenly assign www.fakeserver.com to the group trusted by Verisign.
Now, what is the upshot of all this? What we have lost here, from the client's point of view, is the assurance that the server is who they say they are. Other aspects of SSL (secure encryption, inability for other parties to intercept connection, client validation) still work. A successful workaround would involve the person operating the client manually inspecting the certificate chain, and checking that *all* the sites on it are ones he/she trusts, not just the top one.
That's because doubles are the same size as ints, and CPUs process the fastest when using their int size.
You could save memory as you say by casting them to floats as you go. Or if you were really serious about only using floats, you could write a new library for float arithmetic and then use that. Can you do that in FORTRAN?
Troll - win98 counts its uptime as milliseconds, so it can't report an uptime of over 2^32 milliseconds (less than 2 months)
Windows regularly process thousands of messages per second. Checking each one for security credentials would slow down everything. I would even go so far as to say that on older hardware, when Windows 3.1 was being developed, it would have made the system completely unusable.
No - the problem still occurs even if VirusScan does check the size of the edit control. It occurs even if VirusScan never looks at the contents of the edit control. The edit control automatically allocates memory when it receives a PASTE message. Even if you then access the edit control and say "give me your first 4 bytes" (or "truncate to 4 bytes"), the entire paste is still in the allocated memory, which is enough for this exploit to work.
To prevent this paste you would have to either filter WM_PASTE messages (and any other messages that have similar functions), and/or filter the message which sets the maximum length of text.
Well, some functions don't provide sanity checking and some do. The advantage of C is that you can select a function that runs as fast as possible (and isn't slowed down by bounds checks).
/* or: sprintf(buf1, "%s", buf2); */
If you program in C then you MUST understand exactly what the upshot of each of your statements will be, otherwise you are liable to write insecure code.
sprintf() and strcpy() are useful sometimes, for example:
char buf1[16]; char buf2[16];
buf2[sizeof(buf2)-1] = '\0';
strcpy(buf1, buf2);
is perfectly safe. No possibility of error. It will run faster than if the third line were
strncpy(buf1, buf2, sizeof(buf1));
and much much faster than
snprintf(buf1, sizeof(buf1), "%s", buf2);
On the other hand, gets() has no possible secure usage that I can think of. Nevertheless it can be useful in learning C (if only as a way of introducing the topic of buffer overflows!)
Oh yes they would; timers can only run callback functions (they can't simply set event flags or anything). Timers are used all over the place; any event that has to happen which isn't in response to some other event (this includes any animation, flashing items, etc.)
NB - It would be possible to implement animation by a thread with Sleeps, but you have to be pretty gay to make a whole thread just for that
The only way to fix this, at the OS level, would be to implement some sort of ACLs for windows messages. Apart from making things staggeringly slower, it would break all existing applications. Most Windows applications use timers.
High-level IDEs (eg Visual C++, Visual Basic, C++Builder etc.) automatically provide message queue processing for each control. Even with the message you are suggesting, one would have to filter the calls to DispatchMethod. To secure an application against this attack would be a monumental piece of work -- to write your own message handler for every control, and filter out all possible harmful messages (or alternatively, filter in exactly the messages you want). And then for the dangerous messages (WM_TIMER etc). you have to figure out whether it was a real timer in your application, or a spoofed timer message.
I'd hate to see the source code of a program like this.
Yes, you must be living on a flat earth. Whenever I throw a ball it takes an elliptical path
The source I downloaded (http://developer.kde.org/~sewardj/valgrind-1.0.0. tar.bz2) didn't have a 'vg_schedular.c' ; and the closest-looking filename ('vg_scheduler.c') did not contain any calls to socket or bind, or the strings 11, 0B or 0xB. Of course this is not to say that the file does not do this, because any backdoor-writer worth his salt would obfuscate it under lots of #defines and such. I would like to see the original poster (or anyone else with the same compiler) paste the compiler output, including the line numbers involved, and also say where he got the source from.
Even open source developers get mod points sometimes...
Seeing a celestial body with your own eyes is so utterly far much more amazing than seeing pictures of it, no matter what the quality.
I'm afraid they aren't.. the Scouring is gonna be cut. Saruman gets killed on a spiky wheel somewhere. I think the rationale was that the long build-down after the climax wouldn't suit the movie format.
You forget the scene where she served the hobbits hot grits for breakfast
If you paid 50 grand for linux then there would be plenty of dealers to configure it for you
Windows 95 does, Windows NT doesn't; in NT you can change resolution and colour depth as much as you like and all the applications are happy