Forgot your password?

Comment: Re:Summary. (Score 2) 301

by Lisandro (#46722995) Attached to: Theo De Raadt's Small Rant On OpenSSL

Think it this way. Using the current memory management implementation on OpenSSL adding a test case for the exploit would not have triggered any issues. At all. Using a regular malloc() on any half-decent *nix system would've immediately triggered an alert when run through Valgrind or similar.

Reusing memory with a freelist on a sensitive library like OpenSSL is a problem waiting to happen.

Comment: Re:Some possible ways (Score 1) 745

by Lisandro (#46271645) Attached to: Mathematician: Is Our Universe a Simulation?

Again, wrong. You're (purposely?) mixing up redefinition of results vs the redefinition of algebra. Redefining the algebra behind division (and all other basic operations in the process) is a valid approach to tackle the division by zero problem. Redefining its result alone is not. And, incidentally, none of those two will do you any good with IEEE numbers.

You simply can't just say " x / 0 = 42 " without redefining division, substraction and multiplication. And all other operations in the process.

division by zero is sometimes undefined, but there is no natural reason for it to give an error. For example, IEEE defines floating point division by zero as infinity, whereas dividing an integer by zero is defined as an error.

The example is wrong since IEEE does not define division by zero as infinite to be a valid result. And, under the algebra rules used by IEEE floats there IS a very good reason for it to give an error. This is not a philosophical discussion; you don't need to sit down with a Fields medal mathematicians in order to understand why the above statement is incorrect.

Comment: Re:Some possible ways (Score 1) 745

by Lisandro (#46270921) Attached to: Mathematician: Is Our Universe a Simulation?

It can be defined at any time in any way. You could define it as always equaling 42.

No, you cannot. It is not only not useful, it breaks math - you can't arbitrarily define a division result without breaking all related operations, like multiplication and substraction.

What you can do, as maxwell demon correctly pointed out, is to extend algebra so division by zero makes sense. This means, of course, to redefine all basic operations, and since IEEE floats are built around basic linear algebra there is a natural reason why division by zero gives errors: it its either indeterminate or has no solution. When IEEE defined division by zero as infinity was not to provide a valid solution for it but to ease error handling afterwards.

You should consider toning down the patronizing. Your original statement is incorrect.

Comment: Re:Some possible ways (Score 3, Informative) 745

by Lisandro (#46264331) Attached to: Mathematician: Is Our Universe a Simulation?

Division by zero is mathematically undefineable.

If A * B = C and C / B = A, you can't have B being zero without C being also zero (in which case the equation is valid for all values of A, a.k.a undefined). For every other value of C the equation has no solution. The only reason IEEE defined division by zero as infinity was to make errors easier to handle.

Comment: Not "must reads" per se, but books i've enjoyed... (Score 1) 796

by Lisandro (#45846493) Attached to: Ask Slashdot: What Are the Books Everyone Should Read? no particular order.

"Passage", by Connie Willis. Yes, "The Doomsday Book" is probably her better known work, but "Passage" is immensely gripping. A light read but quite thought provoking.
"Contact" by Isaac Asimov. Forget the movie (which wasn't half bad either); the book is great and focuses more on the characters' religion and spirituality. I've recommended it to a number of relatives who hate sci-fi and they all thanked me.
"Atlas shrugged" by Ayn Rand. Yes, it is long. Yes, some dialogs are way longer than they should. And yes, it is heavy handed. Love it or hate it, it will make you rethink
"Brave new word" by Aldous Huxley. People keep saying we live in a 1984-future, but Huxley's work reflects our present much better than Orwell could ever imagine. Or fear.
"Rendevouz with Rama" by Arthur C. Clarke. Perhaps the first real page-turner i ever read; if the first chapter grabs your attention you'll have a hard time putting this one down.
"A man without a country" by Kurt Vonnegut. A series of essays by Vonnegut which are short, humorous, depressing and thought-provoking all at the same time. Highly recommended.
"Starship troopers" by Robert A. Heinlein. Also heavy-handed and infused with Heinleins' view of politics, but it is very well-written and leaves you pondering after the last page is done. Rico's discussions with his history and philosophy teacher are though-provoking indeed.
"In cold blood" by Truman Capote. Non-fictional account of the Clutter murders on 1959 which is so well written and researched that feels like a novel.
"Foucault's Pendulum" by Umberto Eco. Long and slow in parts, but stylishly definitely one of the best written works i've ever read. The thinking man's version of "The Da Vinci Code".

Comment: Re:lol (Score 1) 62

by Lisandro (#45765297) Attached to: Enlightenment DR 0.18: Improved Compositing, Wayland Support

Pun aside, i had E17 segfaulting a couple times (i've been using me as my DE of choice for over a year now) and it handles crashes surprisingly nice. The screen goes black for a second while the windows manager restarts and you get all your applications back, running as if nothing happened, along a "Sorry, i've crashed" dialog.

FORTRAN is for pipe stress freaks and crystallography weenies.