I remember being *glad* to switch away from NS4 *to* IE5 ! And telling everybody about it...

Oh the irony...

" a struct which contains pointers to its state and all of the functions related to actions the scheduler needs to take."

Hmmm. Looks like the usual implementation of polymorphism common to mere mortals coding in object-oriented languages (like C++ and vtables).

Or is there something more to it ?

Yep, too obtuse for me. But with your (+1 informative) comment and link, it's a lot clearer.

And funny.

Thanks.

Been there, done that. 64bit kernels bring me nothing I need, as long as I have PAE.

As for "there is no excuse for code that isn't 64bit clean", that may be true in your mom's basement but for a company that needs to turn out a profit supporting an additional platform is a huge cost. We don't have any customer asking for 64bits versions of our software, so why bother ?

Besides, the first affordable 64bit processors weren't available until the early 2000s and they weren't that much of a success (remember itanium ?).

http://en.wikipedia.org/wiki/Physical_Address_Extension

Each process addresses at most 4GB of data (or is it limited by most OSes to 2GB ?), but the whole OS can address more (with the proper hardware support, which has been standard since ... long ago, I don't remember).

I do.

Most of my company's code is not 64-bits clean, so it's easier for me to run a 32bits OS (Linux) to run 32bits processes, even though I have 8GB of RAM. I usually load 6 to 8 processes simultaneously, as I work on very splittable data sets.

Besides, I expect fewer problems with proprietary software (nvidia drivers, flash) on 32bits Linux.

Your mileage may vary.

Heh. Statistics... :)

I guess that when they say "their completion times was below average" means that they took longer which is considered bad (wtf, how about appreciating the design&graphics ?), so the "completion time grade" was below average. Still, it's a pretty messed-up sentence.

Oh and by the way, there is still a possibility that the original sentence is correct if we assume abysmal results by the Solvers. I prefer my explanation, thoough.

Nice explanation, especially the "OOP in C" analogy. Thanks.

There is also apt-cache policy *packagename* which is mighty useful when tracking several Debian branches (e.g. unstable+experimental).

A good system is indeed one that can evolve and adapt to shifting requirements.

A good product/project manager is one that can say "NO" and prevent feature creep. But you need some backbone to say no to the boss/client.

Err.... You *did* realize there were three pages in the article ? TeX is at the bottom of page 2.

Don't let the fact that DB2 is an IBM product restrain you from mocking Oracle.

You obviously don't remember gnome 2.0

Well, perhaps having having written his post in latin instead of italian would have helped ?

WTF are you smoking ? Of course it's not "part of either language", it's part of the POSIX C API.

It's a system-level library, if you prefer.