jcl

The MSN Search, AlltheWeb, Teoma, and Altavista bots are also lurking, but they apparently don't have accounts. Probably wouldn't post much anyway.

Interesting. The connector that came with the HP/Compaq nx9010 I'm using now also doesn't seat snugly. In fact, it fell out just now when I sat up. Nothing useful to add, except that I'm not sure I'd trust a replacement part from HP.

You're far from the only person who's wanted that sort of operator chaining to work. There are a geat many people who really want a < b < c to mean (a < b) and (b < c) especially in numerical or otherwise math-oriented code, but there are very few languages in which it does. The only one that comes to mind is Perl 6. (The Lisp family supports something equivalent, but not with that syntax.) It's a bit of a sensitive subject in programming language design. [Edit: Really bad typo. Though 'operator changing' is accurate, since the meaning of the operators would change depending o

The description of the 'and' operator reads "The expression x and y first evaluates x; if x is false, its value is returned; otherwise, y is evaluated and the resulting value is returned." if i[0] and i[1] and i[2] in pool: is equivalent to if ((i[0] and i[1]) and i[2]) in pool: every element of i is true, so this reduces like so if ((i[0] and i[1]) and i[2]) in pool: --> if (i[1] and i[2]) in pool: --> if i[2] in pool:

Indeed. The board does brute-force formatting with non-breaking spaces and line breaks instead of using the pre element or CSS, and sets a proportional font in code blocks. Really makes a mess of things. Sadly the WWW is slightly hostile to source code; it takes a rediculous amount of effort to make everything work all the time. On the bright side, XHTML2 looks like it will greatly improve the situation with the new blockcode (source code block, all formatting preserved), code (inline source code), var (variable name), l (line), kbd (sample user input), and samp (sample program output) ele

fdisk. You can extend an ext2 (and I assume 3) partition using resize2fs and some other utilities. Other filesystems, I don't know.

As long as IE dominates the market, they're more likely to lose customers if they follow the W3C recommendations.

The aesthetics of multiline strings always suck. By the way, you can toss all of the explicit newlines and tabs if you want. def instructions(): print """ Welcome to the game of tic-tac-toe --a game of man against machine-- you first need to choose to go first or second. If you choose to go first you will be X's. If you choose to go second you will be O's. you will choose your 'move' on the board by using the following key. """ instruc_list = [ '1','2','3','4','5','6','7','8','9' ] print_board(instruc_list) print ' ' (This board eats tabs characters, but they do wo

Slackware jumped from 4.0 to 7.0. Release 6.1 was a beta of what became 7.0. Anyway, almost every release of Slack is available through the mirrors.

The right shift operation shifts the bits in a value so many positions right. The value 0100 right shifted by one is 0010, by two is 0001, by three is 0000. Left shift does the opposite. For added amusement, shifts come in both arithmetic and logical varieties: arithmetic shifts retain or propogate the sign bit (keeping the sign of the value the same) while logical shifts do not. Besides their pure bit-twiddling value, arithmetic shifts can be used to multiply and divide by powers of two. Python, following C, uses the operators << and >> for integer left and right shift respect

Minor issue. POSIX does not in fact specify anything about the layout of the exit status word, so you can't rely on the right shift. Whether Python uses the two-octet format on all systems I don't know. Best to stick with WEXITSTATUS.

You're right. The exit status is a 16-bit value: the lower eight bits are the number of the signal that caused the process to exit or zero if it exited without a signal, and the next eight bits are the exit status. Both the 16-bit value and the second 8-bit value are called 'exit status' in the interest of confusion. It isn't Python's fault, it's replicating the behavior of the POSIX system() function. To extract the exit status (the useful one), you can use the function os.WEXITSTATUS() or shift the value eight bits right. The name 'WEXITSTATUS' is also POSIX's fault.

The return value of os.system() is the exit status of the command. ifconfig = os.system('/sbin/ifconfig | grep essi')

History suggests that it will die. Copy-protection that's strong enough to impede the determined seems to invariable result in discs that can't be played by a regular audio systems, and copy-protection that's weak enough to be compatible is no problem to circumvent. Either way it ends up dead even if it's still on the market. In this case it sounds like it's the copy-protection is partially software-enforced -- the discs themselves obviously can't track the number of copies made of a song -- and software copy-protection is hardly protection at all. On the other hand, even ineffective copy-pr

This is about the dozenth time they've rolled out copy-protected CDs in the last decade. None of the other programs survived much beyond the initial rollout and this one won't either. If it isn't withdrawn after consumer complaints, it'll be cracked.

Ubuntu should be fine. GNOME might be a wee bit heavy but people run it on less. Hardware support may be not be complete, in particular power-management might be a problem, but it's likely that everything you need will work. As for the WiFi card, you'll just have to try it. There are Linux drivers for some Linksys cards and there are ways to use the Windows drivers under Linux.

Yup. GoogleBot is Google's indexing softbot. If you watch it long enough you'll see it crawl all of the forums collecting pages. I don't think so, but some people may disagree.

That seems extraordinarily unlikely. PowerPC 970 is moribund and there's nothing to replace it. The only processors IBM is going to be shipping in volume at the right price are the console processors. Cell is so unbalanced it may not even be completely adequate for it's designed purpose and Xenon isn't much better. It's not impossible, but it might not work in Apple's favor. Portablity means seeking out the common denominator and OS X certainly ain't it. Cocoa would be in direct competition with Qt and GTK+ and it would not win. Without Cocoa, OS X is just an oddball, semi-closed source

Which, of course, is why Apple decided to use it just when Intel decided to abandon it. Apple and Intel were really made for each other: both try increadibly hard to produce great products only to fall short. Actually I guess that describes the entire PC industry. Well, at least Apple is in good company now. (Intel is sort of a tragic. They're a slave to x86. They've been trying to kill it since 1980, but it only grows stronger.)

Indeed. The software side hasn't been doing that well either. They've had have some major SNAFUs with OS X over the years and it seems like the defect rate is increasing. You haven't seen the lastest from Anandtech have you? The Xserve performance is catastrophic. Linux on roughly equivalent x86 hardware is over 10x faster on their multitasking benchmarks. And not because of the hardware. They don't really have any choice. Both AMD and Intel are putting multicore 64-bit processors at the head of their product lines. Microsoft is a bit of wild card. They would benefit from Apple's gro

It's misleading in places (one glaring example being the size of the IA-32 register set) but it's mostly just outdated. The Pentium M and PPC970 really changed the playing field. Intel is now pushing toward shallow pipelines, low-power, and concurrency, while the desktop PowerPCs were growing deeper and more power hungry every iteration. In case you haven't found them, or if anyone is interested: G4: L1: 32/32 KiB L2: 512 KiB G5: L1: 64 KiB I-cache/32 KiB D-cache L2: 512 KiB Pentium 4 6xx (as used in the Apple dev boxes): L1: 12k uops I-cache/16 KiB D-cache L2: 2 MiB Xeon (Pentium 4): L1: S

I don't use Python, but I've never seen any sign of C outside of the runtime and native modules. It may well have some kind of header-based automagic foreign-function interface capability. But so do some implementation of Common Lisp, and if you compare CL to C in public you'll want to keep your asbestos coveralls handy

Heh. You wouldn't believe how many times people have pointed me to that article. It's become a quarterly tradition. I never did think much of it, and it's now completely out of date. That shouldn't a problem. Main memory latency is uniformly lower on the x86 side and the Intel's NetBurst systems can match, and sometimes greatly exceed, the throughput of the G5 desktops. Cache sizes tend to be large (notable exceptions being the NetBurst L1 caches and all of Via's designs) and performance is generally quite good by necessity. It's often forgotten today, but the NetBurst wasn't just a lame

So I take it you're confident that none of your cow-orkers will stop in here

There was no indication he was a threat, IMO. The explanation he gave for why he wanted the information was plausible, and the request itself and that he made it there of all places I take as a sign that he lacks the motivation and, uh, wisdom to represent a real danger if he did have any malicious intent. If there was a chance in Hell that he could actually cause any damage on his own he wouldn't have posted that question, either because he wouldn't need the help or because he would be bright enough to know that it would be a waste of time. If he's a burgeoning SK he'll eventually find all