iowa through to denver is about the flatest, boringest drive in the entire world.
It may well be the flattest, boringest drive through the USA, or even through North America, but I seriously doubt that it's the flattest, boringest drive in the whole world.
Try driving Katherine -- Tennant Creek, NT some day. There is NOTHING between these towns, seriously NOTHING, except about 400 miles of straight road. There might be a few scattered towns, but nothing with more than a few hundred people in them.
Another long, boring stretch of road is on the Nullabor plain. Guess why it's called the nullabor plain? Hint: It's latin, null=no, arbor=tree. No trees, dead straight road, several hundred miles. And when I say dead straight, I mean DEAD STRAIGHT. No turns. Not even a bend. It's the longest straight stretch of road in the world.
I don't know; this was maybe a year ago (just before slack 8.1 came out); I haven't been back since. I love BSD's ports system (I have a couple of netBSD boxen), but when I used it, gentoo seemed like a great idea, poorly implemented.
Look, gentto might work for you, and work well for you; that isn't under dispute. My experience, a long time ago, was that what should have been a simple thing broke my system. I went back to slack, and never looked back. I like slackware, and have no reason to change from here, YMMV.
Oh, yes I agree that in principle the gentoo way of doing things is better. However, the implementation when I used it left a fair bit to be desired; there seemed to be dependency conflicts; ie, upgrading one thing fixed the dependencies for that, but broke something else that depended on the old version. There didn't seem to be all that much quality control.
Things may be better now, but in the words of Samuel L Jackson, I'd never know because I'd never eat the filthy mother fucker.
Yes, that's the process in Slackware, and it Just Works. However, when I tried almost the same thing (only difference was that I used the RH source RPM for 2.4.20 kernel), it wouldn't boot. Firstly it needed an initial ramdisk, then I tried using the old one, that had mismatched kernel version numbers (big surprise there!); then I fixed that up, still wouldn't boot, etc etc. Note that each time we were compiling a kernel from clean, which was taking 1/2 hr to do, so we eventualy gave up and just installed the latest beta...
I agree RPM tends to break on the kernel, but then I always install the latest kernel right after an install so I don't think about it. And a new stable kernel version later, a make oldconfig isn't too hard... I've never installed a kernel any other way, what's hard about doing it "manually"?
Being a slackware user at home, and forced to use redhat at work, I've been battling with trying to upgrade the kernel to v2.4.20 on the RH8.0 machine at work. We need 2.4.20 for a driver that's not in anything earlier; after 2 days piss-farting around compiling the kernel from the 2.4.20 source rpm from 8.1 beta, we finally bit the bullet and just upgraded the whole system. If you can tell me an easier way to upgrade the kernel, I'd most appreciate it...
Gentoo is not for me. I tried it a while ago, and it installed well, and went well, until I did an "emerge kdemultimedia", which then went ahead and upgraded almost the entire system over dialup, breaking the installation in the process. Back to Slack, where I can again have a rock solid box where I know what's in it, and can make my own decisions about what to upgrade (or not).
If you hadn't upgraded the bus etc on the newer systems, it wouldn't be very fast.
Which is exactly my point. The P4 3GHz should be *much* faster than 20x the speed of the P166, due to the better memory bus, larger cache, etc etc etc, but it's not. It's only 26x faster than the P166 at encoding MPEG2.
Not only that, but the pentiums were severely disadvantaged by choosing a motherboard without the usual L2 cache that older P1 motherboards had.
Given the above, it's somewhat surprising that the P4-3GHz was only 26 times fatser than the P166MMX at encoding MPEG-2. Besides the software probably being optimised for the P4, the superior cache and memory speed of the P4 I'd have thought would have influenced the speed difference more than a simple clock calculation. 26 * 166 nearly is 3GHz - not much progress there other than clock speed (for this app).
I dunno, I'll donate that box of floppies I have around here somewhere. Maybe if we all looked behind our collective couches we'd find enough floppies to take over the... ahem... back up the internet.
The internet can't be that big that it can't fit on a couple of floppy disks, surely?
Getting the most out of hardware is sometimes difficult when you dont fit the standard gamer user profile. I hope to see more reviews like this
Amen to that. My home system is quite similar to the review system (dual athlon MP), and I don't play games. Thus, looking at reviews of video cards is usually (for me) pointless. They give framerates in all the latest 3d games, but don't really give usability indications. Maybe if I'd seen this review before I bought my computer I'd have bought a different video card instead of the GF4 I decided on.
I don't know how much of a difference 2x processing will make in most games, and I'm certainly not likely to even consider that route for a gaming system.
Errr, that was the whole point - to look at video cards from a non-gamers point of view. Not all of us use our systems for gaming, you know.
If you want gaming reviews, go to a gaming site...
Especially since http is faster to connect to than ftp.
I disagree. Sure, it's easy to browse via http and get one or two files, but when you're trying to suck down the entire directory, http blows (excuse the pun).
What's faster for getting a whole directory than:
wget -t 0 -c ftp://ftp.server.name/path/to/dir/*
Doesn't work with http, because the directory listing doesn't work with wget, at least the version I have.
[disclaimer] I don't game, but I share an office with a guy who does. [/disclaimer]
Another issue that people have not mentioned yet is that 60fps is all well and good, but that's only either an average or a best-case. Once the screen fills with details, you can be sure that the framerate will drop. That's the issue - not so much the framerate itself, but an indication that as the detail goes up (more things to kill, whatever), the framerate won't drop to unusable levels.
The hardware side, that is. I'm the guy who makes functional circuits at short notice to do "something"; dead-bug style is my favourite. Many times have I made a working circuit out of surface-mount parts, usually in three dimensions.
It's actually a very good, robust technique for quick-n-dirty prototypes when there isn't the time for a circuit board (which is typically a two-week turn-around).
The company isn't giving Rick a loan. Rather, he's agreed to work for a paycheck. His credit history is irrelevant. The only thing that is relevant is his performance on the job.
And that's the funny part - it's Rick who should be asking the employer for the credit rating; after all, it's the employer who is going to owe him money after he's worked for them for a week.
There's a IN SOVIET RUSSIA joke in there somewhere, but I won't go there...
Umm, pretty much any PC/104 embedded system. In fact, pretty much any embedded system you'd like to buy. I have on my desk at the moment an intel xscale in a SO-DIMM form factor, which I have sitting on its own custom board.
That might be the case where you are, but here in Australia the defacto industry standard for EDA is Protel, which is windows-only. Our mech. eng guys use Solidworks, which is a windows-based app, and I also use Altera Max+Plus II Baseline, which is windows-only.
We simply aren't willing to pay more than around $7-8k a seat when there are apps available in that price range that work "well enough".
Slashdot Mirror
I fix most of that kind of problem by playing dumb: "Sorry, I don't know how to use MS Word, it doesn't run under Linux"
I saw a bumper sticker on a ute (ute == truck for merkins) the other day, which said:
iowa through to denver is about the flatest, boringest drive in the entire world.
It may well be the flattest, boringest drive through the USA, or even through North America, but I seriously doubt that it's the flattest, boringest drive in the whole world.
Try driving Katherine -- Tennant Creek, NT some day. There is NOTHING between these towns, seriously NOTHING, except about 400 miles of straight road. There might be a few scattered towns, but nothing with more than a few hundred people in them.
Another long, boring stretch of road is on the Nullabor plain. Guess why it's called the nullabor plain? Hint: It's latin, null=no, arbor=tree. No trees, dead straight road, several hundred miles. And when I say dead straight, I mean DEAD STRAIGHT. No turns. Not even a bend. It's the longest straight stretch of road in the world.
THOSE are long, boring drives.
Were you using a pre-release or beta?
I don't know; this was maybe a year ago (just before slack 8.1 came out); I haven't been back since. I love BSD's ports system (I have a couple of netBSD boxen), but when I used it, gentoo seemed like a great idea, poorly implemented.
Look, gentto might work for you, and work well for you; that isn't under dispute. My experience, a long time ago, was that what should have been a simple thing broke my system. I went back to slack, and never looked back. I like slackware, and have no reason to change from here, YMMV.
Oh, yes I agree that in principle the gentoo way of doing things is better. However, the implementation when I used it left a fair bit to be desired; there seemed to be dependency conflicts; ie, upgrading one thing fixed the dependencies for that, but broke something else that depended on the old version. There didn't seem to be all that much quality control.
Things may be better now, but in the words of Samuel L Jackson, I'd never know because I'd never eat the filthy mother fucker.
Yes, that's the process in Slackware, and it Just Works. However, when I tried almost the same thing (only difference was that I used the RH source RPM for 2.4.20 kernel), it wouldn't boot. Firstly it needed an initial ramdisk, then I tried using the old one, that had mismatched kernel version numbers (big surprise there!); then I fixed that up, still wouldn't boot, etc etc. Note that each time we were compiling a kernel from clean, which was taking 1/2 hr to do, so we eventualy gave up and just installed the latest beta...
I agree RPM tends to break on the kernel, but then I always install the latest kernel right after an install so I don't think about it. And a new stable kernel version later, a make oldconfig isn't too hard... I've never installed a kernel any other way, what's hard about doing it "manually"?
Being a slackware user at home, and forced to use redhat at work, I've been battling with trying to upgrade the kernel to v2.4.20 on the RH8.0 machine at work. We need 2.4.20 for a driver that's not in anything earlier; after 2 days piss-farting around compiling the kernel from the 2.4.20 source rpm from 8.1 beta, we finally bit the bullet and just upgraded the whole system. If you can tell me an easier way to upgrade the kernel, I'd most appreciate it...
Gentoo is not for me. I tried it a while ago, and it installed well, and went well, until I did an "emerge kdemultimedia", which then went ahead and upgraded almost the entire system over dialup, breaking the installation in the process. Back to Slack, where I can again have a rock solid box where I know what's in it, and can make my own decisions about what to upgrade (or not).
in order to develop an entire fucking gaming environment
I'm not a gamer, but if there was a fucking gaming environment, I'd be there!
Which is why my .bashrc contains the following:
You can take the boy out of DOS, but you can't take DOS out of the boy...
If you hadn't upgraded the bus etc on the newer systems, it wouldn't be very fast.
Which is exactly my point. The P4 3GHz should be *much* faster than 20x the speed of the P166, due to the better memory bus, larger cache, etc etc etc, but it's not. It's only 26x faster than the P166 at encoding MPEG2.
Not only that, but the pentiums were severely disadvantaged by choosing a motherboard without the usual L2 cache that older P1 motherboards had.
Given the above, it's somewhat surprising that the P4-3GHz was only 26 times fatser than the P166MMX at encoding MPEG-2. Besides the software probably being optimised for the P4, the superior cache and memory speed of the P4 I'd have thought would have influenced the speed difference more than a simple clock calculation. 26 * 166 nearly is 3GHz - not much progress there other than clock speed (for this app).
I don't know, maybe he has a BO problem...
I dunno, I'll donate that box of floppies I have around here somewhere. Maybe if we all looked behind our collective couches we'd find enough floppies to take over the ... ahem ... back up the internet.
The internet can't be that big that it can't fit on a couple of floppy disks, surely?
Again.
Amen to that. My home system is quite similar to the review system (dual athlon MP), and I don't play games. Thus, looking at reviews of video cards is usually (for me) pointless. They give framerates in all the latest 3d games, but don't really give usability indications. Maybe if I'd seen this review before I bought my computer I'd have bought a different video card instead of the GF4 I decided on.
I don't know how much of a difference 2x processing will make in most games, and I'm certainly not likely to even consider that route for a gaming system.
Errr, that was the whole point - to look at video cards from a non-gamers point of view. Not all of us use our systems for gaming, you know.
If you want gaming reviews, go to a gaming site...
Especially since http is faster to connect to than ftp.
I disagree. Sure, it's easy to browse via http and get one or two files, but when you're trying to suck down the entire directory, http blows (excuse the pun).
What's faster for getting a whole directory than:
Doesn't work with http, because the directory listing doesn't work with wget, at least the version I have.
[disclaimer] I don't game, but I share an office with a guy who does. [/disclaimer]
Another issue that people have not mentioned yet is that 60fps is all well and good, but that's only either an average or a best-case. Once the screen fills with details, you can be sure that the framerate will drop. That's the issue - not so much the framerate itself, but an indication that as the detail goes up (more things to kill, whatever), the framerate won't drop to unusable levels.
But as I said, I don't game, so what do I know
The hardware side, that is. I'm the guy who makes functional circuits at short notice to do "something"; dead-bug style is my favourite. Many times have I made a working circuit out of surface-mount parts, usually in three dimensions.
It's actually a very good, robust technique for quick-n-dirty prototypes when there isn't the time for a circuit board (which is typically a two-week turn-around).
when i just want to browse Slashdot [...] it's Windows
Um, Why?? IMO Galeon is a *much* better browser than MSIE, so why would you prefer to browse under Windows than Linux
The company isn't giving Rick a loan. Rather, he's agreed to work for a paycheck. His credit history is irrelevant. The only thing that is relevant is his performance on the job.
And that's the funny part - it's Rick who should be asking the employer for the credit rating; after all, it's the employer who is going to owe him money after he's worked for them for a week.
There's a IN SOVIET RUSSIA joke in there somewhere, but I won't go there...
"What do you guys do with your computers?" The answers were amazing [...] adding users and scanning for viruses.
That's because they all were too embarrassed to say: "Look at Porn!"
it is better of to start out clean and nice and pretty
That's why I use slackware when given the choice (have to run redhat at work...)
Be honst, what computer gives this opportunity?
Umm, pretty much any PC/104 embedded system. In fact, pretty much any embedded system you'd like to buy. I have on my desk at the moment an intel xscale in a SO-DIMM form factor, which I have sitting on its own custom board.
That might be the case where you are, but here in Australia the defacto industry standard for EDA is Protel, which is windows-only. Our mech. eng guys use Solidworks, which is a windows-based app, and I also use Altera Max+Plus II Baseline, which is windows-only.
We simply aren't willing to pay more than around $7-8k a seat when there are apps available in that price range that work "well enough".