For example, this means that performance of these filesystems is sometimes much, much better when you have got a huge number of files in a single directory.
The key word there is "sometimes". Stephen Tweedie recently commented on the Linux Kernel mailing list that resierfs is significantly faster on empty filesystems, but slows down as the filesystem approaches 90% usage (which is pretty typical for a production box).
For good practice you might want to get a PC and install FreeBSD or one of the Linuxes
I'd avoid Linux. I may well love the OS, and have been using it since the Linus boot/root disk days, but I'd advise something else for learning how to admin the box. Linux makes life too easy, with the consequence that you get used to the niceties and are then stuck when confronted with an OS that doesn't have them (and most of the paid Unix admin jobs will have such an OS).
Writing bash scripts, for example, gives you some syntactic sugar, but little in the way of real added value over and above plain Bourne shell. But it means your scripts won't be portable, and when confronted with an OS without bash, you're stuck.
I'd recommend OpenBSD or Solaris, or preferably both. Both can be acquired at zero cost for PC hardware, and hence make good choices to play with. Try to do everything you do without resorting to adding extra toys to the system (via the ports collection or sunfreeware.com, for example). Some might claim that's making your life hard for the sake of it, but I'd say it gives you invaluable experience that you'll welcome later in your Unix admin career. Get exposure to as many different versions of Unix as you can lay your hands on, and learn the differences between them. I've met (and in fact, interviewed recently) too many admins that only know Linux, or only know Solaris. Ultimately, Unix is Unix, but if you can show exposure to a wide variety, you're demonstrating an ability to deal with the variance between systems. I've met AIX admins who didn't know how to use a system without smit/smitty, and hence are useless on any other version of Unix.
The Compaq DL360 is a rockin server that can handle dual 1.3 Ghz PIII's and can come with the Lights Out Board as well. Oh yeah it's 1U as well.
I stand by my original comment. The Lights Out board is overpriced, and useless in a 1U server. Yes, it can work in some (e.g., the DL360 that you mentioned), but only if you're prepared to sacrifice your single available PCI slot. Our servers tend to already have that used for external SCSI controllers, and hence the Lights Out board simply isn't an option. Our Sun hardware, by way of contrast, lets us have remote lights out management, an external SCSI connector, and still leaves a PCI slot free. Sun may be way behind the price/perfomance curve, but when it comes to making a managable server, they're still head and shoulders above the Intel crowd, and that makes up for an awful lot...
If you happen to have Compaq servers check out their Lights Out Management boards.
Yeah, but this is the sort of thing that should be built into the machine, not an extra option that you have to pay for. Furthermore, it's useless in a 1U rackmount server, which is arguably where you need it most.
Anything that isn't made by Belkin. Many people swaer by them, but for me, they've caused no end of trouble. Half the time, they don't switch when you request them to, other times, they'll switch of their own accord. And they suck at handling non-PC hardware. I can't use them with my SPARC, and my Alpha works intermittently at best through them.
PostScript is software. Not just PostScript itself, incidentally, but even PostScript fonts, which are actually programs that describe the font outline, rather than just a set of data points. This was a deliberate decision on the part of Adobe, specifically to afford font designers copyright protection over their works.
This is no different from downloading a tarball with a Makefile inside. You are downloading a script from the net and running it as root.
Speak for yourself. None of the code I compile myself is either compiled or installed as root. If the final deliverable calls for a setuid root program, then yes, I'll run the final make install as root, but only after checking the Makefile to see what it does.
The situation I would really warn against is running an unexamined script that isn't provided by a known author
But that's exactly what you're advocating. Since you don't know if the DNS for the download domain has been hijacked, you haven't a clue who the author of the script you're piping into a root shell is.
OK, so they took from BSD, gave to GPL, then mixed the result with something as absurdly proprietary as Curl.
But not only that, they're disappearing down the same route as Ximian, namely chasing Microsoft in the "features and convenience are more important than security" game. The more users get used to seeing installation instructions that involve piping the output of an arbitratory web download into a root shell, the more they'll start to believe that's just the way it's done. You or I might know better, but the average user that'll be thinking ofdoing this probably won't.
As far as I am aware there are no vendors that offer an ssh-like encrypted login for network equipment.
Cisco do. But given that we were quoted £12000 per router to add ssh support, we decided to stick with telnet, and roll over to Linux routers as time and circumstances permit (there are still some areas where Cisco kit wins out, but not as many as there used to be)
Which completely misses the point. Yes, it means that you can display remote applications on your directfb machine. But you can't do the reverse -- display your directfb application on a remote machine. At least, not without kludging multiple display targets into gtk, something akin to what the libggi folks did long ago.
I've been a fan of his for longer than I care to mention - his games were the only good thing about my Atari Falcon030
Which immediately shows you to not have been a fan of his work for as long as you might think. Real Minter fans look back with fondness on Gridrunner on the Vic 20 and Mutant Camels on the C64...
one can go from a Sun to an HP to an IRIX box, and, using CDE, be productive on each platform. I've been there
I, too have been there, and found myself equally unproductive on all of them until I got myself a decent fvwm2 setup configured:-) Note that despite the "common" in the name, there's not a great deal in common between the different CDE versions. Each vendor puts everything on different menus, for a start. Sun's menu layout is probably the worst I've come across, with Tru64 being among the better ones. Sure, you have common keyboard accelerators in all versions of KDE, and because they're all running dtwm, the windows look and behave the same. But the end user experience really isn't that consistent. Which kind of defeats the whole point...
On an LCD a pixel is a pixel, and they're sooooo crisp compared to a CRT. They say the pixels are blocky, the rest of us call that clarity. Awesome clarity compared to a CRT.
Call it what you will, but a high end CRT still looks better than a modern LCD. That slight blur with CRTs actually *improves* the image quality, due to the way the human eye interprets it, much in the same way that ink spread in a laser printer improves the output quality. Plus, of course, I've yet to see an LCD screen that can match a decent CRT on gamut, contrast, refresh rate or price. Furthermore, they tend to suck horribly when you're running them at anything other than their natural resolution. I'll guarantee that in time LCDs with surpass a CRT in just about every department, but for demanding users, they're just not there yet.
I took the pepsi challenge, and WMA won hands down.
Which, of course, is utterly irrelvant. I'm not going to reboot just so I can listen to music in WMA format. I want to listen while I get real work done. Personally, I like Ogg Vorbis, and with the headphones I use at work, I can't tell the difference between ogg and MP3. Thus I use Ogg, to ensure that in 10 years time, I'll still be able to legally listen to tracks I encode today. Not only that, but friends I've introduced to Ogg think it's sounds better then MP3.
Now, if only I knew how to input the Pound symbol on my US keyboard
More to the point, I wish CmdrTaco didn't. Then, perhaps, he'd use the correct £ HTML character entity, rather than a Latin1 pound (0xA3). Try setting the character encoding to something else (e.g., Cyrillic ISO-8859-5) and then look at the title of this story to see why.
We've been evaluating it here. It's OK, but little more. The user interface sucks in a major way. But it does at least give some degree of Outlook-like calendaring.
it shouldn't be too big a deal to set it instead to download the fscking banner and funnel it to/dev/null.
Except for those outside the USA who typically have to pay for bandwidth by the minute. I'm one of the lucky few with unmetered access, but for everyone else, not downloading banners actually saves money, as well as making for a less irritating browsing experience...
They can just check the access.log - however they will never know if it actually has been displayed.
Yep, or more likely, by the use of a web server module that does the sme thing without having to actually parse the logs. But that's not what they're claiming. In the article, they say that they "make contact" with the user's browser to determine if the ad has actually been displayed. The only way I can think of doing this is by embedding some JavaScript that checks to see if the page has been rewritten en route, and if so, posts something back to the web server, which can then modify its content accordingly. But even that won't be particularly effective, and your favourite blocking proxy should just be able to filter out the offending javascript anyway. And even if it didn't, it still wouldn't catch proxies that just serve a blank image instead of the requested ad. As far as the browser is concerned, it's been given the image it requested. I'm sceptical, but then all of my assumptions are based on having a sane browser. Who knows what MS have put in IE to give content providers control over the browser?
Re:Emacs - wrong, lying and criminally insane BUT.
on
VIM 6.0 is Out
·
· Score: 3, Insightful
more productive.:-)
It's late, I've just got back from work, and right at this moment, I don't care. So I'll respond. It's only more productive because you don't know how to use vi properly:-)
Slashdot Mirror
The key word there is "sometimes". Stephen Tweedie recently commented on the Linux Kernel mailing list that resierfs is significantly faster on empty filesystems, but slows down as the filesystem approaches 90% usage (which is pretty typical for a production box).
I'd avoid Linux. I may well love the OS, and have been using it since the Linus boot/root disk days, but I'd advise something else for learning how to admin the box. Linux makes life too easy, with the consequence that you get used to the niceties and are then stuck when confronted with an OS that doesn't have them (and most of the paid Unix admin jobs will have such an OS).
Writing bash scripts, for example, gives you some syntactic sugar, but little in the way of real added value over and above plain Bourne shell. But it means your scripts won't be portable, and when confronted with an OS without bash, you're stuck.
I'd recommend OpenBSD or Solaris, or preferably both. Both can be acquired at zero cost for PC hardware, and hence make good choices to play with. Try to do everything you do without resorting to adding extra toys to the system (via the ports collection or sunfreeware.com, for example). Some might claim that's making your life hard for the sake of it, but I'd say it gives you invaluable experience that you'll welcome later in your Unix admin career. Get exposure to as many different versions of Unix as you can lay your hands on, and learn the differences between them. I've met (and in fact, interviewed recently) too many admins that only know Linux, or only know Solaris. Ultimately, Unix is Unix, but if you can show exposure to a wide variety, you're demonstrating an ability to deal with the variance between systems. I've met AIX admins who didn't know how to use a system without smit/smitty, and hence are useless on any other version of Unix.
Fair point. My only experience has been with DL320s, which only have one.
I stand by my original comment. The Lights Out board is overpriced, and useless in a 1U server. Yes, it can work in some (e.g., the DL360 that you mentioned), but only if you're prepared to sacrifice your single available PCI slot. Our servers tend to already have that used for external SCSI controllers, and hence the Lights Out board simply isn't an option. Our Sun hardware, by way of contrast, lets us have remote lights out management, an external SCSI connector, and still leaves a PCI slot free. Sun may be way behind the price/perfomance curve, but when it comes to making a managable server, they're still head and shoulders above the Intel crowd, and that makes up for an awful lot...
Yeah, but this is the sort of thing that should be built into the machine, not an extra option that you have to pay for. Furthermore, it's useless in a 1U rackmount server, which is arguably where you need it most.
Anything that isn't made by Belkin. Many people swaer by them, but for me, they've caused no end of trouble. Half the time, they don't switch when you request them to, other times, they'll switch of their own accord. And they suck at handling non-PC hardware. I can't use them with my SPARC, and my Alpha works intermittently at best through them.
PostScript is software. Not just PostScript itself, incidentally, but even PostScript fonts, which are actually programs that describe the font outline, rather than just a set of data points. This was a deliberate decision on the part of Adobe, specifically to afford font designers copyright protection over their works.
Speak for yourself. None of the code I compile myself is either compiled or installed as root. If the final deliverable calls for a setuid root program, then yes, I'll run the final make install as root, but only after checking the Makefile to see what it does.
The situation I would really warn against is running an unexamined script that isn't provided by a known author
But that's exactly what you're advocating. Since you don't know if the DNS for the download domain has been hijacked, you haven't a clue who the author of the script you're piping into a root shell is.
But not only that, they're disappearing down the same route as Ximian, namely chasing Microsoft in the "features and convenience are more important than security" game. The more users get used to seeing installation instructions that involve piping the output of an arbitratory web download into a root shell, the more they'll start to believe that's just the way it's done. You or I might know better, but the average user that'll be thinking ofdoing this probably won't.
In 1995 or so, I saw a job being advertised, for which one of the minimum requirements was 10 years' Java experience...
Try here.
Cisco do. But given that we were quoted £12000 per router to add ssh support, we decided to stick with telnet, and roll over to Linux routers as time and circumstances permit (there are still some areas where Cisco kit wins out, but not as many as there used to be)
Which completely misses the point. Yes, it means that you can display remote applications on your directfb machine. But you can't do the reverse -- display your directfb application on a remote machine. At least, not without kludging multiple display targets into gtk, something akin to what the libggi folks did long ago.
...gonna have to dust off the 'ol ST one of these days and see if it still works
The PC version worked fine under DOSEMU last time I tried...
Which immediately shows you to not have been a fan of his work for as long as you might think. Real Minter fans look back with fondness on Gridrunner on the Vic 20 and Mutant Camels on the C64...
I tmay be humour, but it's the only way to get decent HTML out of word...
I, too have been there, and found myself equally unproductive on all of them until I got myself a decent fvwm2 setup configured
Simple question: were you expecting your site to get slashdotted when details of this interview were posted?
Call it what you will, but a high end CRT still looks better than a modern LCD. That slight blur with CRTs actually *improves* the image quality, due to the way the human eye interprets it, much in the same way that ink spread in a laser printer improves the output quality. Plus, of course, I've yet to see an LCD screen that can match a decent CRT on gamut, contrast, refresh rate or price. Furthermore, they tend to suck horribly when you're running them at anything other than their natural resolution. I'll guarantee that in time LCDs with surpass a CRT in just about every department, but for demanding users, they're just not there yet.
Which, of course, is utterly irrelvant. I'm not going to reboot just so I can listen to music in WMA format. I want to listen while I get real work done. Personally, I like Ogg Vorbis, and with the headphones I use at work, I can't tell the difference between ogg and MP3. Thus I use Ogg, to ensure that in 10 years time, I'll still be able to legally listen to tracks I encode today. Not only that, but friends I've introduced to Ogg think it's sounds better then MP3.
More to the point, I wish CmdrTaco didn't. Then, perhaps, he'd use the correct £ HTML character entity, rather than a Latin1 pound (0xA3). Try setting the character encoding to something else (e.g., Cyrillic ISO-8859-5) and then look at the title of this story to see why.
We've been evaluating it here. It's OK, but little more. The user interface sucks in a major way. But it does at least give some degree of Outlook-like calendaring.
Except for those outside the USA who typically have to pay for bandwidth by the minute. I'm one of the lucky few with unmetered access, but for everyone else, not downloading banners actually saves money, as well as making for a less irritating browsing experience...
Yep, or more likely, by the use of a web server module that does the sme thing without having to actually parse the logs. But that's not what they're claiming. In the article, they say that they "make contact" with the user's browser to determine if the ad has actually been displayed. The only way I can think of doing this is by embedding some JavaScript that checks to see if the page has been rewritten en route, and if so, posts something back to the web server, which can then modify its content accordingly. But even that won't be particularly effective, and your favourite blocking proxy should just be able to filter out the offending javascript anyway. And even if it didn't, it still wouldn't catch proxies that just serve a blank image instead of the requested ad. As far as the browser is concerned, it's been given the image it requested. I'm sceptical, but then all of my assumptions are based on having a sane browser. Who knows what MS have put in IE to give content providers control over the browser?
It's late, I've just got back from work, and right at this moment, I don't care. So I'll respond. It's only more productive because you don't know how to use vi properly