/var is for variable data. While web pages change, it's not exactly 'variable'.
/usr is good for the binaries, but the data is updated too much. Besides, it's common to mount/usr read only for security reasons.
/opt? That's for package data. An apache package could live there, but again the data changes, and opt could be mounted read only for the same reason/usr is.
/home? doesn't really fit either. There's a apache user for it, but it doesn't give multiple users easy access to the data. Besides, it's common (sorta) to mount/home with no_exec. So much for php, or perl.
So where should the data files for all applications live? There's not really a good place for them.
The only thing that someone can do cleanly is to have a seperate disk volume for that application (or several volumes). Call it/apache or/web.
I personally keep DNS data in/var/named, but I really shouldn't for the same reasons apache data shouldn't live there. I only do it because everyone else does.
We theoretically could claim class action law suit. Who's going to pony up the legal fees to pay the lawyers? You?
Besides, it's likely that no court would even see the case because the SCO vs. IBM case hasn't come to fruition. And *IF* SCO wins, there's no class action case at all. All their actions to 'harm' us would be perfectly in their rights.
It's frustrating, but until this thing is seen by a judge, nothing can happen.
1) If you have 56 internet facing DNS servers, it might be time to re-visit your design (with the possible exception of very large ISP's). Given BIND's history of security flaws, minimizing exposure is key.
2) With that many servers, if you're not doing it with package management (solaris pkgadd, rpm, deb, hp..., AIX..., etc... all of them have at least a rudimentry package management tool, even if it's tar), you might want to re-visit your design.
3) Deploying BIND without some forthought is going to get companies in trouble. If you can't be bothered, you really should just use DJB's DNS cache and DNS server.
Realistically, this isn't targeted at the average joe (or even the average/. reader (well, not yet)). This is targeted at the HA cluster's we're spec-ing out at work:
3 gigabit interfaces (2 in redundant failover mode, one for backups), 6 fiber channel disks on 2 diskplanes, and 2 qlogic HBA's to our EMC array.
Each one of those cards has quite the ability to saturate a single PCI bus. Thankfully, the boxes we're putting them into have 4 different PCI busses, so we can put 1 fiber or 1 HBA onto each.
End why is it that all the legal stuf seems to have become so important in the Open-Source/Unix® world.
{sigh} You obviously havn't been around the Unix wolrd for very long. The unix world has been beset by hideous legal issues since at least the early 80s:
Just who owns the unix trademark? I can think of at least 5 different owners.
The BSD 4.4 -> BSD Lite stupidity that stopped the *BSD's cold for at least a year (without that, it's quite possible that/. would be a BSD focused forum)
IIRC, there were several Xwindows legal challenges, OpenWin never came to fruition, etc, etc, etc
and those are just the ones I can think of on a couple hours of sleep.
I see (with much excitement) that tivo's now grok home networking, and playing of MP3's. Poking around the website it seems it only works with their software on "PC's and Mac's". The obvious question is can I hack it to play from Linux? If so, they just got a new convert.
You're used to 'home storage' prices. Look at pricewatch, find a good brand of EIDE, and just get it.
They're looking at 'enterprise storage'. We have 11 tera of raw disk on an EMC. It cost $2 million. The useable storage out of it is around 3-4 tera, after counting mirroring, and third mirror break off for backups, etc, etc, etc.
These drives use MCA (iirc) interconnects to a disk backplane, and fiber channel interconnects between disk boards and the front end san switch. The computers are fiber connected into the san switch as well, and the JNI cards (client end of a SAN connection) for this are NOT cheep.
To Online storage companies, downtime costs serious money. They can't afford the downtime. That's why their storage costs real money. Then they pass it on to you.
If you need real amounts of data, you don't want a hosting service, you want a CoLo service (They give you rack space, and an internet connection. You provide the box). If you want, you can put a desktop with 2x140 gb drives, and you'll get what reliability you can out of it (most IDE drives are warrenteed for 1 year for a reason). If you want the thing to last, get a server class, rack mountable server from (dell|compaq|ibm|penguin computers). You'll be happy you did. Mirror the drives (preferably in hardware) so you can loose a disk without killing your service.
And it might already be there. Please let me know.
I want ARBITRARY keyboard shortcuts. I want to be able to write a shell script (or any executable), and have it execute when I hit (ctrl)-(alt)-w (My keystroke to bring up a vertically maximized terminal window).
I was quite scared with gnome 2.x when they seemed to take this feature away, but I found out how to do it eventually (gconf-editor under the metacity stuff).
And sendmail will stop. No remembering obscure syntax. No weird rules. All of it (should) be encoded in that script.
The problem with traditional BSD init files is that it is deeply challenging to add/remove the code to start/stop a package because it has to insert code into a single script. If there's an out-er 'if' statement that the code doesn't expect, the program might not start. Well, it's somewhat easy to add. It's almost imposible to cleanly remove, especially if the admin has modified the script in any way. With SysV, you just drop the script in place, and add some symlinks. When you remove, you delete these files.
Granted, I havn't played with a recent *BSD to see how the init scripts have evolved. So, note the 'traditional' above.
Re:Advice for switching wife's computer to Gentoo
on
Gentoo Reviewed
·
· Score: 1
I love Gentoo, but you're not going to get mozilla and evolution installed this evening.
Just getting gnome going took from 10:00 PM to 8:30 AM on my 850 MHz laptop with 256 MB. If you do a stage3 build, you'll have CLI done by the end of an evening. Either gnome or KDE will take the time while you sleep and do work the following day. Mozilla will take 2ish hours, and evolution will take 4-5ish (IIRC. I havn't done that yet on this hardware).
My suggestion: Get computer up to CLI, and Friday after work run:
emerge kde gnome gdm mozilla evolution
It probably will be done by Sunday.
Re:My experiences with Gentoo
on
Gentoo Reviewed
·
· Score: 1
Reguarding performance, there is a difference. Anything CPU intensive (and especially math intensive) is going to get a nice boost. LAME went from 2x encode to 4-5x encode by recompiling after playing with compiler settings (from defaults to using -march, -mcpu, -pipe, -fomit-frame-pointer, and a few others)
And reguarding 'whatever crappy defaults', that's also a yes and a no. I for one hate the fact that MTR will link to GTK if it finds it. It should forever be a command line app. So I can easily do this:
USE="-gnome -gtk" emerge mtr
Now, as for configuration defaults, gentoo tends to take whatever the source code comes with (for example, no bluecurve).
Unfortunatly, the 'I cost them more money in support' argument doesn't wash. Accountants split money into different 'buckets'. Revenue comes in, and is devided into capital, Operations and Maintance, and Profit. (I'm over simplifying).
Capital money is money given to projects once. It's usually a bigish wad, and it's for a year. You have this much money to get this many new things to do this good work for us (IE: We need to implement a new source code version tracking package. Needs this much hardware, this much software, and will cost this much in support the first year).
O&M money is 'mantaiance' money. It costs this much to keep this web server up and running, and supplied with bandwidth. Salaries come out of this pool as well. O&M is usually a sunk cost. Just the cost of doing business. Support for their product is certainly a sunk cost. They invest a flat fee in salaries, headcount, phones, etc. From my CVS example above, maintance on the hardware, and software would be added to the O&M budget for the second year of it's life.
To the accountants, someone being on hold for 6 hours, and wasting 1 hour of 4 employees time isn't 'wasting' more money than the revenue from purchasing the product. It's just averaged in with all those people who DIDN'T call intuit, and just used TurboTax on one computer (like me (i bought it before I heard of the DRM games)).
Both of you are kind of right (Note: I've played more MK64 than I should have).
Bowser, Wario, and DK have the fastest top end. The problem is that they take a long time to get up there.
Toad, Peach, and Yoshi have a slower top end, but they get up to it much much faster.
Discounting powerslides for the moment, the heavy guys also lost a decent bit of speed in the curves. This is where the light guys really shine. Powerslides equalize things somewhat. Usually in a race, the little guy will still come out on top, but it's a narrower victory.
Now, if we're talking battle mode, then I break out the Wario, and you're all fried.;-)
What audiophiles realize is that equipment makes a huge difference in sound quality. What they don't realize is that people's audible memory (memory for the quality of the sound) is about as bad as our memory can be. If you go to an audio shop, they usually 'switches' that let you listen to different gear, instantly. You're listening to one receiver, then 'switch, you're music is coming through a different one, in about one second. Then you can hear the differences. However, if you drive 5 minutes to someone's house and listen to their rig, you're not going to be able to tell which sounded 'better'.
That said, stereo and home theater rigs are going to sound better than computer rigs, if for no other reason than better speakers. Computer speakers can't use large magnets, where sterio speakers can.
If you switch back and forth between CD source, and MP3 source for the same track, you'll be able to tell a difference (unless you encoded to 256 kbit). However, if you used a decent encoder, and a decent bit rate (128+) it will sound acceptable.
To give a final answer to the question above, the reason you listen to MP3's over your stereo is convienience. No changing CD's, no waiting on the 200 disk changer to rotate 180 degrees to find the disk you want, better UI for selecting disks/tracks, better randomizer functions, custom playlists, etc. It's fantastic for background noise in a party, and great when you're puttering around doing things.
If you're just relaxing, and focusing on the music, then you'll probably want the CD.
While Nintendo does have a broadband adaptor out, titles for it are a little thin on the ground. Traditionally, Nintindo has always focused on 'kart' style multiplayer. 4 people on a couch, trash talking, and having a great time. F-Zero, Mario party, and several others are examples of this.
I just hope the new Mario Kart can live up to expectations. The n64 version is still amazingly fun to play multiplayer.
The new Kart will be internet'able. I don't know if armada will be.
"It's hard to beat the bandwidth of a stationwagon filled with DAT tapes"
If you're talking: 1) several servers and 2) 200ish gigs per, welcome to needing a real backup solution.
One thing to keep in mind is the three 'kinds' of backups. You will need to cover (or choose not to) all three.
1) DR. Disaster recovery. A full image of ALL data, usually duplicated so you have a in house copy and a remote copy. Full system images, and a software package that can blast a full system image to a box, or full data and config backups that require an OS install before your restore. Usually this is somewhat light on tapes, since you'll only keep 2-3 weeks of them.
2) File Recovery. Someone deleted something that they shouldn't have and need it restored. Or the Database equivelent: "We dropped this table 5 weeks ago, and discovered just now this random important process that hits it every 2 months. Can we restore the DBF file so we can get that table, data and schema back?" Sometimes DR feeds into File Recovery. You just keep the tapes longer. More expensive in tapes though, and you have data you'll not use (like OS images) wasting tape space. It's easier though.
3) Archival. EG: The IRS mandates that we keep this data for N years (where N is usually greater than 7). Thankfully, this is a thin ammount of data, but it's important none the less. CD/DVD rock for this, but tapes are good too (so long as you're under 10 years. Media and reader issues will kill you after that).
Good luck. Backups are a huge pain. Be sure to test the DR portion of it at least once a year. You'll be thankful you do.
1) It's hyper expensive 2) the support organization is far from fantastic 3) It requires expensive talent to maintain
The answer to all of these issues is 'Name recognition'. Joe Average has at least heard of Oracle. CEO's and CIO's associate it with quality.
Currently, for 'high end' databases where 'high end' is defined by either CEO's or serious amounts of data, Oracle and DB2 (nee UDB or Universal Data Base) are wins.
To a corperation, nothing is more important than their data. If it can't be protected by a name they trust (whether or not that trust is misplaced) it will not penetrate the DataCenter.
Ok. I've been eyeing the sunfire line (cinema grand, and theater grand lines). From what I've HEARD, their pre-amp's arn't that good at the video switching, but are find for audio. The thing that really interests me though is their 'step down transformer' amps. How do they sound to you? The lack of heat, and the nice power ratings are good pros, but it's a rather un-orthodox way of amping a signal (and in this day of surround sound, I'm not sure if the assumptions will hold up).
I know that I really should just go and listen to them (and I will if I ever go to buy one), but I'm lazy, so I'm asking you.:-)
*I* heard her say you were an idiot i *HEARD* her say you were an idiot i heard *HER* say you were an idiot i heard her say you *WERE* an idiot
Four rather different meanings depending on the inflection. All languages are tonal, to a certain degree (even esperanto), and programing languages (being written) need to live with out the tonality. Therefore, they'll always be different.
The Native Posix Threading Lib (IIRC) is something Ingo cooked up in addition to the O(1) scheduler, and a few other goodies. Previously, posix threads could only have a couple hundred threads going concurently on ANY hardware. It just couldn't scale past that.
With NPTL, Ingo on a dual proc box (granted, a nice one) was able to get 16,000 concurent threads going, and the IO system wasn't suffocated, the CPU's wern't useless, and you could still browse the web.
Granted, these threads wern't doing much, but they were alive, and switching in and out of context.
This means the foundation can scale to effectivly any size, and so long as the hardware can keep up, you'll be fine. You can now unleash your massivly multithreaded java apps (and what not. That's just the easiest example).
This doesn't help you if you need more than 4-8 CPU's on an intel platform, but it gets you a lot closer. If you want something that can parallelize that far, you really need something like Sun's e12k or e15k. IIRC, the DoD commissioned an e15k farm with a total of 4096 CPU's to model the first few nanoseconds of nuclear explosions. It had to be a single system image for various reasons, so don't go crying beowulf.
Slashdot Mirror
/var is for variable data. While web pages change, it's not exactly 'variable'.
/usr is good for the binaries, but the data is updated too much. Besides, it's common to mount /usr read only for security reasons.
/opt? That's for package data. An apache package could live there, but again the data changes, and opt could be mounted read only for the same reason /usr is.
/home? doesn't really fit either. There's a apache user for it, but it doesn't give multiple users easy access to the data. Besides, it's common (sorta) to mount /home with no_exec. So much for php, or perl.
So where should the data files for all applications live? There's not really a good place for them.
The only thing that someone can do cleanly is to have a seperate disk volume for that application (or several volumes). Call it /apache or /web.
I personally keep DNS data in /var/named, but I really shouldn't for the same reasons apache data shouldn't live there. I only do it because everyone else does.
Unfortunatly, how?
We theoretically could claim class action law suit. Who's going to pony up the legal fees to pay the lawyers? You?
Besides, it's likely that no court would even see the case because the SCO vs. IBM case hasn't come to fruition. And *IF* SCO wins, there's no class action case at all. All their actions to 'harm' us would be perfectly in their rights.
It's frustrating, but until this thing is seen by a judge, nothing can happen.
How else can we 'take the lead' from SCO?
1) If you have 56 internet facing DNS servers, it might be time to re-visit your design (with the possible exception of very large ISP's). Given BIND's history of security flaws, minimizing exposure is key.
2) With that many servers, if you're not doing it with package management (solaris pkgadd, rpm, deb, hp..., AIX..., etc... all of them have at least a rudimentry package management tool, even if it's tar), you might want to re-visit your design.
3) Deploying BIND without some forthought is going to get companies in trouble. If you can't be bothered, you really should just use DJB's DNS cache and DNS server.
Subject: You have been selected...
You have been selected as a recipient of spam. Go to this website to collect your damages. Make money fast.
Laws in those states prohibit such contests. If you read almost any contest rules, they will have the exact same verbage.
Some friends of mine snaged Brute Force for the xbox, and to listen to them, it blows their minds. I havn't played it yet, but they give rave reviews.
Realistically, this isn't targeted at the average joe (or even the average /. reader (well, not yet)). This is targeted at the HA cluster's we're spec-ing out at work:
3 gigabit interfaces (2 in redundant failover mode, one for backups), 6 fiber channel disks on 2 diskplanes, and 2 qlogic HBA's to our EMC array.
Each one of those cards has quite the ability to saturate a single PCI bus. Thankfully, the boxes we're putting them into have 4 different PCI busses, so we can put 1 fiber or 1 HBA onto each.
End why is it that all the legal stuf seems to have become so important in the Open-Source/Unix® world.
{sigh} You obviously havn't been around the Unix wolrd for very long. The unix world has been beset by hideous legal issues since at least the early 80s:
Just who owns the unix trademark? I can think of at least 5 different owners.
The BSD 4.4 -> BSD Lite stupidity that stopped the *BSD's cold for at least a year (without that, it's quite possible that /. would be a BSD focused forum)
IIRC, there were several Xwindows legal challenges, OpenWin never came to fruition, etc, etc, etc
and those are just the ones I can think of on a couple hours of sleep.
Where should I look to get a refurbished s2 tivo? Froogle.google.com couldn't find one.
I see (with much excitement) that tivo's now grok home networking, and playing of MP3's. Poking around the website it seems it only works with their software on "PC's and Mac's". The obvious question is can I hack it to play from Linux? If so, they just got a new convert.
You're used to 'home storage' prices. Look at pricewatch, find a good brand of EIDE, and just get it.
They're looking at 'enterprise storage'. We have 11 tera of raw disk on an EMC. It cost $2 million. The useable storage out of it is around 3-4 tera, after counting mirroring, and third mirror break off for backups, etc, etc, etc.
These drives use MCA (iirc) interconnects to a disk backplane, and fiber channel interconnects between disk boards and the front end san switch. The computers are fiber connected into the san switch as well, and the JNI cards (client end of a SAN connection) for this are NOT cheep.
To Online storage companies, downtime costs serious money. They can't afford the downtime. That's why their storage costs real money. Then they pass it on to you.
If you need real amounts of data, you don't want a hosting service, you want a CoLo service (They give you rack space, and an internet connection. You provide the box). If you want, you can put a desktop with 2x140 gb drives, and you'll get what reliability you can out of it (most IDE drives are warrenteed for 1 year for a reason). If you want the thing to last, get a server class, rack mountable server from (dell|compaq|ibm|penguin computers). You'll be happy you did. Mirror the drives (preferably in hardware) so you can loose a disk without killing your service.
And it might already be there. Please let me know.
I want ARBITRARY keyboard shortcuts. I want to be able to write a shell script (or any executable), and have it execute when I hit (ctrl)-(alt)-w (My keystroke to bring up a vertically maximized terminal window).
I was quite scared with gnome 2.x when they seemed to take this feature away, but I found out how to do it eventually (gconf-editor under the metacity stuff).
/etc/init.d/sendmail stop
And sendmail will stop. No remembering obscure syntax. No weird rules. All of it (should) be encoded in that script.
The problem with traditional BSD init files is that it is deeply challenging to add/remove the code to start/stop a package because it has to insert code into a single script. If there's an out-er 'if' statement that the code doesn't expect, the program might not start. Well, it's somewhat easy to add. It's almost imposible to cleanly remove, especially if the admin has modified the script in any way. With SysV, you just drop the script in place, and add some symlinks. When you remove, you delete these files.
Granted, I havn't played with a recent *BSD to see how the init scripts have evolved. So, note the 'traditional' above.
I love Gentoo, but you're not going to get mozilla and evolution installed this evening.
Just getting gnome going took from 10:00 PM to 8:30 AM on my 850 MHz laptop with 256 MB. If you do a stage3 build, you'll have CLI done by the end of an evening. Either gnome or KDE will take the time while you sleep and do work the following day. Mozilla will take 2ish hours, and evolution will take 4-5ish (IIRC. I havn't done that yet on this hardware).
My suggestion: Get computer up to CLI, and Friday after work run:
emerge kde gnome gdm mozilla evolution
It probably will be done by Sunday.
Reguarding performance, there is a difference. Anything CPU intensive (and especially math intensive) is going to get a nice boost. LAME went from 2x encode to 4-5x encode by recompiling after playing with compiler settings (from defaults to using -march, -mcpu, -pipe, -fomit-frame-pointer, and a few others)
And reguarding 'whatever crappy defaults', that's also a yes and a no. I for one hate the fact that MTR will link to GTK if it finds it. It should forever be a command line app. So I can easily do this:
USE="-gnome -gtk" emerge mtr
Now, as for configuration defaults, gentoo tends to take whatever the source code comes with (for example, no bluecurve).
Unfortunatly, the 'I cost them more money in support' argument doesn't wash. Accountants split money into different 'buckets'. Revenue comes in, and is devided into capital, Operations and Maintance, and Profit. (I'm over simplifying).
Capital money is money given to projects once. It's usually a bigish wad, and it's for a year. You have this much money to get this many new things to do this good work for us (IE: We need to implement a new source code version tracking package. Needs this much hardware, this much software, and will cost this much in support the first year).
O&M money is 'mantaiance' money. It costs this much to keep this web server up and running, and supplied with bandwidth. Salaries come out of this pool as well. O&M is usually a sunk cost. Just the cost of doing business. Support for their product is certainly a sunk cost. They invest a flat fee in salaries, headcount, phones, etc. From my CVS example above, maintance on the hardware, and software would be added to the O&M budget for the second year of it's life.
To the accountants, someone being on hold for 6 hours, and wasting 1 hour of 4 employees time isn't 'wasting' more money than the revenue from purchasing the product. It's just averaged in with all those people who DIDN'T call intuit, and just used TurboTax on one computer (like me (i bought it before I heard of the DRM games)).
Both of you are kind of right (Note: I've played more MK64 than I should have).
;-)
Bowser, Wario, and DK have the fastest top end. The problem is that they take a long time to get up there.
Toad, Peach, and Yoshi have a slower top end, but they get up to it much much faster.
Discounting powerslides for the moment, the heavy guys also lost a decent bit of speed in the curves. This is where the light guys really shine. Powerslides equalize things somewhat. Usually in a race, the little guy will still come out on top, but it's a narrower victory.
Now, if we're talking battle mode, then I break out the Wario, and you're all fried.
What audiophiles realize is that equipment makes a huge difference in sound quality. What they don't realize is that people's audible memory (memory for the quality of the sound) is about as bad as our memory can be. If you go to an audio shop, they usually 'switches' that let you listen to different gear, instantly. You're listening to one receiver, then 'switch, you're music is coming through a different one, in about one second. Then you can hear the differences. However, if you drive 5 minutes to someone's house and listen to their rig, you're not going to be able to tell which sounded 'better'.
That said, stereo and home theater rigs are going to sound better than computer rigs, if for no other reason than better speakers. Computer speakers can't use large magnets, where sterio speakers can.
If you switch back and forth between CD source, and MP3 source for the same track, you'll be able to tell a difference (unless you encoded to 256 kbit). However, if you used a decent encoder, and a decent bit rate (128+) it will sound acceptable.
To give a final answer to the question above, the reason you listen to MP3's over your stereo is convienience. No changing CD's, no waiting on the 200 disk changer to rotate 180 degrees to find the disk you want, better UI for selecting disks/tracks, better randomizer functions, custom playlists, etc. It's fantastic for background noise in a party, and great when you're puttering around doing things.
If you're just relaxing, and focusing on the music, then you'll probably want the CD.
While Nintendo does have a broadband adaptor out, titles for it are a little thin on the ground. Traditionally, Nintindo has always focused on 'kart' style multiplayer. 4 people on a couch, trash talking, and having a great time. F-Zero, Mario party, and several others are examples of this.
I just hope the new Mario Kart can live up to expectations. The n64 version is still amazingly fun to play multiplayer.
The new Kart will be internet'able. I don't know if armada will be.
"It's hard to beat the bandwidth of a stationwagon filled with DAT tapes"
If you're talking: 1) several servers and 2) 200ish gigs per, welcome to needing a real backup solution.
One thing to keep in mind is the three 'kinds' of backups. You will need to cover (or choose not to) all three.
1) DR. Disaster recovery. A full image of ALL data, usually duplicated so you have a in house copy and a remote copy. Full system images, and a software package that can blast a full system image to a box, or full data and config backups that require an OS install before your restore. Usually this is somewhat light on tapes, since you'll only keep 2-3 weeks of them.
2) File Recovery. Someone deleted something that they shouldn't have and need it restored. Or the Database equivelent: "We dropped this table 5 weeks ago, and discovered just now this random important process that hits it every 2 months. Can we restore the DBF file so we can get that table, data and schema back?" Sometimes DR feeds into File Recovery. You just keep the tapes longer. More expensive in tapes though, and you have data you'll not use (like OS images) wasting tape space. It's easier though.
3) Archival. EG: The IRS mandates that we keep this data for N years (where N is usually greater than 7). Thankfully, this is a thin ammount of data, but it's important none the less. CD/DVD rock for this, but tapes are good too (so long as you're under 10 years. Media and reader issues will kill you after that).
Good luck. Backups are a huge pain. Be sure to test the DR portion of it at least once a year. You'll be thankful you do.
Why do people use oracle?
1) It's hyper expensive
2) the support organization is far from fantastic
3) It requires expensive talent to maintain
The answer to all of these issues is 'Name recognition'. Joe Average has at least heard of Oracle. CEO's and CIO's associate it with quality.
Currently, for 'high end' databases where 'high end' is defined by either CEO's or serious amounts of data, Oracle and DB2 (nee UDB or Universal Data Base) are wins.
To a corperation, nothing is more important than their data. If it can't be protected by a name they trust (whether or not that trust is misplaced) it will not penetrate the DataCenter.
Ok. I've been eyeing the sunfire line (cinema grand, and theater grand lines). From what I've HEARD, their pre-amp's arn't that good at the video switching, but are find for audio. The thing that really interests me though is their 'step down transformer' amps. How do they sound to you? The lack of heat, and the nice power ratings are good pros, but it's a rather un-orthodox way of amping a signal (and in this day of surround sound, I'm not sure if the assumptions will hold up).
:-)
I know that I really should just go and listen to them (and I will if I ever go to buy one), but I'm lazy, so I'm asking you.
Never. Written languages are too vague:
*I* heard her say you were an idiot
i *HEARD* her say you were an idiot
i heard *HER* say you were an idiot
i heard her say you *WERE* an idiot
Four rather different meanings depending on the inflection. All languages are tonal, to a certain degree (even esperanto), and programing languages (being written) need to live with out the tonality. Therefore, they'll always be different.
The Native Posix Threading Lib (IIRC) is something Ingo cooked up in addition to the O(1) scheduler, and a few other goodies. Previously, posix threads could only have a couple hundred threads going concurently on ANY hardware. It just couldn't scale past that.
With NPTL, Ingo on a dual proc box (granted, a nice one) was able to get 16,000 concurent threads going, and the IO system wasn't suffocated, the CPU's wern't useless, and you could still browse the web.
Granted, these threads wern't doing much, but they were alive, and switching in and out of context.
This means the foundation can scale to effectivly any size, and so long as the hardware can keep up, you'll be fine. You can now unleash your massivly multithreaded java apps (and what not. That's just the easiest example).
This doesn't help you if you need more than 4-8 CPU's on an intel platform, but it gets you a lot closer. If you want something that can parallelize that far, you really need something like Sun's e12k or e15k. IIRC, the DoD commissioned an e15k farm with a total of 4096 CPU's to model the first few nanoseconds of nuclear explosions. It had to be a single system image for various reasons, so don't go crying beowulf.
What prevents me from doing 'esc-! cat ~/.ssh/*', or something more nefarious?
xauth? (which is called by ssh, or can be set up manually)
The world has come someways since 'xhost +'.