Um, damn you're ignorant. I don't mean to be so rude, but dude.
Anyway, monitors are NOT the problem. They are dumb devices. As advanced as monitors get (aside from digital flat panels) is to respond with their capabilities when the video card asks. It's called DDC and has been implemented in most monitors made in the last five years. XFree86 doesn't directly use this information expect for the physical screen size because it would totally hose things up if the monitor was turned off, or not connected, when X queried it.
Apple sure has it easy. They only have to make Quartz run with their own monitors.
Again you've confused monitors with video hardware. Quartz doesn't care what monitor it's feeding. At most it cares about some of the basic capabilities of the monitor, like size, resolution, and maybe refresh rate.
- RustyTaco
Re:Xfree is sufferring from poor PR
on
Xfree86 4.2.0 Out
·
· Score: 1
That's why you can run XFree86 on SCO, SVR4, *BSD, Solaris and Linux.
Don't for get Darwin x86 AND PPC, Aqua, and one or two windows ports hanging around. I believe there is both a GDI And a DirectX port hanging around. I think the OS/2 port is mostly working still to.
- RustyTaco
Why can't the xfree86 guys just add anti-alising to the existing font rendering code?
Because there are no shades of grey in a black and white (bitmap) image. X11R6 core fonts are just bitmaps. Replacing them with greyscale images would be a protocol violation AND would be EVERY app.
well, running 2.4.17 means it hasn't yet really told about its stability (since it's so recent a kernel). i've been running 2.4.0-test2 for a quite long time now (in a server environment) and in spite of some problems, it's still going and hasn't actually yet crashed on me.
Hehe, me too, but on x86 not PPC. Up until about 2:30 hours ago I had 176 days on 2.4.6-ac5. Wouldn't have bothered with rebooting to 2.4.17 except I'd created a partition for the unused 13G of space on the drive and we needed to be able to use it.
Fact is, with the Riot you're getting a device with 4 times the storage space, plus an FM tuner, for the same price. Oh, and you get a carrying case, too!
Remind me again why a carrying case is a good thing? Seems to me that if something is big and fragile enough to need a carrying case it would be just a bit of a PITA to deal with. The FM tuner is nice, yes, but I'm just not seeing where having a carrying case for something that's supposed to be carried in a pocket is a good thing.
For 20GB and maximum USB throughput (1.5MBytes/sec) it would take approx. 3 hours 47 minutes and 33 seconds. Ouch.
I have to step in here and remind everybody that that is a very raw speed. USB has an extreamly heavy protocol which limits your actual transfer rates to 600-800kps, IF, and only IF, there is nothing else connected to the bus.
Yep, that's right, the mere existance of a mouse plugged into the same bus will slow down the transfer a little bit. Probably not much, but some.
- RustyTaco
Also, it's not hard to imagine Disney using some proprietary security solutions such as a RADIUS server for added security on top of WEP. RADIUS can be configured to change the WEP key every so many minutes.
It's called LEAP. And yes, it's a cisco-specific thing at the moment. They are working with the appropriate standards bodies(IEEE) to get something like it into the spec books. MS has some preliminary version implemented in XP, though I don't really think 802.1x has be ratified yet.
It does do exactly what you suggested though. It uses some sort of hash-based authentication so that your actual identifying information never sees the air(encrypted or not). It's backed by a RADIUS server and changes the per-client WEP key at a user defined interval.
It makes the likes of airsnort pretty much useless, if used right. Airsnort needs a fairly large amount of traffic to analize, but with a key lifetime of 5-10minutes it just isn't going to get the amount of traffic it needs before the key changes. And even if it does, the key still changes. Back to square one.
Don't just think music here: consider that, if paired with moderately decent voice recognition software back on the desktop computer, you could have close to instant transcription of speeches, lectures, meetings, etc.
Dragon Systems allready has something like this. A small digital voice recorder that you sync with Dragon Dictate on your PC. I think the package was going for $300-400, last I checked.
Are there any Firewire or USB Hard Drives that are of mobile rack format? I
would like to have hot swappable storage, but not with an external box. Mobile
rack would be nice.
Yes. Granite Digital has them. I just bought an external 5.25" ATA->firewire enclosure yesterday at Comdex. They did have 3.5" hotswapable bays with the same ATA100->1394 bridge chip.
Here's a link.
Seriously, most of your problems are with/.'s implementation, not the core idea. Require a strong identity check to create an account, there's your credibility.
As for the rest, do you see the little drop down box that says "Threshold: -1" and the one that says "Oldest First"? Select Treshold 3, Highest Scores first and click "Change". Now you don't have to deal with the riff-raff in your preciously limited available time.
Several members of the Tucson Free UNIX Group setup a Linux system at Corbell elementary school in Tucson, AZ. The kids (6+ years old) all had no problems using the system. As a matter of fact they looked forward to working in the computer lab.
Thank you for mentioning it. I'm guessing you are one of the ones who was involved earlier in the project.
These were debian boxes with KDE on the desktop.
It has been a while since you've been down there. We're now running icewm from a gdm login prompt. It's just cleaner and MUCH more usable on the old P90's
Special scripts were written to prevent user error (i.e. A perl script which would only allow one instance of Netscape to be spawned since Netscape has a very slow startup time.)
And that script has how been adapted to keep kiddies from starting multiple WINE sessions(used for a small handful of edutainment apps).
and NIS with NFS home directories were setup so that any user could log onto to any node and get their desktop/user data.
This is still in place, with a plot afoot to migrate from NIS to LDAP. For those who have never experiances a properly setup networok these are important points. You can't really begin to realize how lame and broken most windows apps are until you get used to ploping down infront of whatever computer is nearest and just working.
This was done about four years ago and with the DRAMATIC improvement of Linux on the desktop, I would think that this task would be even easier today.
The Linux side, hell yes. Either with a "magic" boot disk on the P90's or an F12 network boot on the new PIIIs we can drop right into a shared NFS root Debian install and have our way with the local systems. Great for bringing new harddrives up to speed(We've lost a lot of the new drives allready) or just syncing the system up.
The Dark(NT4) side of the new systems is quite the opposite. Nightmarish to even get functioning properly, especially when it came to the network printers.
All-in-all I think we've spent three times as much time fighting with the damn NT sides this school year than with the (default) Linux side. Most of the new systems(District/state, not our purchases) also came with 2k Pro, so we might give that a wirl, but I'm not expecting much more cooperation from it.
The administration benefits alone are worth using Linux versus Windows. You SSH-in and remotely fix any box at anytime. No PC-Anywhere licenses, no waiting until you come into the office tomorrow, it rocks!
Damn skippy. It's even better as the new systems are Wake-On-Lan capable. Plus, we have a total of 2.3 real installs to maintain. 56 of the new systems are identical, and are image appropriatly. The other 33 of the PIIIs are, um, well, s/imps2/ps2/ in gpm.conf is the extent of the changes that need to be done from the 733's image. Then there are the 32 P90's you remember, which are gutted 733 images so that it can be made to fit on the 540M harddrives.
Most of the working getting the PIIIs up to speed for this year was "apt-get dist-upgrade" and compiling a fresh kernel.
I also have setup all my systems at my house with Linux and all of my computer illiterate family members has never had a problem using the systems.
Sweet.
The key is that you as the sysadmin need to setup the environment for the user and make it stupid simple for them. Then you show them how to use it. A little scripting and creation of a desktop icon on your part will go a long way to helping your users have a good experience.
Damn skippy again. Thats what I do roughly one (3-10 hour) saturday every month. Mind you that is for the 65+33 PIII's, the 32 P90's, the 3 PPro servers, and then playing around a bit with the new(to us) UltraSparc server.
More detailed info here, until the district restarts their sick IP rearanging games again.
Here's a crazy idea: If you don't want to worry about frying your chip DON'T buy a frickin' spaceheater! Jeez, there isn't anything nessesarily hot(>30C) about circuitry. If done right you'd barely be able to tell it's on. Of course, Intel & AMD are more concerened with rediculously high clock speeds than power & heat efficient and durable chips. Flip chips don't help with higher external pinouts, or really effect overall manufacturing much. The sole reason they are used is to disapate the disgusting amount of waste heat generated by the raw(unfinished) core designs.
- RustyTaco
Happily having to turn off screen blanking to tell if his G3/500 laptop is on.
Why not both? Where better to comemorate the names of those lost than on the base of the new towers. Cover exterior walls of the first one or two floors of the new towers with those engraved steel plates you mentioned.
Um, I have a quick suggestion. Do a: ls -l/usr/bin some time. There are a hell of a lot more programs in there than the big 7 you mentioned. Try looking up the authors for all those non-monsterous projects and see if you can find ANY of them getting paid to work on them.
1. Any shop with over 5 identical machines should have Ghost or Drive image. You install the OS, apps, etc. Make an image via a network boot disk. Put boot disks in machines, boot to them, blow image on. Change SID, rename machine, reboot. Add to domain. Done. All the big cloning software packages support multicast as well. MS also provides some tools
Or, with Linux, rsync and a kernel disk with NFS root enabled, or just a nice initrd. Boot the disk, have it run tar, rsync, partimage, whatever to clone from a network-stored image. Reboot. Done, because you've configured the systems so that their only identifiying marks are the IP and hostname, which they conviently get from the DHCP server.
3. Terminal services are viable for NT/2k.
Kinda like X, just 15 years later.
4. Application installs - login scripts, as well as all kinds of software packages.
apt-get -y install "blah"
Or, just leave everything in the image if you have the disk space.
Or, NFS mount everything from a server (cluster).
Congratulations, you've just spent the same amount of time you would have spent tearing through shrink wrap, fiddling with CDs, and entering 20 digit CD keys.
Instead of having only one tree, there are two: the -stable and the -current.
You mean even and odd?(2.2 & 2.3, 2.4 & 2.5)
It's allready there. There isn't any real active development allowed on 2.4, period. It's all fixes and clean ups. Once 2.5 hits then we'll see nifty stuff like a new console subsystem, making everything happy with hotplugging, maybe a few new FS's being merged and other subsystem reorganisations.
If I go and tip someone, I usually don't expect that person to think it's anyghing else but a friendly guesture.
But what you do do is introduce an uncomfortable new element into the mix. Much like playing with a switchblade in public. You might say you're just playing, but the within striking distance.
That is what makes people nervous, you're changing the rules of the game.
- RustyTaco
Yes. I volunteer doing exactly that. They love it and it meant they got to use the P90s that were donated instead of the Mac LCs the district still hadn't replaced.
District finally did come through with Compaq PIII's, and those run Debian much better than the P90's:) They still have a dark side(NT4), but there wasn't anything we could do to avoid the NT4 & Office 2k Premium licences.
The problem i see with open source, especially Linux, is that no person thought of bringing the same to that world. No one has created their own CLR that everyone can add into, sharing code much more easily and helping each other more: which would be a truly open source ideal.
Um, open up an xterm the next time you're drooling at a pretty KDE or Gnome theme. You obviously have never learned your way around a simple (ba)sh prompt. Integration of multiple "components" is incredibly strait forward and effective. And you don't need $800 worth of MS devel software and 256M of ram to use it.
Instead, in ONLY 1 shocking example, there still isn't even a unified ODBC standard in Linux - totally unbelievable. Perl, Python, Lisp, on and on, each have to create their own interfaces to databases - tens of thousands of lines of code re-written over and over again to do the SAME thing.
First off, it's no where near that amount of code. Secondly: Perl is not C is not Python is not Lisp, on and on. Every language has it's own style and semantics. If the native database APIs don't fit well into that style and semantics then the API is broken. Every languages set of libraries makes the database's native API and fits it as well as posible into the style of the language. And, if you actually look at it you'll see most of those "different" database access APIs actually just bind the "official" C library into a native library and then provide the appropriate wrappers in the language itself.
One last note on this: They may be doing the same thing, but the are doing it differently, to complement the strengths of the native language.
Do you see, just from this 1 example, that with.NET Microsoft is going to be eating Open Source's lunch? OSS is wasting time re-writing, while MS builds a pluggable component architecture, letting programmers everywhere leverage each other's work, no matter what the language.
for i in `find -name *.mp3`; do
mpg123 -s $i | oggenc -r -B16 -C2 -R44100 - -o `sed s/.mp3/.ogg/`
done
As you were saying? I could throw a simple id3 tag translation in there too, but I don't feel like it.
I think that is why the writer was complaining there is no leadership in OSS. Why didn't someone think of this before for the OSS world? Why are you still programming in the dark ages, like in the ODBC example? And Mono isn't the answer, as Ximian won't be around long enough to make it happen, and it isn't innovative at all, it is just a copycat.
Like DBD::ODBC? Or any of the other DBI DBDs? The API is still DBI regardless of what the backend happens to be.
Mono is the answer, it was just a stupid question.
I think the internet is a bad medium for things like tv, video, phone etc. anyway, centralized broadcasts are much more efficient.
That right there is a major flaw in the article. The main complaint is that they can't broadcast their latested crap to everybody on the planet from the same MegaServer(tm). The problem is, they can't do it with any other media either. That's why we have local TV and radio transmitters, even if their just shoving the same "Survivor" crap through our heads.
All it would take is a network of repeaters ala Akami and they'd have at least as reliable a system as they have now, if not more.
Mountain(slow link) in the way? So the broadcast sucks for those people.
Fiber-seeking-backhoe take out a line? Happens all the time with any sort of land line. (My cable(TV&modem) was attacked by a backhoe just last week)
Environmental interference(Water, storms, wind, bad cables, trafic overload)? It happens
Person wasn't available to watch it immediatly? Gotcha there.
Rather get it perfect than realtime? No way with conventional delivery systems. So you may have to wait till a few minutes after the live cast finished to have it all. But then you can watch/listed/experiance it in it's entirety without worrying about any of the above problems.
With multicast aware routers(not even worrying about QOS) the digital stream flows over the IP part of the (cable) network as smoothly as an NTSC channel.
They can't acheive that level of quality and reliability with the current intrastructures. Satalite outages, storm interference, hell, even digitial decoding artifacts in the analog channels
Of course, that's all for the massive webcast senario. Just respecting the TOS field and LARTing idiots setting thier pr0n transfers to "realtime" would take care of most of the latency for true realtime stuff.
Slashdot Mirror
Anyway, monitors are NOT the problem. They are dumb devices. As advanced as monitors get (aside from digital flat panels) is to respond with their capabilities when the video card asks. It's called DDC and has been implemented in most monitors made in the last five years. XFree86 doesn't directly use this information expect for the physical screen size because it would totally hose things up if the monitor was turned off, or not connected, when X queried it.
Again you've confused monitors with video hardware. Quartz doesn't care what monitor it's feeding. At most it cares about some of the basic capabilities of the monitor, like size, resolution, and maybe refresh rate.
- RustyTaco
Don't for get Darwin x86 AND PPC, Aqua, and one or two windows ports hanging around. I believe there is both a GDI And a DirectX port hanging around. I think the OS/2 port is mostly working still to.
- RustyTaco
- RustyTaco
- RustyTaco
- RustyTaco
Yep, that's right, the mere existance of a mouse plugged into the same bus will slow down the transfer a little bit. Probably not much, but some.
- RustyTaco
It does do exactly what you suggested though. It uses some sort of hash-based authentication so that your actual identifying information never sees the air(encrypted or not). It's backed by a RADIUS server and changes the per-client WEP key at a user defined interval.
It makes the likes of airsnort pretty much useless, if used right. Airsnort needs a fairly large amount of traffic to analize, but with a key lifetime of 5-10minutes it just isn't going to get the amount of traffic it needs before the key changes. And even if it does, the key still changes. Back to square one.
- RustyTaco
- RustyTaco
- RustyTaco
Proud anti-x86 bigot since he got his iBook.
Here's a link.
- RustyTaco
Kinda like a disk cache? Only less reliable
- RustyTaco
Taco boy doesn't let you moderate yet does he?
/.'s implementation, not the core idea. Require a strong identity check to create an account, there's your credibility.
Seriously, most of your problems are with
As for the rest, do you see the little drop down box that says "Threshold: -1" and the one that says "Oldest First"? Select Treshold 3, Highest Scores first and click "Change". Now you don't have to deal with the riff-raff in your preciously limited available time.
- RustyTaco
It has been a while since you've been down there. We're now running icewm from a gdm login prompt. It's just cleaner and MUCH more usable on the old P90's
And that script has how been adapted to keep kiddies from starting multiple WINE sessions(used for a small handful of edutainment apps). This is still in place, with a plot afoot to migrate from NIS to LDAP. For those who have never experiances a properly setup networok these are important points. You can't really begin to realize how lame and broken most windows apps are until you get used to ploping down infront of whatever computer is nearest and just working.
The Linux side, hell yes. Either with a "magic" boot disk on the P90's or an F12 network boot on the new PIIIs we can drop right into a shared NFS root Debian install and have our way with the local systems. Great for bringing new harddrives up to speed(We've lost a lot of the new drives allready) or just syncing the system up.
The Dark(NT4) side of the new systems is quite the opposite. Nightmarish to even get functioning properly, especially when it came to the network printers.
All-in-all I think we've spent three times as much time fighting with the damn NT sides this school year than with the (default) Linux side. Most of the new systems(District/state, not our purchases) also came with 2k Pro, so we might give that a wirl, but I'm not expecting much more cooperation from it.
Damn skippy. It's even better as the new systems are Wake-On-Lan capable. Plus, we have a total of 2.3 real installs to maintain. 56 of the new systems are identical, and are image appropriatly. The other 33 of the PIIIs are, um, well, s/imps2/ps2/ in gpm.conf is the extent of the changes that need to be done from the 733's image. Then there are the 32 P90's you remember, which are gutted 733 images so that it can be made to fit on the 540M harddrives.
Most of the working getting the PIIIs up to speed for this year was "apt-get dist-upgrade" and compiling a fresh kernel.
Sweet.
Damn skippy again. Thats what I do roughly one (3-10 hour) saturday every month. Mind you that is for the 65+33 PIII's, the 32 P90's, the 3 PPro servers, and then playing around a bit with the new(to us) UltraSparc server.
More detailed info here, until the district restarts their sick IP rearanging games again.
- RustyTaco
Here's a crazy idea: If you don't want to worry about frying your chip DON'T buy a frickin' spaceheater! Jeez, there isn't anything nessesarily hot(>30C) about circuitry. If done right you'd barely be able to tell it's on. Of course, Intel & AMD are more concerened with rediculously high clock speeds than power & heat efficient and durable chips. Flip chips don't help with higher external pinouts, or really effect overall manufacturing much. The sole reason they are used is to disapate the disgusting amount of waste heat generated by the raw(unfinished) core designs.
- RustyTaco
Happily having to turn off screen blanking to tell if his G3/500 laptop is on.
That is REALLY bad. Really, really bad... but I can't help but chuckle at it.
- RustyTaco
Why not both? Where better to comemorate the names of those lost than on the base of the new towers. Cover exterior walls of the first one or two floors of the new towers with those engraved steel plates you mentioned.
- RustyTaco
Um, I have a quick suggestion. Do a: ls -l /usr/bin some time. There are a hell of a lot more programs in there than the big 7 you mentioned. Try looking up the authors for all those non-monsterous projects and see if you can find ANY of them getting paid to work on them.
- RustyTaco
Kinda like X, just 15 years later.
apt-get -y install "blah"
Or, just leave everything in the image if you have the disk space.
Or, NFS mount everything from a server (cluster).
Congratulations, you've just spent the same amount of time you would have spent tearing through shrink wrap, fiddling with CDs, and entering 20 digit CD keys.
- RustyTaco
"if the bible proves the existance of god, then superman comics prove the existance of superman" - Usenet
Sorry, just had to share.
- RustyTaco
It's allready there. There isn't any real active development allowed on 2.4, period. It's all fixes and clean ups. Once 2.5 hits then we'll see nifty stuff like a new console subsystem, making everything happy with hotplugging, maybe a few new FS's being merged and other subsystem reorganisations.
- RustyTaco
That is what makes people nervous, you're changing the rules of the game.
- RustyTaco
Yes. I volunteer doing exactly that. They love it and it meant they got to use the P90s that were donated instead of the Mac LCs the district still hadn't replaced.
:) They still have a dark side(NT4), but there wasn't anything we could do to avoid the NT4 & Office 2k Premium licences.
District finally did come through with Compaq PIII's, and those run Debian much better than the P90's
- RustyTaco
First off, it's no where near that amount of code. Secondly: Perl is not C is not Python is not Lisp, on and on. Every language has it's own style and semantics. If the native database APIs don't fit well into that style and semantics then the API is broken. Every languages set of libraries makes the database's native API and fits it as well as posible into the style of the language. And, if you actually look at it you'll see most of those "different" database access APIs actually just bind the "official" C library into a native library and then provide the appropriate wrappers in the language itself.
One last note on this: They may be doing the same thing, but the are doing it differently, to complement the strengths of the native language.
for i in `find -name *.mp3`; do
mpg123 -s $i | oggenc -r -B16 -C2 -R44100 - -o `sed s/.mp3/.ogg/`
done
As you were saying? I could throw a simple id3 tag translation in there too, but I don't feel like it.
Like DBD::ODBC? Or any of the other DBI DBDs? The API is still DBI regardless of what the backend happens to be.
Mono is the answer, it was just a stupid question.
- RustyTaco
Um, who's this "us" you speak of? I'm not familiar with them. Nor am I familiar with "we". Could you please explain?
- RustyTaco
That right there is a major flaw in the article. The main complaint is that they can't broadcast their latested crap to everybody on the planet from the same MegaServer(tm). The problem is, they can't do it with any other media either. That's why we have local TV and radio transmitters, even if their just shoving the same "Survivor" crap through our heads.
All it would take is a network of repeaters ala Akami and they'd have at least as reliable a system as they have now, if not more.
Mountain(slow link) in the way? So the broadcast sucks for those people.
Fiber-seeking-backhoe take out a line? Happens all the time with any sort of land line. (My cable(TV&modem) was attacked by a backhoe just last week)
Environmental interference(Water, storms, wind, bad cables, trafic overload)? It happens
Person wasn't available to watch it immediatly? Gotcha there.
Rather get it perfect than realtime? No way with conventional delivery systems. So you may have to wait till a few minutes after the live cast finished to have it all. But then you can watch/listed/experiance it in it's entirety without worrying about any of the above problems.
With multicast aware routers(not even worrying about QOS) the digital stream flows over the IP part of the (cable) network as smoothly as an NTSC channel.
They can't acheive that level of quality and reliability with the current intrastructures. Satalite outages, storm interference, hell, even digitial decoding artifacts in the analog channels
Of course, that's all for the massive webcast senario. Just respecting the TOS field and LARTing idiots setting thier pr0n transfers to "realtime" would take care of most of the latency for true realtime stuff.
- RustyTaco