A guy I knew in college wrote a pretty good one for a compilers class or something like that. Basically all variable names became a number of underscores ("_" and "__", etc), all one line, eliminated as much whitespace as possible. Don't know if he also did things like typedefs to make it more confusing, etc. I'm sure there's more that he did too...
Very cool. It looks like they've done a lot of what my system does, plus most of what I've been planning for future enhancements.;-) Sounds like it's time for me to play around with that a bit and see if it meets my needs, and if not, contribute to the project. Mine's always been very specific to my setup and I've always wanted a more general implementation.
For remote control, have a look at Irman, which is supported under Linux via LIRC. Definitely the way to go.
As for Dolby Digital, I'll bet that some high end sound cards support it, but I have no links to back this up. Also, Linux support may very well not exist for these. (hmm, I'm assuming you would be running Linux but you don't say... however this is/. afterall);-)
For other features, networking is a must, to control it from elsewhere and perhaps to get program data like the Tivo does. I've always envisioned my house of the future having a central audio server somewhere with all my CDs as high quality MP3s. Then in each room would be some sort of networked terminal for song selection, and a pair of speakers connected to the server's sound card. Have as many sound cards as rooms where you potentially want to have something different playing at the same time. The same could be done with video. I suppose someday soon disk space will be cheap enough that we'll store terabytes of DVD video the same way we store gigs of MP3s now.
Then, it'll need intelligent software to control it. The software should take command-line input, gui input, networked input from remote clients, ir remote control input, and of course, voice input. Fast search/play capabilities are a must, but a menu-driven UI could be available for those who want it. My current system at home (not quite the audio server described above, but slowly getting there) has a custom perl script I wrote that plays songs, albums, etc based on regular expression searches. So rather than wading through a menu to get to Rock -> Van Halen -> 1984 -> Hot For Teacher, I can simply type "playmp3 ^Hot" and it will play everything that matches. Great convenience. Same thing for playing movies on DVD (hmm, do they make 200-disc DVD changers or anything like that?) and selecting TV programs to watch.
Just some random ideas, that's all I have to say...
Maybe the precompiled binaries are optimized for 686, but on my dual pentium-233 (Linux 2.2.16), Mozilla M17 is much slower than NS4 all the time. Not just in startup, but during the entire browsing/surfing session. I'm too lazy/apathetic to get the source and recompile, so who knows?
I'm downloading K-Meleon as I type to try on my NT laptop at work, and one of these days I'll compile and try Galeon to see if that's any faster at home.
I don't know what it's like in your area, but if you have any alternatives (Cable, other DSL providers, etc) I would look into them seriously, and let the company you're dealing with know. My girlfriend ordered PacBell DSL and after 6-7 weeks of phone calls and no install date, she gave up and got Cable. That was installed in a week, everything went flawlessly.
One thing that a lot of people don't seem to know about DSL is that your provider doesn't necessarily have to be your phone company. I chose a relatively small provider in my area that got lots of great reviews. They have to deal with the telco regarding line issues, not me. And they've been great with service, much better than the telco itself. But I don't know if such providers exist in your area or if the local laws allow the telco to have a monopoly on DSL service...
Check out DSL Reports for more info on your area, and post a bad review of the company if you're sick of 'em!
True, mostly, but not always. Some high end compilers can do certain optimizations that will allow certain types of algorithms to run faster on a multiprocessor machine. For example, consider this loop:
for (i = 0; i < n; i++)
a[i] = b[i] + 5;
Successive iterations of this loop aren't dependent on previous iterations, so in theory, if you had n processors, you could do all n iterations in parallel at once. But of course, most algorithms aren't this parallelizable. For example:
for (i = 1; i < n; i++)
a[i] = a[i-1] + 5;
And of course this is the answer to the poster's question. While some problems may be highly parallelizable, others aren't - each step of the algorithm may depend on the results of the previous step. In this case, throwing more processors at it does you zero good. It's like if you're driving across the country and you decide to take 100 of your friends, all in their own cars, to try to get there faster. It's still going to take the same amount of time.
But in the past few weeks, Andrew
Morton has written a clean and successful low-latency patch that is likely to be accepted. During
testing, it has been found to be stable and get sub-2 msec latencies. We're already beating BeOS,
and we've been beating Mac OS and Windows for quite some time now.
You know, right after I hit submit, I realized I forgot to add the disclaimer of "please correct me if I'm wrong, etc, etc...";-) Thanks for the updated information, this is certainly exciting news. Has this stuff been tested under relatively high loads? With Linux 2.2, I can still fairly easily interrupt simple audio output when a couple processes decide to eat CPU or thrash the disk for a bit. Not that you would run critical realtime stuff on a highly loaded machine, as you point out below.
As for cron, it's a DDT (Don't Do That!). There's no need to run cron on a DAW, so disable it.
True, I'd certainly want to kill off (or not start to begin with) any non-essential processes on a DAW. (I wonder if MacOS X will have anything cron-like running by default that could interfere with realtime stuff) I guess my point was with preemptive multitasking, the user has to do some work to be sure nothing is running in the background that might wake up and disrupt realtime processing. Of course, maybe someone could come up with a Linux-DA distribution or something with all background services, etc turned off. But the one thing I still love about cooperative multitasking in this situation is that I don't have to worry if I forgot to kill a couple background processes. They just won't get any CPU until I'm done. It's just the peace of mind of knowing that nothing will interrupt me, without having to doublecheck.;-)
They're sick and tired of the high latencies
(even with DirectSound) under Windows;
This is precisely why I don't think Linux will be a viable platform for serious multimedia development for some time to come. I love Linux, I've been running it on my main PCs for 7 years now. But realtime multimedia development demands realtime response, which is difficult if not nearly impossible to achieve, guaranteed, in a pre-emptively multitasking OS. I've never used BeOS but had heard good things about it in this regard, and SGI has some nice guaranteed-rate I/O stuff and realtime process priorities in IRIX. But Linux at this point isn't that close.
What happens when I've been working on a song all night, it's 4 in the morning and I'm in the middle of recording that once-in-a-lifetime perfect take, and all of a sudden the nightly updatedb cron job kicks in, thrashing the disk and losing sync? This is absolutely not acceptable. The beauty of cooperative multitasking in the current MacOS (and the reason I own a Mac) is that the sound/video/multimedia application can take over the entire machine if it wants, allowing no other process to screw it up. Sure, in everyday use I definitely want pre-emptive multitasking, but this is one instance where cooperative can be a good thing. When I'm laying down a track or capturing video or whatever, I don't give a crap if some background task wants to run. It isn't gonna get 1 CPU cycle til I'm done capturing my data. And that's the way I want it.
That said, I'm insanely curious how Apple plans to address this issue in MacOS X. I think if multimedia producers upgrade and then start getting bitten by occasional strange latencies due to Darwin's background processes, it may alienate a lot of them. I know I read somewhere that the QuickTime developers at Apple are pissed that they can't get full control of the CPU. This suggests that OS X won't have guaranteed realtime scheduling, so I wonder if the multimedia market will force Apple to maintain Classic MacOS for awhile to come...
Whoops, looks like I started to stray a little offtopic towards the end, oh well...
Interesting. This is beginning to stray a bit offtopic, but what the heck.
I had a very similar thing happen to my Mac over the weekend. MacOS began hanging and wouldn't complete the boot process, yet LinuxPPC still ran like a champ. It perplexed the hell out of me since even the MacOS install CD wouldn't complete booting, so it wasn't that hard drive. Turns out I fixed it when I pulled all the memory chips and put them back in, but I wonder if something similar was going on? Probably not, but interesting anyway...
I agree that it can definitely be time-consuming, which is why I ultimately abandoned mine. I made one for myself and my roomie at the time back in '96. It probably didn't take more than about a day to grab the source for everything I needed (I based mine on Slackware and grabbed a lot of their source packages, getting newer versions if I could find them) and put it all together into a working system. It was a fun little project and there was definitely some satisfaction in knowing that every binary on that system was compiled on that system.
But eventually packages progressed, getting better, newer features, and it became a royal pain to try to keep up. I finally gave up after spending way too much time sorting out dependencies and tracking down old files from older versions of packages.
By '98 or so I had switched to RedHat due to rpm and have been reasonably happy with that, applying a lot of my own customizations. Even now, though, every time I upgrade I like to do a clean install and then apply my custom changes - it turns out to be quite a project for a couple hours. One reason why I still haven't gone up to 6.2. Plus I'd like to try Debian and maybe a BSD or two one of these days... If only I had the time.;-)
Bottom line, it's a lot of fun but be prepared to spend some time on it.
Yep, here's my review of MM Internet. I connect to the office using VPN from my work laptop, which is on my internal home network and masqueraded out to the DSL via my Linux box. All perfectly legal by MM's terms of service. Take that, @Home!;-)
I ran into this problem with my old dialup and practically pulled my hair out trying to get sendmail to use my dynamic dialup hostname in its greeting. I finally gave up and went with qmail. You can configure its "default host" on the fly simply by putting it in/var/spool/qmail/control/defaulthost. (all qmail configuration is done via text files in that directory).
I hacked up a quick script to interrogate my dialup hostname via nslookup on the IP and put that in the above qmail config. Then I added that script to/etc/ppp/ip-up. I also added a line to/etc/ppp/ip-down to put my internal hostname back for when I was offline. A hack for sure, but it worked.
Now, if your provider doesn't have reverse DNS setup for its dynamic IPs, the only other thing I can think of is do what I do now. I own bikegods.org (hosted elsewhere), so I just set my qmail defaulthost to that for all time. If a mailer on the other end wants to contact postmaster, etc, it will still get back to me. Do something similar if you own a domain (or know someone who doesn't mind you doing that).
Sounds like most of the Vegas casinos use NT for their billboards. I got a nice chuckle back in '98 when I saw the BSOD on the billboard at the MGM Grand. Too bad I didn't have a camera on me at the time...
The days when you had to worry about what file format is used to store your mail are over. IMAP is designed to allow you to access your mail folders anywhere, anytime, from any program.
This all sounds nice, but what about when I'm not connected to my central server and I need information that's stored in my mail messages? This is precisely why my work laptop is configured to store mail locally rather than IMAP, which is the company-wide default. I do a fair amount of traveling to customer sites, and when I'm in a meeting and I need to pull up an email, I can't get it from IMAP, so it has to be in a local file.
I don't suppose IMAP supports synchronization of local and remote copies of the mailbox so that I can have all of its benefits when I'm connected, yet still have access to messages when I'm in the middle of nowhere? Perhaps this can be done by the client, but I'm not sure if any of them have such a feature. Anyone know?
A former roommate of mine used to do something similar:
Telemarketer: [sales spiel] Him: I'm interested, let me go find my credit card... Him: [checking back every 5 or 10 minutes] I almost found my credit card, hold on, just another minute... Repeat until the poor sucker finally gives up.
They need to start selling is a blocked anonymous call rejection override override.
I've got this already, essentially. I have a voice modem in my linux box, so I hacked up a perl script to make it my answering machine. Whenever callerid data comes up as Unknown or Private, the computer answers immediately (usually just after the first ring, sometimes before the phone itself actually rings). Works like a charm. Telemarketers never bug me now, and on the rare occasion that someone I know has callerid blocking, they can either leave a message or call back using *82.
I'd love to. I live about 15 miles from work (LA area) and the ride would be perfect to keep me in shape for my weekend biking. But there's no safe route with bike lanes or clear sidewalks 100% of the way, and there's no way in hell I'll trust my life to the maniacs on the road. Just another reason why LA is so car-centric...
You got that wrong. You can't win is the first law of thermogoddamnics, not thermodynamics. You can't break even is the second law of thermogoddamnics.
Encode using LAME at 256 kbits. You can still fit 6-7 MP3 albums on 1 CD. At that bitrate I dare you to notice a difference between the original CD and the MP3 version, they sound absolutely identical to me.
Agreed. Actually if you use LAME's VBR (variable bitrate) capabilities, you get the best of both worlds. Sections that need it are encoded at a higher bitrate for quality and those that don't are encoded at a lower rate to save space. I've done this on my entire CD collection and usually end up with files that are about 30-40% larger than straight 128kbit/s mp3s.
Like you say though, I dare anyone to take a "blindfolded" test between this and the original and consistently identify both. I can definitely hear the difference in 128kbit/s files and it bugs me enough that I can't listen to most mp3s available on the web. I really hope VBR or at least a higher rate becomes more "standard" before major artists start putting their stuff out on mp3. It'd be a real shame if 128kbit/s is the best you can get when that happens.
Actually what I hate more than this "lock-in" stuff (minor annoyance, easily worked around) are sites that have continually refreshing ads in a separate frame. First, the "hold down Back button and select page from history" solution won't work (the first location in the list will get you completely out of the frameset, not necessarily what you wanted). And because of this, you're stuck hitting Back 50 times while the ad frame cycles back before the "real" frame goes back to the previous page. You may be able to work around this by opening the real frame in a new window, but not if the braindead web designers give you no navigation tags in the real frame, forcing you to use a third, navigation frame. It's just poor website design and it pisses me off sometimes.;-)
So many times at sites like this I've wished the Stop button would also stop the timer for meta refresh tags (and I think it should - everything should stop: loading, animations, refresh timers, javascript timers, etc). This small feature would be a godsend for such situations.
Re:Lynx: I don't get the "best experience possible
on
Web Site "Lock-In"
·
· Score: 1
Actually I got the same message with NS 4.72 on IRIX with JavaScript on (Java off). Seems Home Depot's web designers don't know what they're doing, eh?
Wow, if I had moderator points today, I'd bump this comment up another notch. I just tried DSL Reports and found lots of great information. After checking with several ISPs, I was under the impression that I was too far from the CO for DSL. It turns out they were all using erroneous information from Covad and my CO is 262 feet away (which I was kinda suspecting...). Time to look into DSL again!
I think the problem with the "Ask questions later" approach is that the military is liable to shoot down what effectively looks like a big missile hurling down toward the US... Sort of a crappy ending to a great vacation:)
Perhaps, but current algorithms for analyzing rocket plumes (as seen from military satellites) are certainly advanced enough to know the difference between a real missile threat and some guy's backyard spacecraft's retro rockets. I could tell you more, but then I'd have to kill you.;-)
Slashdot Mirror
Too bad the language he's working on now can't be used in the contest. It's a pretty nifty piece of work.
Thanks for the link!
As for Dolby Digital, I'll bet that some high end sound cards support it, but I have no links to back this up. Also, Linux support may very well not exist for these. (hmm, I'm assuming you would be running Linux but you don't say... however this is /. afterall) ;-)
For other features, networking is a must, to control it from elsewhere and perhaps to get program data like the Tivo does. I've always envisioned my house of the future having a central audio server somewhere with all my CDs as high quality MP3s. Then in each room would be some sort of networked terminal for song selection, and a pair of speakers connected to the server's sound card. Have as many sound cards as rooms where you potentially want to have something different playing at the same time. The same could be done with video. I suppose someday soon disk space will be cheap enough that we'll store terabytes of DVD video the same way we store gigs of MP3s now.
Then, it'll need intelligent software to control it. The software should take command-line input, gui input, networked input from remote clients, ir remote control input, and of course, voice input. Fast search/play capabilities are a must, but a menu-driven UI could be available for those who want it. My current system at home (not quite the audio server described above, but slowly getting there) has a custom perl script I wrote that plays songs, albums, etc based on regular expression searches. So rather than wading through a menu to get to Rock -> Van Halen -> 1984 -> Hot For Teacher, I can simply type "playmp3 ^Hot" and it will play everything that matches. Great convenience. Same thing for playing movies on DVD (hmm, do they make 200-disc DVD changers or anything like that?) and selecting TV programs to watch.
Just some random ideas, that's all I have to say...
I'm downloading K-Meleon as I type to try on my NT laptop at work, and one of these days I'll compile and try Galeon to see if that's any faster at home.
One thing that a lot of people don't seem to know about DSL is that your provider doesn't necessarily have to be your phone company. I chose a relatively small provider in my area that got lots of great reviews. They have to deal with the telco regarding line issues, not me. And they've been great with service, much better than the telco itself. But I don't know if such providers exist in your area or if the local laws allow the telco to have a monopoly on DSL service...
Check out DSL Reports for more info on your area, and post a bad review of the company if you're sick of 'em!
True, mostly, but not always. Some high end compilers can do certain optimizations that will allow certain types of algorithms to run faster on a multiprocessor machine. For example, consider this loop:
Successive iterations of this loop aren't dependent on previous iterations, so in theory, if you had n processors, you could do all n iterations in parallel at once. But of course, most algorithms aren't this parallelizable. For example:
And of course this is the answer to the poster's question. While some problems may be highly parallelizable, others aren't - each step of the algorithm may depend on the results of the previous step. In this case, throwing more processors at it does you zero good. It's like if you're driving across the country and you decide to take 100 of your friends, all in their own cars, to try to get there faster. It's still going to take the same amount of time.
You know, right after I hit submit, I realized I forgot to add the disclaimer of "please correct me if I'm wrong, etc, etc..." ;-) Thanks for the updated information, this is certainly exciting news. Has this stuff been tested under relatively high loads? With Linux 2.2, I can still fairly easily interrupt simple audio output when a couple processes decide to eat CPU or thrash the disk for a bit. Not that you would run critical realtime stuff on a highly loaded machine, as you point out below.
As for cron, it's a DDT (Don't Do That!). There's no need to run cron on a DAW, so disable it.
True, I'd certainly want to kill off (or not start to begin with) any non-essential processes on a DAW. (I wonder if MacOS X will have anything cron-like running by default that could interfere with realtime stuff) I guess my point was with preemptive multitasking, the user has to do some work to be sure nothing is running in the background that might wake up and disrupt realtime processing. Of course, maybe someone could come up with a Linux-DA distribution or something with all background services, etc turned off. But the one thing I still love about cooperative multitasking in this situation is that I don't have to worry if I forgot to kill a couple background processes. They just won't get any CPU until I'm done. It's just the peace of mind of knowing that nothing will interrupt me, without having to doublecheck. ;-)
This is precisely why I don't think Linux will be a viable platform for serious multimedia development for some time to come. I love Linux, I've been running it on my main PCs for 7 years now. But realtime multimedia development demands realtime response, which is difficult if not nearly impossible to achieve, guaranteed, in a pre-emptively multitasking OS. I've never used BeOS but had heard good things about it in this regard, and SGI has some nice guaranteed-rate I/O stuff and realtime process priorities in IRIX. But Linux at this point isn't that close.
What happens when I've been working on a song all night, it's 4 in the morning and I'm in the middle of recording that once-in-a-lifetime perfect take, and all of a sudden the nightly updatedb cron job kicks in, thrashing the disk and losing sync? This is absolutely not acceptable. The beauty of cooperative multitasking in the current MacOS (and the reason I own a Mac) is that the sound/video/multimedia application can take over the entire machine if it wants, allowing no other process to screw it up. Sure, in everyday use I definitely want pre-emptive multitasking, but this is one instance where cooperative can be a good thing. When I'm laying down a track or capturing video or whatever, I don't give a crap if some background task wants to run. It isn't gonna get 1 CPU cycle til I'm done capturing my data. And that's the way I want it.
That said, I'm insanely curious how Apple plans to address this issue in MacOS X. I think if multimedia producers upgrade and then start getting bitten by occasional strange latencies due to Darwin's background processes, it may alienate a lot of them. I know I read somewhere that the QuickTime developers at Apple are pissed that they can't get full control of the CPU. This suggests that OS X won't have guaranteed realtime scheduling, so I wonder if the multimedia market will force Apple to maintain Classic MacOS for awhile to come...
Whoops, looks like I started to stray a little offtopic towards the end, oh well...
It was a fake. Check out the FAQs and shot by shot description. This guy did an amazing job.
I had a very similar thing happen to my Mac over the weekend. MacOS began hanging and wouldn't complete the boot process, yet LinuxPPC still ran like a champ. It perplexed the hell out of me since even the MacOS install CD wouldn't complete booting, so it wasn't that hard drive. Turns out I fixed it when I pulled all the memory chips and put them back in, but I wonder if something similar was going on? Probably not, but interesting anyway...
But eventually packages progressed, getting better, newer features, and it became a royal pain to try to keep up. I finally gave up after spending way too much time sorting out dependencies and tracking down old files from older versions of packages.
By '98 or so I had switched to RedHat due to rpm and have been reasonably happy with that, applying a lot of my own customizations. Even now, though, every time I upgrade I like to do a clean install and then apply my custom changes - it turns out to be quite a project for a couple hours. One reason why I still haven't gone up to 6.2. Plus I'd like to try Debian and maybe a BSD or two one of these days... If only I had the time. ;-)
Bottom line, it's a lot of fun but be prepared to spend some time on it.
All DSL/cable providers should be this good...
I hacked up a quick script to interrogate my dialup hostname via nslookup on the IP and put that in the above qmail config. Then I added that script to /etc/ppp/ip-up. I also added a line to /etc/ppp/ip-down to put my internal hostname back for when I was offline. A hack for sure, but it worked.
Now, if your provider doesn't have reverse DNS setup for its dynamic IPs, the only other thing I can think of is do what I do now. I own bikegods.org (hosted elsewhere), so I just set my qmail defaulthost to that for all time. If a mailer on the other end wants to contact postmaster, etc, it will still get back to me. Do something similar if you own a domain (or know someone who doesn't mind you doing that).
Sounds like most of the Vegas casinos use NT for their billboards. I got a nice chuckle back in '98 when I saw the BSOD on the billboard at the MGM Grand. Too bad I didn't have a camera on me at the time...
This all sounds nice, but what about when I'm not connected to my central server and I need information that's stored in my mail messages? This is precisely why my work laptop is configured to store mail locally rather than IMAP, which is the company-wide default. I do a fair amount of traveling to customer sites, and when I'm in a meeting and I need to pull up an email, I can't get it from IMAP, so it has to be in a local file.
I don't suppose IMAP supports synchronization of local and remote copies of the mailbox so that I can have all of its benefits when I'm connected, yet still have access to messages when I'm in the middle of nowhere? Perhaps this can be done by the client, but I'm not sure if any of them have such a feature. Anyone know?
Telemarketer: [sales spiel]
Him: I'm interested, let me go find my credit card...
Him: [checking back every 5 or 10 minutes] I almost found my credit card, hold on, just another minute...
Repeat until the poor sucker finally gives up.
Been there, done that. ;-)
I've got this already, essentially. I have a voice modem in my linux box, so I hacked up a perl script to make it my answering machine. Whenever callerid data comes up as Unknown or Private, the computer answers immediately (usually just after the first ring, sometimes before the phone itself actually rings). Works like a charm. Telemarketers never bug me now, and on the rare occasion that someone I know has callerid blocking, they can either leave a message or call back using *82.
I'd love to. I live about 15 miles from work (LA area) and the ride would be perfect to keep me in shape for my weekend biking. But there's no safe route with bike lanes or clear sidewalks 100% of the way, and there's no way in hell I'll trust my life to the maniacs on the road. Just another reason why LA is so car-centric...
Shouldn't that be "thermogoddamnits?" ;-)
Agreed. Actually if you use LAME's VBR (variable bitrate) capabilities, you get the best of both worlds. Sections that need it are encoded at a higher bitrate for quality and those that don't are encoded at a lower rate to save space. I've done this on my entire CD collection and usually end up with files that are about 30-40% larger than straight 128kbit/s mp3s.
Like you say though, I dare anyone to take a "blindfolded" test between this and the original and consistently identify both. I can definitely hear the difference in 128kbit/s files and it bugs me enough that I can't listen to most mp3s available on the web. I really hope VBR or at least a higher rate becomes more "standard" before major artists start putting their stuff out on mp3. It'd be a real shame if 128kbit/s is the best you can get when that happens.
So many times at sites like this I've wished the Stop button would also stop the timer for meta refresh tags (and I think it should - everything should stop: loading, animations, refresh timers, javascript timers, etc). This small feature would be a godsend for such situations.
Actually I got the same message with NS 4.72 on IRIX with JavaScript on (Java off). Seems Home Depot's web designers don't know what they're doing, eh?
Wow, if I had moderator points today, I'd bump this comment up another notch. I just tried DSL Reports and found lots of great information. After checking with several ISPs, I was under the impression that I was too far from the CO for DSL. It turns out they were all using erroneous information from Covad and my CO is 262 feet away (which I was kinda suspecting...). Time to look into DSL again!
Perhaps, but current algorithms for analyzing rocket plumes (as seen from military satellites) are certainly advanced enough to know the difference between a real missile threat and some guy's backyard spacecraft's retro rockets. I could tell you more, but then I'd have to kill you. ;-)