I tried mp3 and wma at 96kbs, and they sounded pretty much the same to me, (pretty much like crap) and the wma was bigger (by only about 20K).
Amazing! In the other news there was an user that got pretty much equal speed with 56kbps modem made by Dynalink and with 56kpbs modem made by Genius...
Why do you think that two file formats with the same bitrate should create different file sizes? (Not counting different header size etc.) When comparing compressed file formats one should compare sizes of compressed files that decompress to similar results. What is similar result is totally subjective for lossy formats, of course. For example, if you think that 128kpbs MP3 sounds like 196kpbs WMA then it's MP3 that you should use. _________________________
Resources are used. I have to request your ad (time), download your ad (bandwidth), store your ad in my cache (storage). TV inflicts no such overhead.
TV vs. web isn't that simple. TV ad stoles my time (have to wait it end no matter if I'm interested in it or not), it takes my bandwidth (how do you think that video stream comes to your tv?). Granted that TV doesn't store stream, but you can use VCR(/TiVo or like) or disable browser cache to make things equal.
In fact TV ads are much worse than web ads because you can always close those pop up windows as soon as they appear or click stop button to stop animations (doesn't work for flash ads though). There's much software to filter incoming web pages to contain no ads. etc etc etc.
I think the difference is that because web is interactive media you expect it to do only what you've asked for; for example you click link, you expect browser to change page instead of opening ten new windows and playing some songs. Watching tv and seeing ending titles coming you immediatly expect to see some ads next so you change channel or close tv. In fact choise to not to see ads when using tv is always opt out (per ad!) instead of opt in (global!) in web (turn on javascript, java and images). _________________________
I'm using IceWM without crashes too. I don't like Konqueror because it wants to load all that KDE stuff. I use Galeon instead and even though it crashes sometimes it has great crash recovery - it restores all pages/windows/tabs you had open. With Galeon 0.11.0 with nightly build mozilla libs [1] I get the same rendering as mozilla with the UI of Galeon:) The only thing I'm missing is alternative style sheet selection in galeon.
[1] Just setup environment MOZILLA_FIVE_HOME before launching galeon. _________________________
Hell, I can't even link to a GPL library and place my own work into the public domain!
How many times this must be repeated? GPL doesn't prevent software to be distributed under other licenses in the same time. If you want to distribute your software as public domain that's fine. If you're using GPL libraries in it you're required to distribute it [your software, not GPL'd library] under GPL license also. The point is you cannot take the rights away the GPL quarantees. _________________________
this limits you to the ~300x200 res supported by the BTTV capture driver
Umm... There must be something broken with your BTTV. I've Hauppauge WinTV PCI and I can capture up to 924x576 (PAL). With mmap()ed capture CPU time [for capturing] needed per frame is 0-2ms. Another question is whether or not you can compress 25MB/s YUV (or 38MB/s RGB) data stream fast enough... This is the reason most programs capture at lower resolutions. _________________________
You know, those are problem ONLY if you are an unfortunate NN4.x user. For others an extra table means 'huge' 20-50 chars additional load. Hardly a big problem. How about trying Mozilla, Galeon, Konqueror or Opera (or IE or K-meleon if you use that alternative system). _________________________
It's of no use for me. UI is from the time wheel wasn't invented and making fire was something new. I wouldn't call xfig vector graphics program - it doesn't even have eps and/or ps support!
Clearly anybody who recommends xfig as a replacement for illustrator or freehand hasn't used neither of those products.
_________________________
The difference here is that 300ppi display can show a lot more information than 300dpi printer. Other than pure black or white is created with raster in print and with true shading with electronic displays. One pixel can show 16.7M different colors (24bpp) whereas you'd need at least 24 dots for the same amount of colors - perhaps a lot more because you have 4 color components (CMYK) and rastering difficulties - you can print different dots only that near each other.
Therefore I would say that 300ppi display equals to at least 1200dpi printer with real high quality paper. With technologies like ClearType text resolution on display is even higher.
_________________________
I think MSIE5+ has limit of about 4k characters for GET. I haven't reached Netscape 4.x/6.x limits yet... For POST there're no limits I'm aware of.
_________________________
My biggest problems with GIMP are keyboard shortcuts and layer support. Keyboard supports would be otherwise great (redefining and all) but you have to have image window active for them to work. Because GIMP always has that many open windows I need to spend extra time to select correct window before using shortcut... Problem with layers is that they have hard limits. Why doesn't GIMP automatically expand those limits when I draw something in the layer outside those limits? Then there wouldn't be any need to render that ugly looking layer limit dashed rectangle thing neither. GIMP should also remember what windows (brushes, layers etc) I had open last time and how those windows were positioned.
_________________________
I agree. And even if I remember all rules there's a good change I get not so accurate answer doing it in my head... I'd probably get equally valid answer simply guessing and it's much faster algorithm! Success rate 1 of 7. How about changing to 10-based system and discarding months altogether instead of inventing YAWA (yet another weekday algorithm). Simply calculate days from the start of year. That would be great solution to mm/dd/yyyy vs. dd.mm.yyyy problem also. One could simply write like 2001#157@500 (internet time). One can always dream...
_________________________
Actually with any X system I've seen that sentence could be continued with any software name. There're four reasons to use windows: good font rendering (IMO better than Macintosh for example), IE, Adobe's software and games. I can play windows games with wine and biggest titles have native ports, there's konqueror, mozilla and galeon and latest GIMP is almost usable when compared to photoshop. I'm still lacking proper font rendering and good vector drawing program for X. I'm afraid that I see freehand or illustrator equivalent on X before seeing decent font rendering. Unfortunately.
I look forward for berlin. _________________________
Um.. are you guys using address http://slashdot.org/ or http://www.slashdot.org/. If the latter then the reason is that redirect is to slashdot.org and cookie is set for www.slashdot.org. It makes me wonder why slash doesn't send domain=.slashdot.org with all cookies? _________________________
Yes, license allows it but because you're
required to include source with binary (*) and
licensee gets the same rights it might be that
you only sell one copy of the software because
licensee might sell it for cheaper or give it
out for free. This guarantees that GPLed
software is free (as in beer) in practice. GPL talks about media costs and that's pretty much what you can charge for GPLed software.
However, if ALL software required isn't under GPL this doesn't matter because if for example the compiler is under different license you cannot do "anything" with the Free software without full, perhaps expensive, license.
(*) To be exact source doesn't have to be bundled but it must be made available. _________________________
There really isn't any RAD programming system for Linux (Klyx ain't there yet.)...
In addition to RAD systems others have already listed there's also Borland Kylix. Don't know what Klyx is - I take that you don't mean KLyX (KDE version of LyX). Still I agree many windows apps have better UI than their unix equivalents. _________________________
But every desktop computer on earth uses ASCII letters out of the box.
Perhaps you mean: every IBM compatible PC? Others may show you ascii text correctly on software but I think that it doesn't map one-to-one in hardware...
_________________________
For me this sounds like a really silly way to define vision. If in the future average person has corrected vision that equals 20/10 in todays scale will it be then 20/20? How do you define Normal Vision?
Better way to describe vision quality would be to tell how long away can a person read letters printed in 20 pt font or how small characters can a person read from 10 feet. Of course we should use symbols (instead of letters) that don't have that much visual clues against other symbols (Think about letter "I" vs "M" for example - pretty much easier to see from distance than "D" vs. "O" or "V" vs "Y"). Vision would then be something like 2.0 pt/ft meaning person could read text written in 10 pt from 5 feet or 20pt from 10 feet.
Of course there's still no way to tell real vision. I mean refractive error is one thing and (best case) vision precision is another. For example I'm near-sighted and I can see in short distances with higher clarity than average person. When needed ability to see far I'm almost blind, though. _________________________
...the work that they put into archiving the posts... It is the same way that free software sell CDs with open source programs on them....
So now that Google has a copy of its own, would it be legal for Google to sell another full copy of the database to someone else? In my opinion, answer should be yes. Exactly because the points you've written. Another question is whether or not Google is allowed to do this by law. Lawyers can make miracles, unfortunately. _________________________
No, it's illegal in the same way that breaking and entering is illegal
No, it's illegal almost the way it would be illegal to play VHS movie with a VCR you made yourself. The difference is that VHS is protected by patents and CSS isn't. Bigger question is whether or not it's legal to decode MPEG2 stream with free software. AFAIK there's plenty of standards dealing with MPEG2.
_________________________
so playing a DVD at 1600x1200 is just as fast as 800x600...
I'm afraid you are wrong. Motion compensation and iDCT does nothing to do with output resolution. DVDs are something like 800x600 (cannot remember the exact res but somebody will tell it) really and what makes the difference is how you scale from original resolution that you get when you decode MPEG2 stream. For example playing DVD with the resolution of 800x600 or 1600x1200 doesn't make any difference (in speed) to me either because I scale the video with hardware (Matrox G400). That's what overlay support means - you can output for example in YUV2 format with the resolution of 640x480 and your graphics adapter converts that into RGB and scales it to 1600x1200 on the fly.
However if you're watching DVD with much less than 800x600 resolution decoding software may skip some computing and that can make difference. Or if your hardware or its drivers suck scaling the video can be done faster with software...
In short if your DVD playback is jerky no matter what resolution you select then you don't have enough CPU power (and this is the only situation where motion compensation and iDCT in hardware helps). If on the ofher hand your playback is jerky only if you select high resolution mode then you have crappy graphics adapter. If the video isn't jerky but looks bad then it's time to change software encoder. There's no magic way to get higher resolution out of video stream that doesn't have enough information. DVDs only have up to 10Mbps stream... I cannot figure out why hardware decoder should have any better image quality than good software decoder - especially if your output is "digital" (that is no tv).
_________________________
...try tracking the progress of a long-lived pipe command...
It would be nice, but in general case this is impossible. Think about something simple like "cat * | sort". How could cat know how many bytes there is still to read after reading 50% of filenames given as parameters? (It could be 99 files of 1 byte and last one of 1GB... and another process could be appending the last one!) Even less can sort say anything about how long it takes before it ends, because it has even less information about information to be sorted. Now think something like "cat/dev/urandom | grep a"...
It's however a great advantage to be able to work with files arbitrary large because pipelined programs only need commands like "read next 4000 chars" and "output these 100 chars". If OS can support 10TB files these little programs will do also. Of course sort doesn't work with arbitrary large files because it needs all data at once. Using pipeline also decreases memory requirements because commands are run in paraller and if program outputs more data than next one can handle it will sleep until next process is ready to process data again. No need to store all data by a single process. _________________________
Take the 'everything is a file' concept a bit further and turn the windowing system into a file system
I like this idea. I once though having a clipboard where contents of it would always be in ~/.etc/clipboard/text/html where the filename would also tell content type. This way you could still have your copied text in clipboard while in between using it to transfer an image (as ~/.etc/clipboard/image/jpeg). In addition clipboard would be persistent between sessions.
However there is a one little but... If we want to be backwards compatible we need to have network transparency. If I start my netscape as a remote client (in my X-server that is - I still wonder who decided this naming convetion) and select copy from my netscape menu where should the copy go? Should it be ~/.etc/text/plain or remote_box:~/.etc/text/plain. Also should I type cat/somepath/windows/netscape/1/content/text/plain in my local system or remote system to get window's content as text?
Before this I would rather see user land mount. I mean something like mount ftp://ftp.kernel.org/ ~/kernel.org without root. If I can copy same contents with an ftp client why should this be even a security issue (of course mount "driver" should be user land program running as me)? Support for at least ftp, http and ssh2 "file systems" would be nice. Of course authentication would be a little tricky, but we have kerberos and all in the future. _________________________
How is this any different than in the past? Vinyl records gave way to cassette tape....better performance
The problem I see with this change is that going to secure system doesn't provide anything valuable to end user. There's only a few people saying that vinyl or cassette is superior to CD so quite clearly it's better way to distribute music and therefore it's logical that other methods gave way to it. On the other hand all later distribution methods - especially all electronic-only-medias such as MP3 - only give end user up to CD quality.
Why would I want to "upgrade" my system when everything I get is lower quality sound and no rights at best? I'm afraid that like so some many times demonstrated, average end user that cannot even hear difference between 96Kbps MP3 and CD audio, selects new and "better" method.
_________________________
Slashdot Mirror
Amazing! In the other news there was an user that got pretty much equal speed with 56kbps modem made by Dynalink and with 56kpbs modem made by Genius...
Why do you think that two file formats with the same bitrate should create different file sizes? (Not counting different header size etc.) When comparing compressed file formats one should compare sizes of compressed files that decompress to similar results. What is similar result is totally subjective for lossy formats, of course. For example, if you think that 128kpbs MP3 sounds like 196kpbs WMA then it's MP3 that you should use.
_________________________
TV vs. web isn't that simple. TV ad stoles my time (have to wait it end no matter if I'm interested in it or not), it takes my bandwidth (how do you think that video stream comes to your tv?). Granted that TV doesn't store stream, but you can use VCR(/TiVo or like) or disable browser cache to make things equal.
In fact TV ads are much worse than web ads because you can always close those pop up windows as soon as they appear or click stop button to stop animations (doesn't work for flash ads though). There's much software to filter incoming web pages to contain no ads. etc etc etc.
I think the difference is that because web is interactive media you expect it to do only what you've asked for; for example you click link, you expect browser to change page instead of opening ten new windows and playing some songs. Watching tv and seeing ending titles coming you immediatly expect to see some ads next so you change channel or close tv. In fact choise to not to see ads when using tv is always opt out (per ad!) instead of opt in (global!) in web (turn on javascript, java and images).
_________________________
$man X
...
...
The X Consortium requests that the following names be used when referring to this software:
X
X Window System
X Version 11
Window System, Version 11
X11
X Window System is a trademark of X Consortium, Inc.
I don't see XWindow there... If you want to whine get your facts right.
_________________________
[1] Just setup environment MOZILLA_FIVE_HOME before launching galeon.
_________________________
How many times this must be repeated? GPL doesn't prevent software to be distributed under other licenses in the same time. If you want to distribute your software as public domain that's fine. If you're using GPL libraries in it you're required to distribute it [your software, not GPL'd library] under GPL license also. The point is you cannot take the rights away the GPL quarantees.
_________________________
Umm... There must be something broken with your BTTV. I've Hauppauge WinTV PCI and I can capture up to 924x576 (PAL). With mmap()ed capture CPU time [for capturing] needed per frame is 0-2ms. Another question is whether or not you can compress 25MB/s YUV (or 38MB/s RGB) data stream fast enough... This is the reason most programs capture at lower resolutions.
_________________________
You know, those are problem ONLY if you are an unfortunate NN4.x user. For others an extra table means 'huge' 20-50 chars additional load. Hardly a big problem. How about trying Mozilla, Galeon, Konqueror or Opera (or IE or K-meleon if you use that alternative system).
_________________________
Clearly anybody who recommends xfig as a replacement for illustrator or freehand hasn't used neither of those products.
_________________________
Therefore I would say that 300ppi display equals to at least 1200dpi printer with real high quality paper. With technologies like ClearType text resolution on display is even higher.
_________________________
I think MSIE5+ has limit of about 4k characters for GET. I haven't reached Netscape 4.x/6.x limits yet... For POST there're no limits I'm aware of.
_________________________
My biggest problems with GIMP are keyboard shortcuts and layer support. Keyboard supports would be otherwise great (redefining and all) but you have to have image window active for them to work. Because GIMP always has that many open windows I need to spend extra time to select correct window before using shortcut... Problem with layers is that they have hard limits. Why doesn't GIMP automatically expand those limits when I draw something in the layer outside those limits? Then there wouldn't be any need to render that ugly looking layer limit dashed rectangle thing neither. GIMP should also remember what windows (brushes, layers etc) I had open last time and how those windows were positioned.
_________________________
I agree. And even if I remember all rules there's a good change I get not so accurate answer doing it in my head... I'd probably get equally valid answer simply guessing and it's much faster algorithm! Success rate 1 of 7. How about changing to 10-based system and discarding months altogether instead of inventing YAWA (yet another weekday algorithm). Simply calculate days from the start of year. That would be great solution to mm/dd/yyyy vs. dd.mm.yyyy problem also. One could simply write like 2001#157@500 (internet time). One can always dream...
_________________________
Actually with any X system I've seen that sentence could be continued with any software name. There're four reasons to use windows: good font rendering (IMO better than Macintosh for example), IE, Adobe's software and games. I can play windows games with wine and biggest titles have native ports, there's konqueror, mozilla and galeon and latest GIMP is almost usable when compared to photoshop. I'm still lacking proper font rendering and good vector drawing program for X. I'm afraid that I see freehand or illustrator equivalent on X before seeing decent font rendering. Unfortunately.
I look forward for berlin.
_________________________
Um.. are you guys using address http://slashdot.org/ or http://www.slashdot.org/. If the latter then the reason is that redirect is to slashdot.org and cookie is set for www.slashdot.org. It makes me wonder why slash doesn't send domain=.slashdot.org with all cookies?
_________________________
Yes, license allows it but because you're required to include source with binary (*) and licensee gets the same rights it might be that you only sell one copy of the software because licensee might sell it for cheaper or give it out for free. This guarantees that GPLed software is free (as in beer) in practice. GPL talks about media costs and that's pretty much what you can charge for GPLed software.
However, if ALL software required isn't under GPL this doesn't matter because if for example the compiler is under different license you cannot do "anything" with the Free software without full, perhaps expensive, license.
(*) To be exact source doesn't have to be bundled but it must be made available.
_________________________
This sounds interesting. Does anybody have more information about this?
_________________________
In addition to RAD systems others have already listed there's also Borland Kylix. Don't know what Klyx is - I take that you don't mean KLyX (KDE version of LyX). Still I agree many windows apps have better UI than their unix equivalents.
_________________________
Perhaps you mean: every IBM compatible PC? Others may show you ascii text correctly on software but I think that it doesn't map one-to-one in hardware...
_________________________
For me this sounds like a really silly way to define vision. If in the future average person has corrected vision that equals 20/10 in todays scale will it be then 20/20? How do you define Normal Vision?
Better way to describe vision quality would be to tell how long away can a person read letters printed in 20 pt font or how small characters can a person read from 10 feet. Of course we should use symbols (instead of letters) that don't have that much visual clues against other symbols (Think about letter "I" vs "M" for example - pretty much easier to see from distance than "D" vs. "O" or "V" vs "Y"). Vision would then be something like 2.0 pt/ft meaning person could read text written in 10 pt from 5 feet or 20pt from 10 feet.
Of course there's still no way to tell real vision. I mean refractive error is one thing and (best case) vision precision is another. For example I'm near-sighted and I can see in short distances with higher clarity than average person. When needed ability to see far I'm almost blind, though.
_________________________
So now that Google has a copy of its own, would it be legal for Google to sell another full copy of the database to someone else? In my opinion, answer should be yes. Exactly because the points you've written. Another question is whether or not Google is allowed to do this by law. Lawyers can make miracles, unfortunately.
_________________________
No, it's illegal almost the way it would be illegal to play VHS movie with a VCR you made yourself. The difference is that VHS is protected by patents and CSS isn't. Bigger question is whether or not it's legal to decode MPEG2 stream with free software. AFAIK there's plenty of standards dealing with MPEG2.
_________________________
I'm afraid you are wrong. Motion compensation and iDCT does nothing to do with output resolution. DVDs are something like 800x600 (cannot remember the exact res but somebody will tell it) really and what makes the difference is how you scale from original resolution that you get when you decode MPEG2 stream. For example playing DVD with the resolution of 800x600 or 1600x1200 doesn't make any difference (in speed) to me either because I scale the video with hardware (Matrox G400). That's what overlay support means - you can output for example in YUV2 format with the resolution of 640x480 and your graphics adapter converts that into RGB and scales it to 1600x1200 on the fly.
However if you're watching DVD with much less than 800x600 resolution decoding software may skip some computing and that can make difference. Or if your hardware or its drivers suck scaling the video can be done faster with software...
In short if your DVD playback is jerky no matter what resolution you select then you don't have enough CPU power (and this is the only situation where motion compensation and iDCT in hardware helps). If on the ofher hand your playback is jerky only if you select high resolution mode then you have crappy graphics adapter. If the video isn't jerky but looks bad then it's time to change software encoder. There's no magic way to get higher resolution out of video stream that doesn't have enough information. DVDs only have up to 10Mbps stream... I cannot figure out why hardware decoder should have any better image quality than good software decoder - especially if your output is "digital" (that is no tv).
_________________________
It would be nice, but in general case this is impossible. Think about something simple like "cat * | sort". How could cat know how many bytes there is still to read after reading 50% of filenames given as parameters? (It could be 99 files of 1 byte and last one of 1GB... and another process could be appending the last one!) Even less can sort say anything about how long it takes before it ends, because it has even less information about information to be sorted. Now think something like "cat /dev/urandom | grep a"...
It's however a great advantage to be able to work with files arbitrary large because pipelined programs only need commands like "read next 4000 chars" and "output these 100 chars". If OS can support 10TB files these little programs will do also. Of course sort doesn't work with arbitrary large files because it needs all data at once. Using pipeline also decreases memory requirements because commands are run in paraller and if program outputs more data than next one can handle it will sleep until next process is ready to process data again. No need to store all data by a single process.
_________________________
I like this idea. I once though having a clipboard where contents of it would always be in ~/.etc/clipboard/text/html where the filename would also tell content type. This way you could still have your copied text in clipboard while in between using it to transfer an image (as ~/.etc/clipboard/image/jpeg). In addition clipboard would be persistent between sessions.
However there is a one little but... If we want to be backwards compatible we need to have network transparency. If I start my netscape as a remote client (in my X-server that is - I still wonder who decided this naming convetion) and select copy from my netscape menu where should the copy go? Should it be ~/.etc/text/plain or remote_box:~/.etc/text/plain. Also should I type cat /somepath/windows/netscape/1/content/text/plain in my local system or remote system to get window's content as text?
Before this I would rather see user land mount. I mean something like mount ftp://ftp.kernel.org/ ~/kernel.org without root. If I can copy same contents with an ftp client why should this be even a security issue (of course mount "driver" should be user land program running as me)? Support for at least ftp, http and ssh2 "file systems" would be nice. Of course authentication would be a little tricky, but we have kerberos and all in the future.
_________________________
The problem I see with this change is that going to secure system doesn't provide anything valuable to end user. There's only a few people saying that vinyl or cassette is superior to CD so quite clearly it's better way to distribute music and therefore it's logical that other methods gave way to it. On the other hand all later distribution methods - especially all electronic-only-medias such as MP3 - only give end user up to CD quality.
Why would I want to "upgrade" my system when everything I get is lower quality sound and no rights at best? I'm afraid that like so some many times demonstrated, average end user that cannot even hear difference between 96Kbps MP3 and CD audio, selects new and "better" method.
_________________________