> 750 Megabyte was amazing > especially considering that the biggest discs at the > time were 650 MB! (read "D'oh!")
For what it's worth, a "650MB" CD actually does have a capacity of around 750MB. The CD-ROM standard apparently allows for a lot of error-correcting redundancy, since songs and video originally put onto the media did not have built-in error correction, and since the media itself was prone to scratches and such.
The XCD spec was pretty cool. It just filled all the sectors fully and relied on software error correction (the kind that you find in modern movie formats). So you could burn something like 804MB to a 700MB CD-R.
> Horrible 'brand'. Once worked in a computer store for a while. > We sold about 20 of their TFTs before we figured out
If it helps, I have at least a little experience with their actual optical writing products, which are what this conversation is about. I have their DVD+RW product (400A or something), and it's pretty fantastic. No coasters on either DVD+R or CD-R writes, and my only failure was when I accidentally bought a pack of 50 DVD-R media, and it didn't burn because it wasn't DVD+R. But I had bought it on the cheap (under half a dollar per disc), so it didn't really hurt to donate it to the local LUG.
--
-JC
http://www.jc-news.com/coding/freedom/
Real virtual desktops == Multiple X sessions
on
The GNOME Roadmap
·
· Score: 1
> You know what I'd like to see? Real virtual desktops. The current "virtual desktops" are > really just virtual screens, not desktops. Full virtual desktops should act as completely > separate desktops, with their own set of icons, etc. Obviously this would not be for > everyone, but I would love to see it as a user-selectable option.
Well, you can kind of do that my editing inittab (or placing an extra xinit entry somewhere in your startup scripts) so that it starts up multiple X sessions. Let's say, one desktop on:0 and another on:1. Then, you can hit CTRL+ALT+F7 to get to the first desktop and CTRL+ALT+F8 to get to the second desktop. And you could probably add an icon that binds to that key combination or does something else to switch to the other desktop. And you could have KDE in one and GNOME in the other, or you could have two XFce desktops (though they'd probably have to be under two separate users, though not necessarily).
There are perhaps problems with this approach (for instance, switching from one vterm to another could take a second or three, and that's annoying if you want instant switching), but it is essentially what you're asking for.
For what it's worth, KDE supports Mac-style menu bars. This works for any KDE app and probably most Qt apps (with the Opera web browser, it works, but you have to tell Opera to hide its menu bar or you see the menu bar in both places). Also, KDE remembers which desktop you asked a program to load in. Not that I'm evangelizing or anything....
--
-JC
FreeDominion, a WIP Civ clone
http://www.jc-news.com/coding/freedom/
> But 'm' is stil faster to type than 'd' (m is closer to the space key which you need to type afterwards, you > can do it by just taping m+space by one wet finger), so mandrake is light-years ahead of debian...
No. 'd' is on the home row. To type 'm', you'd have to move your right index finger before pushing down. Pushing 'd' required no such movement.
That said, I am a proud participant of Mandrake's new index finger exercise program....
> The fact that insignia is commonly accepted as a > singular today makes that glaringly obvious.
FWIW, it's not only accepted by common folk. It's acceptable to use "insignias" as a plural by both the American Heritage Dictionary and Merriam-Webster.
> The thoery of Relativity posits that all measurements are taken from some frame of > reference, and it is impossible for an object to go faster than light for any frame of reference.
No. Well, at least, slightly no. It is impossible for an object to accelerate to (and hence accelerate from below to above) the speed of light. This is different from it being impossible for an object to have a velocity greater than c. The only caveat is that the object has to have *always* had been travelling at a velocity greater than c.
This seems confusing because the measured momentum and energy from you as an observer of this superluminal object would be imaginary numbers. But that doesn't mean that the rest mass needs to be anything other than normal.
But that's just semantics on my part. The rest of your post seems groovy enough at a first glance.:)
> I never truely appreciated autscroll because it > puts too much pressure on me to read fast and i > read at really varying rates.
The autoscroll feature in CSpotRun is variable speed. You can use the Up/Down buttons to make it go faster or slower.
Of course, my way is to set the fonts down to a very small size (like the Tiny font on a 240x320 HandEra 330, perhaps the best monochrome PalmOS PDA in history) so that I only have to press the button to advance the screen once or twice a minute.
Nowadays, I'm lazy. I just use command line tools to convert whatever the format of the book is into html, then split it into 15K-or-so-sized chunks and upload them somewhere so that I can read them with a web browser on my Treo. ^_^
> You can have information go faster than the speed of light. Your electrons don't move very > far. They just bump into the next one, which bumps into the next one. Imagine this: you've got > a 1 light year long pole. You shoot a photon/wave of light parallel to the pole. It hits > then end in 1 year. I push on the end of my pole. The end 1 light year away moves almost > immediately (small lag for compression of the material). I just transferred information faster > than you did. Electrons have a longer lag than a metal pole will, but not enough to slow it down > past light speed.
Your post is inaccurate. When you push on one end of an object, the other end does not immediately start moving. You just produce a wave of compression from one end to the other (kind of like how tapping one end of the object only produces vibration at the other end at the speed of sound, which is certainly not infinite), and the speed of this wave depends on the rigidity of the object. For instance, if you push on one end of a rubber pole, it will take a lot longer from the other end to likewise move than if the pole were made of steel. But there is no object quite rigid enough, even in theory, that does this faster than the speed of light.
Here's a place that explains it better than I can:
http://www.vscht.cz/mat/Pavel.Pokorny/physics/FT L. html
"If you have a long rigid stick and you hit one end, wouldn't the other end have to move immediately? Would this not provide a means of FTL communication?
Well it would if there were such things as perfectly rigid bodies. In practice the effect of hitting one end of the stick propagates along it at the speed of sound in the material which depends on its elasticity and density. Relativity places an absolute limit on material rigidity so that the speed of sound in the material will not be greater than c."
> The 6 GHz is a little fishy to me, and here's why:
> 6 GHz --> 0.17 ns per cycle. Light travels 5 cm (about two inches) in 0.17 ns, and information > cannot travel faster than light. This means that even at the speed of light (electrical signals in > typical electronics propogate at ~0.8 c, IIRC) it will take almost the entire clock cycle to get > information across the chip, never mind whatever time it takes the transistors to respond.
Pipelining is a well-understood technique that was introduced to the x86 world with the 80486 processor (that's the one that was a generation before the Pentium for you new folk). The idea is that each stage of the pipeline acts as a dedicated, specialized processor with limited functionality that hands off its results to the next stage. It is analogous to the Assembly Line, where each worker has a specialized task and hands off each in-progress product to the next worker in the chain.
The key here is that the electrons only have to pass through *each stage* in a single cycle. If your cpu is 4cm across, then the electrical signals (according to your number) would take 0.17ns to cross it. But if the cpu were separated into ten stages, then the signals would only need to traverse 0.4cm during each cycle.
Naturally, the tradeoff is that when you increase the number of stages, then the number of cycles that each instruction needs to complete increases, so you get penalties from erroneous predictions and cache misses and the like.
So your 6GHz limit only applies if your cpu is a one stage processor. Most consumer desktop processors have ten to fifteen stages. The Pentium 4, depending on how you count it, has as many as twenty-eight stages.
> In the meantime, those nursing dreams of 100 GHz chips had better look beyond nanotech to > picotech-- atom-sized transistors.:-P
I don't think that I disagree here, despite my above comments. To do these frequencies without a dramatic decrease in transistor sizes would require an absolutely obscene level of pipelining, to the point that performance would take massive hits and operating temperatures would be quite Venusian.
> Jobs are like relationships. It's always easier > to get a new job offer when you've already got > an existing job -- even a crappy one.
Bullshit. The only time that I ever got jobs offered to me was back when I ran a tech news site that had ten thousand visitors a day, back when I was known enough that people like Anand (from anandtech.com) and Johan (from Aceshardware.com) would pop messages to me just for a friendly chat. Back then, when I was actually sort of *popular*, I'd occasionall get a job offer.
Now, there's *nothing*. I'd send resume after resume for months before getting so much as an interview. My resume has been looked at many times by both professional and amateur resume specialists, so I can only assume that it's at least *decent*. My standards are very low, as after a decade and a half of work experience (and nearly a decade of computer work experience) I'm certainly looking for no more than what is considered average in my area for a recent graduate (I graduated many year ago) -- that being $30K plus benefits in Long Island, a fish-shaped bit of land near to but much larger than Manhattan which "enjoys" a high cost of living. I try to hold back on my luxury purchases. I have broadband internet, a cell phone and my still existing domain being hosted, and I have to pay rent to my family for housing and feeding me. My weekends are a little expensive, as I go to the movies with my friends every week, but I managed to get a wad of lower cost bulk movie tickets that lower that cost. In the end, though, I'm a computer specialist who hesitates to purchase a computer to replace his ailing machine, a Duron-800 that can no longer work reliably in window managed modes, so it's basically running in frame buffer linux. If it weren't for giFT, fbxine, fbi, nethack, cone, links and fbtv, I'd go insane.:P
Hmmm. Looking over the message preview, I realize that I had a point, and it was this: 90% of decent jobs are acquired because you know somebody who can help you fill the position (whether it be a family member or a friend or whatnot). I know this because of the purely unstatistical fact that the least aggressive person in my three man clique got a $53K/yr job for the FAA because his mother worked there, whereas the other two are merely hard workers (one of them quite possibly exceptionally talented) who make under $30K/yr.
> #1 - fresh out of school, learned a bit of c++ and java in school, knows all the academic stuff; ===} 30-38K
Haha. You must not live in the suburbs of New York City.
BS in CompSci, cross-platform C++, Qt, Perl, CGI, HTML, javascript, css, FreeBSD, Linux, Apache, Windows, Qmail, Vpopmail, courier-imap, WAP gateways, helpdesk, MySQL, Nessus, client PC maintenance, server troubleshooting, cat5 crimping, network engineering, Netware 4.x/5.x, remote management tools, web design/development, online journalism, 16 years of general work experience, 9 years of computer-related work experience ===) $27K
It's all who you know, apparently. I know nobody of import.
Wow, we could not possibly have more differing experiences here! My 5.1 sound card has a CM8378, too, but it was Windows that gave me the problems. An earlier version of Mandrake autodetected it and everything went perfectly. It took me *weeks* of fiddling around before I could get the 8378 working in Windows 2000. In the end, I had to boot into Linux to get information about the sound card (including the ID of the chip) before I was able to find drivers that worked in Win2k.
I also have continuing problems getting my Hauppauge TV card working in Windows 2000. And when I swapped my CD-R drive for a DVD+R drive, Windows 2000 failed the hardware detect and refused to reboot (an error with something like "redbook.sys") until I reinstalled.
Then again, once when I moved the location of my CD-R, Mandrake failed to figure out where it moved to, and I had to learn how the "/etc/fstab" file worked. And I can't get Mandrake to work with my Lexmark X83 multifunction printer, largely because Lexmark isn't very active with Linux driver development. And my gyroscopic mouse required me to add a couple lines into two text files before it worked well.
It works both ways, see? I try to avoid complaining about it in either direction, but I do get annoyed when people seem to think that problems only occur on one side of the bridge.
> Now, if it happens to be one of the applications > bundled with Mandrake I can just use the > software installer
> Other times (This is usually where I give up), > the computer starts acting as if it's on crack: > rpm -i annoying-dependency.rpm
I mean no disrespect, but why are you using "rpm -i" to install a downloaded RPM file if you're in Mandrake? If you're in Mandrake, you can use the software installer, just as if you were installing a regular RPM from the CD or from Mandrake's sources. In command line speak, you do: "urpmi thisprogramthatidownloaded.rpm" And it'll do all that automatic dependency fixing.
If you install from source, then it's a little difficult, but not too much. For instance, I was installing QTorrent recently, which is an excellent bittorrent client that uses fewer resources than the standard btdownloadfoo.py and puts everything together in a nicely manageable gui. I unpacked the source tarball and saw that it had the "setup.py" (with these -- and it took me far too long to figure this out, you have to do "python setup.py install"). When I tried to install it, it bitched about there being no "PyQt". So I did a "urpmi PyQt", and it autodownloaded and autoinstalled that particular dependancy. Yay, QTorrent could then install.
It's like that. If./configure says "missing dependencies foo, bar, baz and ook", you type "urpmi foo bar baz ook" as root, then you go back and do./configure again. It's not the sort of thing that a normal user should have to go through (which is why somebody needs to put qtorrent into Mandrake's rpm tree!), but for folks like you, who know enough to do far too difficult dependency hunting, this should be a breeze.
> I seriously wonder why nobody has implemented > binary installation/uninstallation routines for > the Linux desktops yet. What's the damn holdup? > Users need to be able to buy a Linux application > from a store, take it home, and stick in a CD to > get an autoplay installer.
Firstly, I object to your statement about binary installation/uninstallation not existing. That's the whole point of urpmi and apt-get (or their gui counterparts, like Mandrake Install and whatever it is for apt).
But you do have a bit of a point about installs from CDs. Actually, it wouldn't be *too* difficult. I mean, you can make a binary install by just making a.tar.gz file with all the files in the proper relative paths (./bin/,./lib/,./share/progname/,./etc/progname/). The only real stumbling block is that the installer has to look at the OS and find out which variant of linux it is (that is, whether it uses debs or rpms). The installer would just have to add entries to the specific distro's "installed files" system settings so that the distro's normal uninstaller can handle it.
Basically, there would need to be two branches: If the OS is Red hat, Fedora, Mandrake or SuSe, the installer adds to the installed RPMs database. If the OS is Debian or... well, whatever else uses.deb files, then it'd write to the installed DEBs database. I know that Mandrake has a tool to uninstall programs graphically, and the other systems likely do, as well. I mean, heck, if I applied myself for a few days (and I didn't have one of my bizarre depression attacks), *I* could probably write something like this, and I am most certainly not very experienced with programming.
Oh, and Gentoo doesn't count, because those people don't believe in using non-OSS.;P
> Of course, to get that truly working well, you'd > want a sane, robust programming library in the > likes of.NET and Cocoa--none of this absolutely > ridiculous QT/GTK/wxWindows/whatever nonsense > that are merely hacks to get widgets up on X.
You're silly. Almost all the programs that I run in Linux and FreeBSD use a single widget library: Qt. In windows, nearly every program I run has a different widget set. There's Opera, which uses Qt. There's Office, which uses the Office widgets (*NOT* the same as the MS Windows widgets!). There's the built-in windows programs. There's Eudora, which seems to use its own thing. There's Mozilla, which uses its own bizarre stuff. And Ultra VNC seems pretty nonstandard, too. I realize that, for most users, the majority of programs seem to be written with the same widgets, but a large part of that reason is that most users predominantly run programs written by one company -- Microsoft. Of *course* Microsoft apps will look like other Microsoft apps. That's just like how KDE apps look like other KDE apps (and I dare say that KDE is even a little more consistant with this, given that KOffice and other KDE apps use the same widgets).
Bah, finished ranting. Maybe I'll make some kind of generic installer for Linux now....
> As it is now - Real offers the ONLY player that > incorporates all major music formats WMA, MPEG, > ATRAC3, with exception for Apple's AAC.
I may be incorrect, but I think that there are more than one such player. MPlayer, an open-source video player that runs on Windows, Linux and other platforms, plays everything in the universe, including the formats you mentioned above. Xine is a similar effort; I don't know if it works in Windows, but it probably has port either here or coming eventually, and it *probably* supports all those codecs. Meanwhile, Media Player Classic, a Windows-only project that uses the superior (well, to WMP7 or WMP8, at least) interface of Windows Media Player 6 but combines it with a load of codecs, may also support all these formats.
> I certainly did not notice that CAN-SPAM became effective 1/1/04. Or actually, > my filters are still filtering out a very similar number of messages.
That's actually a really good point. Some time late last year, I was getting about 1200 emails a day, nearly all of it spam. It absolutely inundated me and nearly rendered me helpless. I learned how to code mail filtering programs in perl, then I learned how to use stuff like procmail and spamassassin and clamav, and when that stuff is used in conjunction with a mail client with bayesish filtering capabilities (like Mozilla), spam becomes manageable.
But I checked my procmail logs a couple weeks ago, and I discovered that the number of daily emails I'm getting (again, nearly all of them spam) have jumped to over 2000.
I guess the law really didn't do anything after all.
> I dunno, I'm as big a South Park fan as anybody > but season 8 has been consistently bad so far.
That's odd. I only found the 'Passion'-related episode to be weak. The other episodes so far have been extremely funny, moreso than the last two seasons have been.
Of course, maybe I'm just using South Park to compensate for my continuing battle against Futurama-oriented withdrawal symptoms.
> Earth to google... get in easyier[sic] address bar searching
Um, Every single graphical desktop web browser in the universe, save for one (yeah, THAT one), supports keyword features that make it completely pointless for individual websites to bother. In Opera, since last millennium, you could type "g bunch of search terms". Mozilla could be easily configured to use the same syntax, or you could change the keyword from "g" to something else (like "IWannaKnowMoreAbout bunch of search terms"). Konqueror does it like "gg:bunch of search terms" (and, I think, "g:bunch of search terms" to do the "I feel lucky" thing on Google).
I have Opera set up with a few custom keywords. If I type "def someword", then the browser does a search on dictionary.com for "someword". Useful feature. Totally pointless for the feature to be built into the website. Unless you're using a web browser with a 1990s feature set, like lynx or IE.
> Can somebody tell me why it seems everything > requiring a cross-platform C++ gui widgets > seems to be written with QT?
It's largely because Qt is really, really insanely easy to use. The object classes are very intuitive for programming.
Regarding licensing issues, there is a GPL version for native Win32. It's not being actively updated, but it did get far enough to the point where most of my programs (the ones that didn't use other libraries, at least) would cross compile and run pretty decently. There were some speed issues, but most of them disappeared after I figured out how to get the compiled programs running without popping up a stderr command window.
> The only "innovation" KDE and GNOME can profess > is having tons more pointless applets running on > their panels, and running a lot slower than > Windows on the same hardware.
KDE has session management to a level that I've never seen elsewhere. It can set the Z-level of any window to be above of below other windows. You can turn off the titlebar of any window if you need it to disappear. You can fullscreen any maximizable application. Widgets across all KDE applications are instantly, globally skinnable. I can run KDE on Win32, Linux and FreeBSD and the installs are easy compared to any other complex desktop environment. Mouse gestures and shortcut key combos can be easily set for global and application-specific options. The KDE file dialog box is unparalleled -- it has its own bookmark feature, but it also has collapsible and directly editable alternative to the horrifyingly static and bothersome "Places" bar in Win32, and you can type in new directory locations and thusly change directories without accidentally erasing the file name in the dialog box, which is a major failing of Win32 and most other file dialogs. Additionally, the file dialogs have relatively protocol-agnostic access of network-accessible files. You can load and save stuff that are shared via SMB, NFS, FTP, even SSH, as if they were saved on your local disk. Heck, you can browse IMAP as if it were a file sytem! Titlebar buttons can be easily shifted around and customized to avoid the horribly fatal "kill button next to maximize button" flaw in most operating systems and environments. Devices can be set to appear on the desktop when they mount. There's a useful graphical disk usage view included in the default file manager.
And, yeah, there are a ton of incredibly useful panel applets.
> Not to mention using just ONE library instead > of multiple "toolkits" doing the widget work > that should be done by the desktop anyway.
Win32 has this problem, too. Pretty much every single program that I'm using right now on this Windows 2000 install (Opera, Mozilla Mail, cmd.exe, Cygwin/X, vncviewer, OpenOffice.org) has different widgets. What's the big deal if Linux does the same thing? At least KDE has more consistency than, say, Microsoft (whose most popular program -- MS Office -- uses widgets different from the operating system).
> Windows just works, and works with more > hardware without extraconfiguration.
That's not true for everybody. I had to turn ten kinds of hell to get my 5.1 audio card to work in Windows. In the end, I had to boot into Linux (which had detected it properly) to find out what the audio card's chipset was, which finally gave me enough clues to get drivers for the card.
Yes, though, Windows has ubiquitous driver support.
> Ummm rootkits?
I've never been rootkitted, and I'm very horrible with security.
> When more people switch over to > linux you will see more viruses out there for > linux because right now windows is an easy > targetwith a ton of machines out there.
And how will these viruses spread? Unlike Windows, the dominant email apps for Linux do not run applications when you click on them. Unlike Windows, viruses on Linux can't take control of system files. Unlike Windows, Linux computers start up with unnecessary ports turned off.
> People much rather write a virus that will > effect a much larger population.
I don't think you get it. Why do you think that there are multiple means of package management? Why do you think that different distributions handle things differently? Why do you think that Linux advocates and Open Source Software programmers make such great strides towards making sure that applications are available not only on Linux but on other operating systems such as FreeBSD, OS X and QNX? It's because having a monoculture is *BAD*. We don't *want* every computer in the world to run Linux. That would be *stupid*, even though Linux has a far safer security model than Windows. We want operating system usage to be distributed more or less evenly among different models, just like it was in the old days, when viruses *weren't* dangerous, and when a stupid move by one OS maker didn't negatively affect every single computer user on the planet!
> If linux was secure from viruss,why are there > linux virus scanners?
They're to scan for *Windows* viruses. You know, like if you had a mail server on your Linux box? I do, and I filter mail through clamav in order to prevent infection on a Windows box if I chance to check my mail on one.
> with linux you need to figure out why your > soundisnt working and then configure the driver > and what not and edit configfiles.
You mean, "with linux you have at least one way of getting sound working if your sound card doesn't work". This is opposed to "with windows if your source card doesn't work you can only cry a lot and try to reboot or just buy a new sound card".
Linux isn't perfect, and I certainly hope that no single operating system ever captures more than a quarter of the desktop or business market, because that would lead to disaster. But Linux has saved me from a ton of frustration that Windows had caused me, and it's a lot easier to use (especially installing software, which Linux reduced to a simple, single step from the more popular many-step process.
> Yesterday I received 144 pieces of Spam... If > it wasn't for SpamBayes, I probably would have > abandoned email altogether by now. These guys > are rubish.
Ho ho! You should consider yourself lucky, my good fellow. According to my procmail.log file, I'm averaging upwards of two thousand instances of spam mail per day. You can't *possibly* imagine the trauma that this sort of thing can cause.
> The bus, the subway, the train, the bike, > and walking....but that's crazy talk.
It *is* crazy talk, and here's why: It's a forty minute drive to work for me, and a fifty minute drive back (heavier traffic in the afternoon). This is annoying enough, but if I took a train, it'd be:
A) Twenty minute walk to the train station (I'm not trusting enough to tie up a bike for ten hours a day) B) Ten minute train in the wrong direction to get to a central junction C) Five minutes waiting for the switchover train D) Thirty minute train to get to the nearest station E) Thirty minute walk, because the nearest station isn't near to my work at all (neither are any convenient buses).
Yay. That's ninety-five minutes, or about twice as long as my regular commute. Even with the price of gas, the train ends up being more expensive than my car. There's also the fact that I'd have to spend nearly an hour walking -- *each way*! So we get an hour less sleep each day and almost two hours of additional walking each day, resulting in a total zombie worker who can't code or set up BSD servers properly anymore.
Damnit, I hate my job with a nigh impossible passion, and I'm no fan of driving cars, but public transportation would be a worse case scenario (eg, I'd go for unemployment first).
Slashdot Mirror
> 750 Megabyte was amazing
> especially considering that the biggest discs at the
> time were 650 MB! (read "D'oh!")
For what it's worth, a "650MB" CD actually does have a capacity of around 750MB. The CD-ROM standard apparently allows for a lot of error-correcting redundancy, since songs and video originally put onto the media did not have built-in error correction, and since the media itself was prone to scratches and such.
The XCD spec was pretty cool. It just filled all the sectors fully and relied on software error correction (the kind that you find in modern movie formats). So you could burn something like 804MB to a 700MB CD-R.
--
-JC
http://www.jc-news.com/coding/freedom/
> Horrible 'brand'. Once worked in a computer store for a while.
> We sold about 20 of their TFTs before we figured out
If it helps, I have at least a little experience with their actual optical writing products, which are what this conversation is about. I have their DVD+RW product (400A or something), and it's pretty fantastic. No coasters on either DVD+R or CD-R writes, and my only failure was when I accidentally bought a pack of 50 DVD-R media, and it didn't burn because it wasn't DVD+R. But I had bought it on the cheap (under half a dollar per disc), so it didn't really hurt to donate it to the local LUG.
--
-JC
http://www.jc-news.com/coding/freedom/
> You know what I'd like to see? Real virtual desktops. The current "virtual desktops" are
:0 and another on :1. Then, you can hit CTRL+ALT+F7 to get to the first desktop and CTRL+ALT+F8 to get to the second desktop. And you could probably add an icon that binds to that key combination or does something else to switch to the other desktop. And you could have KDE in one and GNOME in the other, or you could have two XFce desktops (though they'd probably have to be under two separate users, though not necessarily).
> really just virtual screens, not desktops. Full virtual desktops should act as completely
> separate desktops, with their own set of icons, etc. Obviously this would not be for
> everyone, but I would love to see it as a user-selectable option.
Well, you can kind of do that my editing inittab (or placing an extra xinit entry somewhere in your startup scripts) so that it starts up multiple X sessions. Let's say, one desktop on
There are perhaps problems with this approach (for instance, switching from one vterm to another could take a second or three, and that's annoying if you want instant switching), but it is essentially what you're asking for.
For what it's worth, KDE supports Mac-style menu bars. This works for any KDE app and probably most Qt apps (with the Opera web browser, it works, but you have to tell Opera to hide its menu bar or you see the menu bar in both places). Also, KDE remembers which desktop you asked a program to load in. Not that I'm evangelizing or anything....
--
-JC
FreeDominion, a WIP Civ clone
http://www.jc-news.com/coding/freedom/
> But 'm' is stil faster to type than 'd' (m is closer to the space key which you need to type afterwards, you
> can do it by just taping m+space by one wet finger), so mandrake is light-years ahead of debian...
No. 'd' is on the home row. To type 'm', you'd have to move your right index finger before pushing down. Pushing 'd' required no such movement.
That said, I am a proud participant of Mandrake's new index finger exercise program....
--
-JC
http://www.jc-news.com/coding/freedom/
> The fact that insignia is commonly accepted as a
> singular today makes that glaringly obvious.
FWIW, it's not only accepted by common folk. It's acceptable to use "insignias" as a plural by both the American Heritage Dictionary and Merriam-Webster.
--
-JC
> The thoery of Relativity posits that all measurements are taken from some frame of
:)
> reference, and it is impossible for an object to go faster than light for any frame of reference.
No. Well, at least, slightly no. It is impossible for an object to accelerate to (and hence accelerate from below to above) the speed of light. This is different from it being impossible for an object to have a velocity greater than c. The only caveat is that the object has to have *always* had been travelling at a velocity greater than c.
This seems confusing because the measured momentum and energy from you as an observer of this superluminal object would be imaginary numbers. But that doesn't mean that the rest mass needs to be anything other than normal.
But that's just semantics on my part. The rest of your post seems groovy enough at a first glance.
--
-JC
http://www.jc-news.com/parse.cgi?coding/main
http://www.jc-news.com/coding/SFi/
http://www.jc-news.com/coding/freedom/
> Though this is up to speculation, I think Anakin was the old hope.
Considering that the balance of Jedi and Sith was a hundred to one before Anakin was born, I don't think that they *needed* an old hope.
--
-JC
http://www.jc-news.com/coding/freedom/
> What media lasts LONGEST?
> I mean, other than paper, or stone.
What bests paper or rock? Obviously, scissors.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> I never truely appreciated autscroll because it
> puts too much pressure on me to read fast and i
> read at really varying rates.
The autoscroll feature in CSpotRun is variable speed. You can use the Up/Down buttons to make it go faster or slower.
Of course, my way is to set the fonts down to a very small size (like the Tiny font on a 240x320 HandEra 330, perhaps the best monochrome PalmOS PDA in history) so that I only have to press the button to advance the screen once or twice a minute.
Nowadays, I'm lazy. I just use command line tools to convert whatever the format of the book is into html, then split it into 15K-or-so-sized chunks and upload them somewhere so that I can read them with a web browser on my Treo. ^_^
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> You can have information go faster than the speed of light. Your electrons don't move very
T L. html
> far. They just bump into the next one, which bumps into the next one. Imagine this: you've got
> a 1 light year long pole. You shoot a photon/wave of light parallel to the pole. It hits
> then end in 1 year. I push on the end of my pole. The end 1 light year away moves almost
> immediately (small lag for compression of the material). I just transferred information faster
> than you did. Electrons have a longer lag than a metal pole will, but not enough to slow it down
> past light speed.
Your post is inaccurate. When you push on one end of an object, the other end does not immediately start moving. You just produce a wave of compression from one end to the other (kind of like how tapping one end of the object only produces vibration at the other end at the speed of sound, which is certainly not infinite), and the speed of this wave depends on the rigidity of the object. For instance, if you push on one end of a rubber pole, it will take a lot longer from the other end to likewise move than if the pole were made of steel. But there is no object quite rigid enough, even in theory, that does this faster than the speed of light.
Here's a place that explains it better than I can:
http://www.vscht.cz/mat/Pavel.Pokorny/physics/F
"If you have a long rigid stick and you hit one end, wouldn't the other end have to move immediately? Would this not provide a means of FTL communication?
Well it would if there were such things as perfectly rigid bodies. In practice the effect of hitting one end of the stick propagates along it at the speed of sound in the material which depends on its elasticity and density. Relativity places an absolute limit on material rigidity so that the speed of sound in the material will not be greater than c."
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> The 6 GHz is a little fishy to me, and here's why:
:-P
> 6 GHz --> 0.17 ns per cycle. Light travels 5 cm (about two inches) in 0.17 ns, and information
> cannot travel faster than light. This means that even at the speed of light (electrical signals in
> typical electronics propogate at ~0.8 c, IIRC) it will take almost the entire clock cycle to get
> information across the chip, never mind whatever time it takes the transistors to respond.
Pipelining is a well-understood technique that was introduced to the x86 world with the 80486 processor (that's the one that was a generation before the Pentium for you new folk). The idea is that each stage of the pipeline acts as a dedicated, specialized processor with limited functionality that hands off its results to the next stage. It is analogous to the Assembly Line, where each worker has a specialized task and hands off each in-progress product to the next worker in the chain.
The key here is that the electrons only have to pass through *each stage* in a single cycle. If your cpu is 4cm across, then the electrical signals (according to your number) would take 0.17ns to cross it. But if the cpu were separated into ten stages, then the signals would only need to traverse 0.4cm during each cycle.
Naturally, the tradeoff is that when you increase the number of stages, then the number of cycles that each instruction needs to complete increases, so you get penalties from erroneous predictions and cache misses and the like.
So your 6GHz limit only applies if your cpu is a one stage processor. Most consumer desktop processors have ten to fifteen stages. The Pentium 4, depending on how you count it, has as many as twenty-eight stages.
> In the meantime, those nursing dreams of 100 GHz chips had better look beyond nanotech to
> picotech-- atom-sized transistors.
I don't think that I disagree here, despite my above comments. To do these frequencies without a dramatic decrease in transistor sizes would require an absolutely obscene level of pipelining, to the point that performance would take massive hits and operating temperatures would be quite Venusian.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> Jobs are like relationships. It's always easier
:P
> to get a new job offer when you've already got
> an existing job -- even a crappy one.
Bullshit. The only time that I ever got jobs offered to me was back when I ran a tech news site that had ten thousand visitors a day, back when I was known enough that people like Anand (from anandtech.com) and Johan (from Aceshardware.com) would pop messages to me just for a friendly chat. Back then, when I was actually sort of *popular*, I'd occasionall get a job offer.
Now, there's *nothing*. I'd send resume after resume for months before getting so much as an interview. My resume has been looked at many times by both professional and amateur resume specialists, so I can only assume that it's at least *decent*. My standards are very low, as after a decade and a half of work experience (and nearly a decade of computer work experience) I'm certainly looking for no more than what is considered average in my area for a recent graduate (I graduated many year ago) -- that being $30K plus benefits in Long Island, a fish-shaped bit of land near to but much larger than Manhattan which "enjoys" a high cost of living. I try to hold back on my luxury purchases. I have broadband internet, a cell phone and my still existing domain being hosted, and I have to pay rent to my family for housing and feeding me. My weekends are a little expensive, as I go to the movies with my friends every week, but I managed to get a wad of lower cost bulk movie tickets that lower that cost. In the end, though, I'm a computer specialist who hesitates to purchase a computer to replace his ailing machine, a Duron-800 that can no longer work reliably in window managed modes, so it's basically running in frame buffer linux. If it weren't for giFT, fbxine, fbi, nethack, cone, links and fbtv, I'd go insane.
Hmmm. Looking over the message preview, I realize that I had a point, and it was this: 90% of decent jobs are acquired because you know somebody who can help you fill the position (whether it be a family member or a friend or whatnot). I know this because of the purely unstatistical fact that the least aggressive person in my three man clique got a $53K/yr job for the FAA because his mother worked there, whereas the other two are merely hard workers (one of them quite possibly exceptionally talented) who make under $30K/yr.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> #1 - fresh out of school, learned a bit of c++ and java in school, knows all the academic stuff;
===} 30-38K
Haha. You must not live in the suburbs of New York City.
BS in CompSci, cross-platform C++, Qt, Perl, CGI, HTML, javascript, css, FreeBSD, Linux, Apache, Windows, Qmail, Vpopmail, courier-imap, WAP gateways, helpdesk, MySQL, Nessus, client PC maintenance, server troubleshooting, cat5 crimping, network engineering, Netware 4.x/5.x, remote management tools, web design/development, online journalism, 16 years of general work experience, 9 years of computer-related work experience
===) $27K
It's all who you know, apparently. I know nobody of import.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
Wow, we could not possibly have more differing experiences here! My 5.1 sound card has a CM8378, too, but it was Windows that gave me the problems. An earlier version of Mandrake autodetected it and everything went perfectly. It took me *weeks* of fiddling around before I could get the 8378 working in Windows 2000. In the end, I had to boot into Linux to get information about the sound card (including the ID of the chip) before I was able to find drivers that worked in Win2k.
I also have continuing problems getting my Hauppauge TV card working in Windows 2000. And when I swapped my CD-R drive for a DVD+R drive, Windows 2000 failed the hardware detect and refused to reboot (an error with something like "redbook.sys") until I reinstalled.
Then again, once when I moved the location of my CD-R, Mandrake failed to figure out where it moved to, and I had to learn how the "/etc/fstab" file worked. And I can't get Mandrake to work with my Lexmark X83 multifunction printer, largely because Lexmark isn't very active with Linux driver development. And my gyroscopic mouse required me to add a couple lines into two text files before it worked well.
It works both ways, see? I try to avoid complaining about it in either direction, but I do get annoyed when people seem to think that problems only occur on one side of the bridge.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> Now, if it happens to be one of the applications
./configure says "missing dependencies foo, bar, baz and ook", you type "urpmi foo bar baz ook" as root, then you go back and do ./configure again. It's not the sort of thing that a normal user should have to go through (which is why somebody needs to put qtorrent into Mandrake's rpm tree!), but for folks like you, who know enough to do far too difficult dependency hunting, this should be a breeze.
> bundled with Mandrake I can just use the
> software installer
> Other times (This is usually where I give up),
> the computer starts acting as if it's on crack:
> rpm -i annoying-dependency.rpm
I mean no disrespect, but why are you using "rpm -i" to install a downloaded RPM file if you're in Mandrake? If you're in Mandrake, you can use the software installer, just as if you were installing a regular RPM from the CD or from Mandrake's sources. In command line speak, you do:
"urpmi thisprogramthatidownloaded.rpm"
And it'll do all that automatic dependency fixing.
If you install from source, then it's a little difficult, but not too much. For instance, I was installing QTorrent recently, which is an excellent bittorrent client that uses fewer resources than the standard btdownloadfoo.py and puts everything together in a nicely manageable gui. I unpacked the source tarball and saw that it had the "setup.py" (with these -- and it took me far too long to figure this out, you have to do "python setup.py install"). When I tried to install it, it bitched about there being no "PyQt". So I did a "urpmi PyQt", and it autodownloaded and autoinstalled that particular dependancy. Yay, QTorrent could then install.
It's like that. If
End communication.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> I seriously wonder why nobody has implemented
.tar.gz file with all the files in the proper relative paths (./bin/, ./lib/, ./share/progname/, ./etc/progname/). The only real stumbling block is that the installer has to look at the OS and find out which variant of linux it is (that is, whether it uses debs or rpms). The installer would just have to add entries to the specific distro's "installed files" system settings so that the distro's normal uninstaller can handle it.
... well, whatever else uses .deb files, then it'd write to the installed DEBs database. I know that Mandrake has a tool to uninstall programs graphically, and the other systems likely do, as well. I mean, heck, if I applied myself for a few days (and I didn't have one of my bizarre depression attacks), *I* could probably write something like this, and I am most certainly not very experienced with programming.
;P
.NET and Cocoa--none of this absolutely
> binary installation/uninstallation routines for
> the Linux desktops yet. What's the damn holdup?
> Users need to be able to buy a Linux application
> from a store, take it home, and stick in a CD to
> get an autoplay installer.
Firstly, I object to your statement about binary installation/uninstallation not existing. That's the whole point of urpmi and apt-get (or their gui counterparts, like Mandrake Install and whatever it is for apt).
But you do have a bit of a point about installs from CDs. Actually, it wouldn't be *too* difficult. I mean, you can make a binary install by just making a
Basically, there would need to be two branches: If the OS is Red hat, Fedora, Mandrake or SuSe, the installer adds to the installed RPMs database. If the OS is Debian or
Oh, and Gentoo doesn't count, because those people don't believe in using non-OSS.
> Of course, to get that truly working well, you'd
> want a sane, robust programming library in the
> likes of
> ridiculous QT/GTK/wxWindows/whatever nonsense
> that are merely hacks to get widgets up on X.
You're silly. Almost all the programs that I run in Linux and FreeBSD use a single widget library: Qt. In windows, nearly every program I run has a different widget set. There's Opera, which uses Qt. There's Office, which uses the Office widgets (*NOT* the same as the MS Windows widgets!). There's the built-in windows programs. There's Eudora, which seems to use its own thing. There's Mozilla, which uses its own bizarre stuff. And Ultra VNC seems pretty nonstandard, too. I realize that, for most users, the majority of programs seem to be written with the same widgets, but a large part of that reason is that most users predominantly run programs written by one company -- Microsoft. Of *course* Microsoft apps will look like other Microsoft apps. That's just like how KDE apps look like other KDE apps (and I dare say that KDE is even a little more consistant with this, given that KOffice and other KDE apps use the same widgets).
Bah, finished ranting. Maybe I'll make some kind of generic installer for Linux now....
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
PS: Your comment about older RPMs are interesting. I may try this out.
> As it is now - Real offers the ONLY player that
> incorporates all major music formats WMA, MPEG,
> ATRAC3, with exception for Apple's AAC.
I may be incorrect, but I think that there are more than one such player. MPlayer, an open-source video player that runs on Windows, Linux and other platforms, plays everything in the universe, including the formats you mentioned above. Xine is a similar effort; I don't know if it works in Windows, but it probably has port either here or coming eventually, and it *probably* supports all those codecs. Meanwhile, Media Player Classic, a Windows-only project that uses the superior (well, to WMP7 or WMP8, at least) interface of Windows Media Player 6 but combines it with a load of codecs, may also support all these formats.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> I certainly did not notice that CAN-SPAM became effective 1/1/04. Or actually,
> my filters are still filtering out a very similar number of messages.
That's actually a really good point. Some time late last year, I was getting about 1200 emails a day, nearly all of it spam. It absolutely inundated me and nearly rendered me helpless. I learned how to code mail filtering programs in perl, then I learned how to use stuff like procmail and spamassassin and clamav, and when that stuff is used in conjunction with a mail client with bayesish filtering capabilities (like Mozilla), spam becomes manageable.
But I checked my procmail logs a couple weeks ago, and I discovered that the number of daily emails I'm getting (again, nearly all of them spam) have jumped to over 2000.
I guess the law really didn't do anything after all.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> I dunno, I'm as big a South Park fan as anybody
> but season 8 has been consistently bad so far.
That's odd. I only found the 'Passion'-related episode to be weak. The other episodes so far have been extremely funny, moreso than the last two seasons have been.
Of course, maybe I'm just using South Park to compensate for my continuing battle against Futurama-oriented withdrawal symptoms.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> Earth to google
Um, Every single graphical desktop web browser in the universe, save for one (yeah, THAT one), supports keyword features that make it completely pointless for individual websites to bother. In Opera, since last millennium, you could type "g bunch of search terms". Mozilla could be easily configured to use the same syntax, or you could change the keyword from "g" to something else (like "IWannaKnowMoreAbout bunch of search terms"). Konqueror does it like "gg:bunch of search terms" (and, I think, "g:bunch of search terms" to do the "I feel lucky" thing on Google).
I have Opera set up with a few custom keywords. If I type "def someword", then the browser does a search on dictionary.com for "someword". Useful feature. Totally pointless for the feature to be built into the website. Unless you're using a web browser with a 1990s feature set, like lynx or IE.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> Can somebody tell me why it seems everything
> requiring a cross-platform C++ gui widgets
> seems to be written with QT?
It's largely because Qt is really, really insanely easy to use. The object classes are very intuitive for programming.
Regarding licensing issues, there is a GPL version for native Win32. It's not being actively updated, but it did get far enough to the point where most of my programs (the ones that didn't use other libraries, at least) would cross compile and run pretty decently. There were some speed issues, but most of them disappeared after I figured out how to get the compiled programs running without popping up a stderr command window.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> The only "innovation" KDE and GNOME can profess
> is having tons more pointless applets running on
> their panels, and running a lot slower than
> Windows on the same hardware.
KDE has session management to a level that I've never seen elsewhere. It can set the Z-level of any window to be above of below other windows. You can turn off the titlebar of any window if you need it to disappear. You can fullscreen any maximizable application. Widgets across all KDE applications are instantly, globally skinnable. I can run KDE on Win32, Linux and FreeBSD and the installs are easy compared to any other complex desktop environment. Mouse gestures and shortcut key combos can be easily set for global and application-specific options. The KDE file dialog box is unparalleled -- it has its own bookmark feature, but it also has collapsible and directly editable alternative to the horrifyingly static and bothersome "Places" bar in Win32, and you can type in new directory locations and thusly change directories without accidentally erasing the file name in the dialog box, which is a major failing of Win32 and most other file dialogs. Additionally, the file dialogs have relatively protocol-agnostic access of network-accessible files. You can load and save stuff that are shared via SMB, NFS, FTP, even SSH, as if they were saved on your local disk. Heck, you can browse IMAP as if it were a file sytem! Titlebar buttons can be easily shifted around and customized to avoid the horribly fatal "kill button next to maximize button" flaw in most operating systems and environments. Devices can be set to appear on the desktop when they mount. There's a useful graphical disk usage view included in the default file manager.
And, yeah, there are a ton of incredibly useful panel applets.
> Not to mention using just ONE library instead
> of multiple "toolkits" doing the widget work
> that should be done by the desktop anyway.
Win32 has this problem, too. Pretty much every single program that I'm using right now on this Windows 2000 install (Opera, Mozilla Mail, cmd.exe, Cygwin/X, vncviewer, OpenOffice.org) has different widgets. What's the big deal if Linux does the same thing? At least KDE has more consistency than, say, Microsoft (whose most popular program -- MS Office -- uses widgets different from the operating system).
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> Windows just works, and works with more
> hardware without extraconfiguration.
That's not true for everybody. I had to turn ten kinds of hell to get my 5.1 audio card to work in Windows. In the end, I had to boot into Linux (which had detected it properly) to find out what the audio card's chipset was, which finally gave me enough clues to get drivers for the card.
Yes, though, Windows has ubiquitous driver support.
> Ummm rootkits?
I've never been rootkitted, and I'm very horrible with security.
> When more people switch over to
> linux you will see more viruses out there for
> linux because right now windows is an easy
> targetwith a ton of machines out there.
And how will these viruses spread? Unlike Windows, the dominant email apps for Linux do not run applications when you click on them. Unlike Windows, viruses on Linux can't take control of system files. Unlike Windows, Linux computers start up with unnecessary ports turned off.
> People much rather write a virus that will
> effect a much larger population.
I don't think you get it. Why do you think that there are multiple means of package management? Why do you think that different distributions handle things differently? Why do you think that Linux advocates and Open Source Software programmers make such great strides towards making sure that applications are available not only on Linux but on other operating systems such as FreeBSD, OS X and QNX? It's because having a monoculture is *BAD*. We don't *want* every computer in the world to run Linux. That would be *stupid*, even though Linux has a far safer security model than Windows. We want operating system usage to be distributed more or less evenly among different models, just like it was in the old days, when viruses *weren't* dangerous, and when a stupid move by one OS maker didn't negatively affect every single computer user on the planet!
> If linux was secure from viruss,why are there
> linux virus scanners?
They're to scan for *Windows* viruses. You know, like if you had a mail server on your Linux box? I do, and I filter mail through clamav in order to prevent infection on a Windows box if I chance to check my mail on one.
> with linux you need to figure out why your
> soundisnt working and then configure the driver
> and what not and edit configfiles.
You mean, "with linux you have at least one way of getting sound working if your sound card doesn't work". This is opposed to "with windows if your source card doesn't work you can only cry a lot and try to reboot or just buy a new sound card".
Linux isn't perfect, and I certainly hope that no single operating system ever captures more than a quarter of the desktop or business market, because that would lead to disaster. But Linux has saved me from a ton of frustration that Windows had caused me, and it's a lot easier to use (especially installing software, which Linux reduced to a simple, single step from the more popular many-step process.
--
-JC
> Yesterday I received 144 pieces of Spam ... If
> it wasn't for SpamBayes, I probably would have
> abandoned email altogether by now. These guys
> are rubish.
Ho ho! You should consider yourself lucky, my good fellow. According to my procmail.log file, I'm averaging upwards of two thousand instances of spam mail per day. You can't *possibly* imagine the trauma that this sort of thing can cause.
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
> The bus, the subway, the train, the bike, ...but that's crazy talk.
;P
> and walking.
It *is* crazy talk, and here's why: It's a forty minute drive to work for me, and a fifty minute drive back (heavier traffic in the afternoon). This is annoying enough, but if I took a train, it'd be:
A) Twenty minute walk to the train station (I'm not trusting enough to tie up a bike for ten hours a day)
B) Ten minute train in the wrong direction to get to a central junction
C) Five minutes waiting for the switchover train
D) Thirty minute train to get to the nearest station
E) Thirty minute walk, because the nearest station isn't near to my work at all (neither are any convenient buses).
Yay. That's ninety-five minutes, or about twice as long as my regular commute. Even with the price of gas, the train ends up being more expensive than my car. There's also the fact that I'd have to spend nearly an hour walking -- *each way*! So we get an hour less sleep each day and almost two hours of additional walking each day, resulting in a total zombie worker who can't code or set up BSD servers properly anymore.
Damnit, I hate my job with a nigh impossible passion, and I'm no fan of driving cars, but public transportation would be a worse case scenario (eg, I'd go for unemployment first).
--
-JC
coder
http://www.jc-news.com/parse.cgi?coding/main
PS: And I don't use my car on days that I don't work, so you can't get me there, either!