There are two very different cases, which tends to lead to people talking past each other. Basically, database portability is important to software companies (who don't want to restrict themselves to selling to Oracle shops), but not to companies working on in-house systems.
Mod parent up. This is perhaps the most insightful comment in this entire sub-thread.
FWIW, I think these days in-house systems (which would include hosted systems with a web front end) are a larger segment of the market. Maybe embedded is bigger, but you're not going to use Oracle in an embedded environment. Well, not usually, anyway....
Stored Procedures can be bad, but they also can be very good. I regard SPs as somewhat like inline assembler in C/C++ code -- it can give you *massive* performance boosts. Of course, there are times where the costs (lack of portability, requires more knowledge of relational theory, etc.) outweigh the benefits, but those are mostly for small databases. When you use 10M+ or even 1B+ row databases, doing table scans (which could take hours or even days) for a simple select join isn't going to cut it, and you are really going to need the optimizations you can get from writing the SQL yourself. And if you're going to be writing database-specific SQL, you might as well store it in the database.
Plus, big databases like Oracle compile SPs and save the execution plan so the SQL text doesn't even have to be interpreted again. Doing all your SQL dynamically to keep database independence is like saying you should write your hardware drivers in JavaScript so you can have platform independence. Plenty of things can be written in JavaScript, but not everything. Use the right tool for the job.
Finally, a properly built database will scale very nicely, thank you. People have problems with overtaxing their database servers when all they're doing are giant SELECT * FROM X queries where the database server has to schlep all the data across to the application server before you can do anything with it. Moving the data across the network costs way more than the actual join/query syntax will, especially if you happen to actually understand the relational data model...
I'm right-handed and I have had my mouse on the left for several years now. At this point, I find it far more comfortable and flexible to use -- I even prefer doing Photoshop-like fine detail work with my left hand now.
Actually, at a recent occupational safety meeting at work about RSI, I found that a large percentage (10%-25% IIRC) of my coworkers have done the same thing.
I agree (sort of) -- most of the people who are die-hard Opera fans switched years ago. Don't get me wrong, Opera is a good browser now, but its "edge" is far smaller now than it was in the days of IE 5 and Mozilla 0.9.
Actually, my browser migration path goes:
- Netscape 3 Gold
- Netscape 4
- IE 5
(switch to Linux as my primary desktop system)
- Mozilla 0.6
- Opera 5
- Galeon
- Mozilla 1.0
(switch back to Windows for most everyday tasks)
- Firefox 1.0
Actually, looking back on that, that's a pretty crazy path....
Anyhow, the point I was trying to make (in a very roundabout way) is that each time I've switched browsers, there's been a "killer app" feature that made me do it. For the Opera to Galeon switch, the "killer app" features were the GNOME integration and the better standards-compliance of the Gecko engine at the time.
Now, the #1 thing that keeps me with Firefox as my primary browser is the Web Developer Toolbar. I can't imagine how anyone who works with web technologies could live without it. However, as someone who works with web tech all day long, having a multiplicity of browsers available is definitely a Good Thing...
Hate the RIAA? Don't like their business model and practices? Then don't do business with them! If you're a music consumer, go to clubs, buy microlabel and band-pressing CDs, swap non-RIAA mp3s, and enjoy. If you're a music producer, release your tunes through some other channel.
That's all well and good, but I'm sure I'm like most people when I say that my dislike of the RIAA's business practices is not enough to change my taste in music.
I'm not aware of any club-level/microlabel/indie bands that come anywhere close to (for example) Liquid Tension Experiment, Racer X, or the Mahavishnu Orchestra. These are all virtuoso-level groups in terms of outright skill. I enjoy their music -- so much so, in fact, that I am willing to pay money, and have paid money, for their CDs.
The parent post deserves an upmod (and the grandparent post a downmod).
Just to clarify: The "an" in the title at the Merc is intentional -- the writer is trying to make a cute play on words. In journalist slang, a person who provides information is a "source" and a source of information that is freely accessible is "open". To the news writer, then, Linus Torvalds is an "open source" about Open Source software.
It doesn't take a rocket scientist to understand English. Its grammar and syntax is not that much more complex than the grammar and syntax of some computer languages, but in most cases it does allow for a great deal more flexiblity in phrasing.
Next, we want to execute this script on startup. To do that, we have to place a link to it in the/etc/rc5.d sub-directory, which is where scripts get executed on startup in RunLevel 5, which is a full multi-user run mode with networking and X-Windows. To create this link, bring up a window in Gnome, and go to the/etc/init.d sub-directory (or wherever you've placed the bash script). Right-click on the file, and drag it onto the desktop. You'll be asked if you want to move, copy, or link to this script. Select "link here."
Now, in that same window, change to the/etc/rc5.d sub-directory, and drag the script link from the desktop into this sub-directory.
Next, you need to rename the script using the following convention-- Look at all the scripts that begin with the letter 'S' (these are services scripts, which get executed after kernel scripts, which begin with the letter 'K'). Find the last script that begins with the letter 'S', which is usually S99local. Verify that this is the highest-numbered script beginning with 'S'. If yes, rename the link to your bash script to:
S100lvm
The script will now execute on startup, and the needed LKMs will be available to Freevo.
[Stupid/. formatting...]
OK, that doesn't jive with what I learned in RHCE class. On a RedHat system, init processes the rc?.d scripts in asciibetical order (so S100lvm would actually come before S99local), passing "start" as the first argument to those scripts that start with S and "stop" to those scripts that start with K. The idea is that your init.d script should accept start and stop arguments and perform accordingly. Once you drop this script into/etc/init.d, you should be able to use chkconfig to set up the proper symlinks in the/etc/rc?.d directories. (A simple "chkconfig lvm on" should put "K" symlinks into rc0.d, rc1.d, rc2.d, and rc6.d, and "S" symlinks in rc3.d, rc4.d, and rc5.d.) Plus, you can use "service lvm start" to avoid typing "/etc/init.d/lvm start" (if that's your thing).
Aside: Also, lvm is a bad acronym for "Load Video Modules", since LVM is the Logical Volume Manager, and RedHat includes support for LVM out-of-box.
Back when I used a gui file manager, I found GMC (the Gnome Midnight Commander) to be fast and intuitive for a Windows user. Of course, then I realized that Eterm + bash is considerably more efficient and useful than any gui could ever be. Even when I'm forced to use Windows (which, thankfully, isn't that often anymore), I find I can manipulate the filesystem faster in a DOS box than in Windows Explorer.
OK, here's my question: I don't run dev kernels on my main box here, so how long can I expect to wait before this patch makes it into a stable kernel? I expect there's no chance it will be backported to 2.4, so the real question is how long until 2.6 (or 3.0?) comes out? I heard "early 2003" several months ago. Is that still a reasonable timeframe?
I've been hearing a lot about WebDAV lately, and the more I learn about it, the more I think it's the perfect solution for remote, crossplatform file sharing. I've been doing some work with a decentralized team, each of us working from our home office, doing technical writing and PR services. We need a remote fileserver for storing common files, especially documents in progress. Something like iDisk would be nice, but none of us are on a mac right now (it's currently a mix of Linux and WindowsXP). I'd run my own, except I'm working from home, and my ISP doesn't like me running servers from my system. I have an account with a great hosting provider, but they have FrontPage extensions installed on their servers, and FP somehow conflicts with mod_dav. My question: Is anyone aware of a commercial DAV service? Even a decent hosting service that would be willing to run mod_dav would be great.
Second, has anyone actually got davfs to work? I've tried both davfs and davfs2 (on a fairly vanilla RedHat 8.0 box), and neither one works. davfs1 gives me mount:no such device errors, while davfs2 says something about an invalid mount point. Any ideas?
IMO, the phrase in the definition that Google's lawyers are taking issue with is "such as": "google: v. To use an Internet search engine such as google.com to look for information." That's what the letter means when it says: "This definition
implies that "google" is a verb synonymous with "search.""
The implication is that I would say, "Hey, I googled X" when I had in fact used some other search engine. AFAIK, this is not a common use. In part because of the widespread popularity of Google, the search engine, when people use "google" as a verb, they always mean Google (at least in my experience). If someone starts using "google" as a verb to simply mean "an Internet search engine" then Google will, naturally, show a legal interest. As the letter says: "We want to make sure
that when people use "Google," they are referring to the services our
company provides and not to Internet searching in general."
If the definition read "google: v. To use the Internet search engine google.com to...", then I doubt McFedries would have received the letter.
IIRC, the "Limited User" option in the WinXP wizard sets the user to be a member of the Users group. What you probably want is the Power Users group (you'll probably have to set this manually). Power Users can install and run most anything, but they have limited access to system-level resources.
The NT security system is fairly granular (it pains me to admit it, but it's actually more flexible than most Linux distributions out of the box), but it's often non-obvious what settings you need to get a particular application to work. Specificially, anything written for the Win9x series assumes it has full control of the machine by default.
It's not impossible to lock down Windows (NT/2k/XP -- it is impossible to lock down 9x/ME), but it takes some time and effort. It's well worth it, though, if you want a secure, stable system.
Or, of course, you can just switch to Linux. These days I think it takes about the same amount of time to get either Windows or Linux set up and running right. The only reason Windows seems easier is that most people throw it on and don't take the time to set it up tightly.
I have no problem with digital recording. You're right, it's possible, with enough bits, to create a (near-)flawless digital signal. Digitally recorded music does tend to have far less noise and distortion, especially after multiple generations.
But that's not what we're talking about here.
As a guitarist, I am playing an instrument. The purpose of the instrument is to translate my musical imagination into sound. Therefore, everything along my signal chain, from the pickups in my guitar, to my preamp, to my effects, to my amp's power section, to my speakers, are all really part of my instrucment, since they all affect the tone of my instrument. Yes, it's true that analog devices create more distortion than digital. And, shockingly enough, as part of my instrument, that analog distortion is exactly what I want! Most modern electric guitarists use a lot of overdrive or distortion, and tubes create a fuller, warmer distortion than transistors can -- especially at low gain (Stevie Ray Vaughan) or super-saturated high gain (Joe Satriani) levels.
Now, a lot of work has been done lately in emulating the waveform of a vacuum tube at high output, but I've yet to hear anything that can exactly duplicate the sound. I've got a digital preamp here that is pretty good, and to an untrained ear might sound the same, but there is a difference. I don't mind the digital, but, given the choice, I'll take a tube amp over a digital one, and I think the vast majority of modern guitarists would agree.
Even Marshall has been making AVT amps with solid-state power sections, but the preamp is still a 12AX7 vaccuum tube. It just wouldn't quite sound the same without it. (Note: I've never played an AVT. They may sound awful. I certainly doubt they sound as good as a JCM2000.)
It reminds me of a clip I once saw of Pete Townsend. He said, roughly, "Buy yourself a nice guitar, like a Les Paul. Plug it into a big Marshall stack. And hit a chord. And then you'll understand."
I think you have some good, valid points. However, some of the "problems" you've mentioned are already solved, or are in the process of being solved.
Move Away From X-Windows
I continue to waffle on this one. Yes, X (BTW, it is X, X11, or the X Window System, never XWindows or X-Windows) has a lot of legacy code & features, and its network-centric model does affect its performance, but I think a lot of people are making a much bigger fuss over X than it really deserves. The XFree86 people have been doing an incredible amount of work lately, and things like fonts are finally getting fixed, while the performance issues were mostly fixed long ago. At this point, I'm not aware of any viable replacement for the X11 protocol (things like fb still need an X11 emulator for compatibility, since 99.999% of all *nix apps use X), so doesn't seem logical to throw X away -- and I haven't even begun to talk about the usefulness of the network transparency features...
The Adoption Of A Single, Standardized Interface Design
I agree that this is important if Linux is to really succeed on the desktop. As much as I like having GTK and Qt apps side by side in Fluxbox, the average user would be completely bewildered by my desktop, and I don't begrudge average users their WIMP interfaces. However, the solution is not to "kill" either KDE or Gnome (fat chance of that happening, anyway) but to do what Red Hat has done with 8.0: re-theme KDE and Gnome with common widget sets, so that even if they're different under the hood, the apps look and feel the same. And it works... although I don't use it on my box, I think Bluecurve is a great idea for Joe User.
Make Graphical Setup "Wizards" For Everything
At this point, the only time you have to touch config files on a Red Hat system is if you're going to be using some obscure settings for some server. Really. Red Hat has created an amazing array of GUI config tools for every administration task under the sun, and they're remarkably easy to use. Need to share files with Windows boxes? Piece of cake -- configure Samba with SWAT. Don't like XF86Config? Use redhat-config-xfree86. Can't figure out named.conf? You can use redhat-config-bind for that! The advantage of doing it this way is that the config files are still there for people like me who actually find using them to be more efficient.
Binary Distributions For Everything
Done. RPM, DEB, heck, even Gentoo has support for binary packages! These days practically every app vendor supplies RPMs. Again, the only time you're going to have to compile anything from source is if you're using some really obscure app. Even then, it's not exactly hard:./configure; make; make install. Someone who's taken the time to search out that rare program that they just can't live without will be patient enough to learn three simple commands.
Workstation Configurations With Dangerous Deamons (ftpd, httpd, etc...) Turned Off By Default
Done. I don't know about all the other distributions, but Red Hat doesn't install or turn on any servers by default, especially if you choose the "Personal Desktop" or "Workstation" installs. With more advanced distributions, like my favorite, Gentoo, you generally have to know what you're doing in order to turn on a dangerous daemon, so you're unlikely to do it by mistake.
Linux Evangelists Stop Insulting MS And Its Users
But they make it so easy! Just kidding. I agree with you here... as long as the open source community is seen as a bunch of squabbiling stoners and teenagers, it's going to have trouble being seen as a serious alternative. However, I think a lot of progress has been made and continues to be made, especially in the public eye. Of course, you can get a very different impression if you read Slashdot. Thank goodness Joe CEO doesn't...
My little brother is (informally diagnosed) ADD, possibly ADHD, and after reading a couple of chapters of Driven to Distraction I'm convinced that I'm at least borderline ADD.
I've always felt, however, that my distractability was not a negative trait -- in fact, with some effort, I've been able to train myself so I can, for instance, participate in two conversations at once, or some similar multi-input task. In fact, I'm convinced that this is the flip side of what the Jargon File calls hack mode, "a Zen-like state of total focus on The Problem" to the exclusion of everything else. This is an ability that I greatly value, even though it's sometimes not fun to finish a project and then suddenly realize that I've forgotten to eat, sleep, or use the bathroom for way too long.
My question, then is: is this a trait (or a mental ability) that is related somehow to the "hacker" mentality? Comments above have indicated that ADD children are often very intelligent and often have logical/mathematical abstract thinking patterns -- both of which are characteristic of hackers. I'd love to see some statistics (if there are any) on the relationship there.
How did you create the partitions? If the partition type flags aren't set to 83 (Linux), then Windows might think that there's nothing on them. (Linux mount can sometimes do the right thing even though the partition flag is wrong, but it looks like Windows isn't that intelligent.) Also, Win2k (IIRC) has a tool in the drive manager that allows you to hide partitions completely... that should stop the problems you're having with Explorer...
Slashdot Mirror
Unless you have a sign bit on your 32-bit word. 2^31 bits = 2GB.
Yes, this does happen.
Well... I hate to burst your bubble, but 13.37 times pi is actually 42.0030937784954923... ad infinitum.
The number you want is probably closer to 13.369015219719221985830700904996......
Funny how this argument always comes up....
Stored Procedures can be bad, but they also can be very good. I regard SPs as somewhat like inline assembler in C/C++ code -- it can give you *massive* performance boosts. Of course, there are times where the costs (lack of portability, requires more knowledge of relational theory, etc.) outweigh the benefits, but those are mostly for small databases. When you use 10M+ or even 1B+ row databases, doing table scans (which could take hours or even days) for a simple select join isn't going to cut it, and you are really going to need the optimizations you can get from writing the SQL yourself. And if you're going to be writing database-specific SQL, you might as well store it in the database.
Plus, big databases like Oracle compile SPs and save the execution plan so the SQL text doesn't even have to be interpreted again. Doing all your SQL dynamically to keep database independence is like saying you should write your hardware drivers in JavaScript so you can have platform independence. Plenty of things can be written in JavaScript, but not everything. Use the right tool for the job.
Finally, a properly built database will scale very nicely, thank you. People have problems with overtaxing their database servers when all they're doing are giant SELECT * FROM X queries where the database server has to schlep all the data across to the application server before you can do anything with it. Moving the data across the network costs way more than the actual join/query syntax will, especially if you happen to actually understand the relational data model...
Just another "me, too!" post...
I'm right-handed and I have had my mouse on the left for several years now. At this point, I find it far more comfortable and flexible to use -- I even prefer doing Photoshop-like fine detail work with my left hand now.
Actually, at a recent occupational safety meeting at work about RSI, I found that a large percentage (10%-25% IIRC) of my coworkers have done the same thing.
I agree (sort of) -- most of the people who are die-hard Opera fans switched years ago. Don't get me wrong, Opera is a good browser now, but its "edge" is far smaller now than it was in the days of IE 5 and Mozilla 0.9. Actually, my browser migration path goes: - Netscape 3 Gold - Netscape 4 - IE 5 (switch to Linux as my primary desktop system) - Mozilla 0.6 - Opera 5 - Galeon - Mozilla 1.0 (switch back to Windows for most everyday tasks) - Firefox 1.0 Actually, looking back on that, that's a pretty crazy path.... Anyhow, the point I was trying to make (in a very roundabout way) is that each time I've switched browsers, there's been a "killer app" feature that made me do it. For the Opera to Galeon switch, the "killer app" features were the GNOME integration and the better standards-compliance of the Gecko engine at the time. Now, the #1 thing that keeps me with Firefox as my primary browser is the Web Developer Toolbar. I can't imagine how anyone who works with web technologies could live without it. However, as someone who works with web tech all day long, having a multiplicity of browsers available is definitely a Good Thing...
I'm not aware of any club-level/microlabel/indie bands that come anywhere close to (for example) Liquid Tension Experiment, Racer X, or the Mahavishnu Orchestra. These are all virtuoso-level groups in terms of outright skill. I enjoy their music -- so much so, in fact, that I am willing to pay money, and have paid money, for their CDs.
Just to clarify: The "an" in the title at the Merc is intentional -- the writer is trying to make a cute play on words. In journalist slang, a person who provides information is a "source" and a source of information that is freely accessible is "open". To the news writer, then, Linus Torvalds is an "open source" about Open Source software.
It doesn't take a rocket scientist to understand English. Its grammar and syntax is not that much more complex than the grammar and syntax of some computer languages, but in most cases it does allow for a great deal more flexiblity in phrasing.
OK, that doesn't jive with what I learned in RHCE class. On a RedHat system, init processes the rc?.d scripts in asciibetical order (so S100lvm would actually come before S99local), passing "start" as the first argument to those scripts that start with S and "stop" to those scripts that start with K. The idea is that your init.d script should accept start and stop arguments and perform accordingly. Once you drop this script into /etc/init.d, you should be able to use chkconfig to set up the proper symlinks in the /etc/rc?.d directories. (A simple "chkconfig lvm on" should put "K" symlinks into rc0.d, rc1.d, rc2.d, and rc6.d, and "S" symlinks in rc3.d, rc4.d, and rc5.d.) Plus, you can use "service lvm start" to avoid typing "/etc/init.d/lvm start" (if that's your thing).
Aside: Also, lvm is a bad acronym for "Load Video Modules", since LVM is the Logical Volume Manager, and RedHat includes support for LVM out-of-box.
Back when I used a gui file manager, I found GMC (the Gnome Midnight Commander) to be fast and intuitive for a Windows user. Of course, then I realized that Eterm + bash is considerably more efficient and useful than any gui could ever be. Even when I'm forced to use Windows (which, thankfully, isn't that often anymore), I find I can manipulate the filesystem faster in a DOS box than in Windows Explorer.
OK, here's my question: I don't run dev kernels on my main box here, so how long can I expect to wait before this patch makes it into a stable kernel? I expect there's no chance it will be backported to 2.4, so the real question is how long until 2.6 (or 3.0?) comes out? I heard "early 2003" several months ago. Is that still a reasonable timeframe?
I've been hearing a lot about WebDAV lately, and the more I learn about it, the more I think it's the perfect solution for remote, crossplatform file sharing. I've been doing some work with a decentralized team, each of us working from our home office, doing technical writing and PR services. We need a remote fileserver for storing common files, especially documents in progress. Something like iDisk would be nice, but none of us are on a mac right now (it's currently a mix of Linux and WindowsXP). I'd run my own, except I'm working from home, and my ISP doesn't like me running servers from my system. I have an account with a great hosting provider, but they have FrontPage extensions installed on their servers, and FP somehow conflicts with mod_dav. My question: Is anyone aware of a commercial DAV service? Even a decent hosting service that would be willing to run mod_dav would be great.
Second, has anyone actually got davfs to work? I've tried both davfs and davfs2 (on a fairly vanilla RedHat 8.0 box), and neither one works. davfs1 gives me mount:no such device errors, while davfs2 says something about an invalid mount point. Any ideas?
Thanks a ton...
IMO, the phrase in the definition that Google's lawyers are taking issue with is "such as": "google: v. To use an Internet search engine such as google.com to look for information." That's what the letter means when it says: "This definition implies that "google" is a verb synonymous with "search.""
The implication is that I would say, "Hey, I googled X" when I had in fact used some other search engine. AFAIK, this is not a common use. In part because of the widespread popularity of Google, the search engine, when people use "google" as a verb, they always mean Google (at least in my experience). If someone starts using "google" as a verb to simply mean "an Internet search engine" then Google will, naturally, show a legal interest. As the letter says: "We want to make sure that when people use "Google," they are referring to the services our company provides and not to Internet searching in general."
If the definition read "google: v. To use the Internet search engine google.com to...", then I doubt McFedries would have received the letter.
The NT security system is fairly granular (it pains me to admit it, but it's actually more flexible than most Linux distributions out of the box), but it's often non-obvious what settings you need to get a particular application to work. Specificially, anything written for the Win9x series assumes it has full control of the machine by default.
It's not impossible to lock down Windows (NT/2k/XP -- it is impossible to lock down 9x/ME), but it takes some time and effort. It's well worth it, though, if you want a secure, stable system.
Or, of course, you can just switch to Linux. These days I think it takes about the same amount of time to get either Windows or Linux set up and running right. The only reason Windows seems easier is that most people throw it on and don't take the time to set it up tightly.
I have no problem with digital recording. You're right, it's possible, with enough bits, to create a (near-)flawless digital signal. Digitally recorded music does tend to have far less noise and distortion, especially after multiple generations.
But that's not what we're talking about here.
As a guitarist, I am playing an instrument. The purpose of the instrument is to translate my musical imagination into sound. Therefore, everything along my signal chain, from the pickups in my guitar, to my preamp, to my effects, to my amp's power section, to my speakers, are all really part of my instrucment, since they all affect the tone of my instrument. Yes, it's true that analog devices create more distortion than digital. And, shockingly enough, as part of my instrument, that analog distortion is exactly what I want! Most modern electric guitarists use a lot of overdrive or distortion, and tubes create a fuller, warmer distortion than transistors can -- especially at low gain (Stevie Ray Vaughan) or super-saturated high gain (Joe Satriani) levels.
Now, a lot of work has been done lately in emulating the waveform of a vacuum tube at high output, but I've yet to hear anything that can exactly duplicate the sound. I've got a digital preamp here that is pretty good, and to an untrained ear might sound the same, but there is a difference. I don't mind the digital, but, given the choice, I'll take a tube amp over a digital one, and I think the vast majority of modern guitarists would agree.
Even Marshall has been making AVT amps with solid-state power sections, but the preamp is still a 12AX7 vaccuum tube. It just wouldn't quite sound the same without it. (Note: I've never played an AVT. They may sound awful. I certainly doubt they sound as good as a JCM2000.)
It reminds me of a clip I once saw of Pete Townsend. He said, roughly, "Buy yourself a nice guitar, like a Les Paul. Plug it into a big Marshall stack. And hit a chord. And then you'll understand."
I continue to waffle on this one. Yes, X (BTW, it is X, X11, or the X Window System, never XWindows or X-Windows) has a lot of legacy code & features, and its network-centric model does affect its performance, but I think a lot of people are making a much bigger fuss over X than it really deserves. The XFree86 people have been doing an incredible amount of work lately, and things like fonts are finally getting fixed, while the performance issues were mostly fixed long ago. At this point, I'm not aware of any viable replacement for the X11 protocol (things like fb still need an X11 emulator for compatibility, since 99.999% of all *nix apps use X), so doesn't seem logical to throw X away -- and I haven't even begun to talk about the usefulness of the network transparency features...
I agree that this is important if Linux is to really succeed on the desktop. As much as I like having GTK and Qt apps side by side in Fluxbox, the average user would be completely bewildered by my desktop, and I don't begrudge average users their WIMP interfaces. However, the solution is not to "kill" either KDE or Gnome (fat chance of that happening, anyway) but to do what Red Hat has done with 8.0: re-theme KDE and Gnome with common widget sets, so that even if they're different under the hood, the apps look and feel the same. And it works... although I don't use it on my box, I think Bluecurve is a great idea for Joe User.
At this point, the only time you have to touch config files on a Red Hat system is if you're going to be using some obscure settings for some server. Really. Red Hat has created an amazing array of GUI config tools for every administration task under the sun, and they're remarkably easy to use. Need to share files with Windows boxes? Piece of cake -- configure Samba with SWAT. Don't like XF86Config? Use redhat-config-xfree86. Can't figure out named.conf? You can use redhat-config-bind for that! The advantage of doing it this way is that the config files are still there for people like me who actually find using them to be more efficient.
Done. RPM, DEB, heck, even Gentoo has support for binary packages! These days practically every app vendor supplies RPMs. Again, the only time you're going to have to compile anything from source is if you're using some really obscure app. Even then, it's not exactly hard: ./configure; make; make install. Someone who's taken the time to search out that rare program that they just can't live without will be patient enough to learn three simple commands.
Done. I don't know about all the other distributions, but Red Hat doesn't install or turn on any servers by default, especially if you choose the "Personal Desktop" or "Workstation" installs. With more advanced distributions, like my favorite, Gentoo, you generally have to know what you're doing in order to turn on a dangerous daemon, so you're unlikely to do it by mistake.
But they make it so easy! Just kidding. I agree with you here... as long as the open source community is seen as a bunch of squabbiling stoners and teenagers, it's going to have trouble being seen as a serious alternative. However, I think a lot of progress has been made and continues to be made, especially in the public eye. Of course, you can get a very different impression if you read Slashdot. Thank goodness Joe CEO doesn't...
+5 Funny!
or +5 Insightful... I can't decide which...
My little brother is (informally diagnosed) ADD, possibly ADHD, and after reading a couple of chapters of Driven to Distraction I'm convinced that I'm at least borderline ADD.
I've always felt, however, that my distractability was not a negative trait -- in fact, with some effort, I've been able to train myself so I can, for instance, participate in two conversations at once, or some similar multi-input task. In fact, I'm convinced that this is the flip side of what the Jargon File calls hack mode, "a Zen-like state of total focus on The Problem" to the exclusion of everything else. This is an ability that I greatly value, even though it's sometimes not fun to finish a project and then suddenly realize that I've forgotten to eat, sleep, or use the bathroom for way too long.
My question, then is: is this a trait (or a mental ability) that is related somehow to the "hacker" mentality? Comments above have indicated that ADD children are often very intelligent and often have logical/mathematical abstract thinking patterns -- both of which are characteristic of hackers. I'd love to see some statistics (if there are any) on the relationship there.
No hard feelings?
What was your point again?
How did you create the partitions? If the partition type flags aren't set to 83 (Linux), then Windows might think that there's nothing on them. (Linux mount can sometimes do the right thing even though the partition flag is wrong, but it looks like Windows isn't that intelligent.) Also, Win2k (IIRC) has a tool in the drive manager that allows you to hide partitions completely... that should stop the problems you're having with Explorer...
I mean, it's only /. karma. And it's only prestige for a /. account.
There are lots of things worth cursing about in the world. This is not one of them.