Busy portal, lots of content and services Ah. So like iGoogle.
Still, I would count it in the category of "search engine and online endeavors."
Lean and clean search interface with no ads. Last I checked, Google has ads. Live does too, in pretty much exactly the same position and format.
Like all sequels, they are attempts to milk the cash cow created by the original franchise Not all sequels. Stargate wasn't that great of a movie, and I'm guessing wasn't that popular -- but SG-1 became a much better show. Buffy the Vampire Slayer was an incredibly campy, shallow movie, but the TV series actually had depth. (In this case, likely because the writer had much more creative control over the series.)
And not really in the same league, but I don't think anyone would call Serenity worse than Firefly.
Chronicles of Riddick -- it's not as if Pitch Black was a particularly good or well-known movie. It wasn't even promoted as a sequel that way. Not saying Riddick was great, but it was better than Pitch Black. But that defies stereotypes anyway -- there was a kind of ok anime, but the best was the videogame.
One more, while I'm at it: Star Trek. Even numbered movies vs odd.
Matrix I actually didn't think the sequels were that bad. In particular, I think what was probably needed was some serious budget cuts and an editor -- the version we saw in the theaters resembles a "Director's Cut".
Trim down the absurdly long action scenes, trim down the rambling dialog, and they could actually be good. Want to see the original be bad? Play the Path of Neo videogame.
Then again, the biggest problem is that it's exactly the same story they told with the original -- The One slowly wakes up, discovers a bigger world, gains new powers, and in the last few minutes of the movie, he has an epiphany and simply solves the problem, Deus Ex Machina style. (The Machine swarm consciousness is even credited as Deus Ex Machina.)
people seem to heap an awful lot of scorn on Mel Brooks because some of his newer movies don't have quite the same magic as Blazing Saddles or Young Frankenstein Weird. I saw Blazing Saddles recently, and it just didn't seem that funny. It was alright, but it seemed to hover somewhere between silly and boring, and was only occasionally brilliant. Maybe you have to be a certain age -- saw Robin Hood (Men in Tights) when I was much younger, and loved it.
Same with Spaceballs, which brings us back on topic...
the kid who played Anakin. Especially the kid. Just superbly terrible acting. Luke was whiny, yeah, but Anakin was dreadful. Yes, it was painful to watch, which doesn't really make for a good movie. But it was also damned realistic.
Have real life people been killed because of it? Dude, real people have been killed because of an Xbox 360. What kind of metric is that?
Oh, and there's a military laser project named after it.
Are people willing to give their life, or alter their concept of what life is about in the most sacred way, because of it? I'd say that's what "religion" means, and enough people call themselves Jedi to make it a religion.
to say that it has "shaped the course of human history" is a bit over the top. Not going to debate that one, but I don't think that's what GP said:
You may not like the movie, but to say it's "just a movie" is like saying "the Bible is just a book"--perhaps in some literal sense it's "a book," but it's one that has shaped the course of human history. So that "shaped the course of human history" is about the Bible, and why it shouldn't just be called "a book". There are other reasons that Star Wars should not just be called "a movie"...
And frankly, it's too early to tell. We've had the Bible for at least a millennium or two. We've only been able to make movies for a little over a century -- and only in color, with sound, for about half that time.
It's simply physically impossible for a movie to have had as much of a chance to become as world-changing (for better or worse) as the Bible is -- it's simply had more time.
Clearly for the first couple of films the right people were in the right place at the right time. Actually, what I loved about the first few films was how clever they were, especially the first one. I don't know if they were actually low-budget, but they looked like it -- in some scenes, the landspeeder was a decorated car, with the wheels removed by smearing vaseline on the film.
My guess is, there wasn't that much of a budget for the first movie. But I don't know, I'm not reading the book either.
Mark Hammil and Carrie Fisher weren't exactly any good. Alec Guinness was pretty damned good, though.
As for continuity? Please! One minute Luke and Leia are about to get hot and heavy, and the next we're told they're brother and sister. ...which they didn't know. Retconning? Maybe. But absolutely consistent, if a bit gross.
Vader as Luke's father was unlikely though plausible, that is until the pathetic explanation that was Episode 3. And what's pathetic about that, exactly?
C3PO and R2D2, using their point of view, is really the most risky and rewarding aspect of star wars. now, i don't think lucas would ever admit it, but i think he was trying to conjure up the same sort of picaresque magic twice... with the character jar jar binks The devil's in the details. Just drop all the analysis for a moment and actually watch it for what's there...
Jar Jar has a high, whiny, irritating voice. He appears to be based on an incredibly offensive stereotype. He looks goofy at best. He's clumsy -- he may try to help, but if he actually does any good, it's only because of pure dumb luck. That's just off the top of my head.
Comparing him to R2 -- R2 is cute. He's got personality, despite being a machine (almost because of it), and initiative. He usually knows what's going on (moreso than 3PO), and is actually helpful.
I actually liked most of the prequels alright -- saw the first when I was young enough to enjoy it (even Jar Jar), and didn't have high hopes for the second and third (by then I was old enough to hate Jar Jar). There were a few really horrible moments, and also a few moments worth watching.
But it does say something when Ryan vs Dorkman is more fun to watch than most of the lightsaber duels in the actual movies.
I've tried 3 different terminal emulators and none do. I've used aterm, xterm, gnome-terminal, and konsole myself. There are many, many others.
How do YOU extend highlighting in an X terminal? I don't, but after some experimenting, I've found that in my browser, shift+click will expand it. Doesn't work in my konsole, though. Weird.
I did, however, find a "set selection end" option on the right-click menu in konsole. That makes it two clicks instead of one, which is annoying, but the feature is there.
Changing the mouse pointer from a pixel stepping "I-bar" to an inverse block of the whole character cell, would be nice (that's what I have in text mode). I don't really have an answer to that, although it is really the same thing -- after all, one pixel you're on one character, one pixel you're on another, either way. In a GUI, it makes sense to simply click somewhere between the character you want and the character you don't.
Of course, that's not a solution, just amazed it was a problem. I'm also not entirely sure I see advantages to your way, other than that you're used to it -- give it a few weeks, maybe?
Another trick I like, though it isn't always applicable, is double-click -- selects a word -- and double-click+drag, to select multiple words.
I also need an instant change from one terminal screen to another terminal screen directly. That's why I mentioned a few that support tabs.
That means no fancy moving or sliding of windows or desktops in and out and no switching screens in rotation one by one. Maybe this is best done by desktops rather than windows. Depending on the window manager, yes, you could do that. I always do it one by one, but I do see shortcuts for all of them which can be set.
And checking again with Konsole, yes, I can switch to an arbitrary "session" (read: tab) with a configurable keystroke. And it's possible to hide the tab bar. Unfortunately, it seems to be limited to 12 such keystrokes/sessions out of the box, but I suspect it would be possible to change that with a config file, and if not, it should be a trivial hack, even if you know nothing about GUI programming.
I can't see any reason why there would be effects you can't turn off.
And this needs to work even if the terminal emulator is in full screen mode Can't see why it wouldn't, though of course the proper solution there is to find a window manager that will let you throw apps into a fullscreen mode.
Also, especially when I'm in a terminal, I barely touch the mouse at all. Not entirely on topic, but there is one feature I desperately miss from my Mac: It distinguished pretty thoroughly between "windows" and "applications". I would usually open four 80x25 Terminals, and I'd have a keystroke to cycle through them -- and still room on the screen for some GUI apps.
I know about alt+tab, but that switches between all windows, or if you're lucky, all apps. I want a way to switch between apps, and then another way to switch between windows belonging to an app (or some arbitrary value of "group").
My programming area is networks and servers. I don't know anything about GUI programming. Do you know anything about FPGA programming? Seems like it might become as big a project either way...
So it looks like, at the end of the day, your biggest problem is the GUI treatment of a cursor, vs good old gpm.
Microsoft marketing is confusing and uses the term "live" for their search engine and for their online endeavors. Then what, praytell, is the MicroSoft Network supposed to be?
Just so long as you at least verify fingerprints via the phone. Fingerprints aren't any more secret than the public key, but at least on the phone, a MITM insertion attack is much more difficult -- they would sound different.
VPN access is per-machine. FTP access is per-user. Making it accessible to anyone on the VPN is equivalent to chmod'ing 777.
It's amazing how many people make this mistake. NEVER implement an unauthenticated protocol, unless you can completely guard access to it -- and by that, I mean use it over pipelines, UNIX sockets, or in wrappers that include authentication.
Oh, and FTP sucks. I can't think of a good reason to use it at all, ever. Use Samba if it's convenient, otherwise things like scp/sftp, rsync, or actual database replication.
Linux virtual console semantics, which work better and faster than terminal emulation under X does ...what?
It must have been a very, very long time since you've tried terminal emulation under X. It's trivial now to get something like Konsole or gnome-terminal, or even a ton of aterms under Fluxbox, to run in a tabbed mode. Given the right combination of terminal emulation and window manager, you could probably get it as good or better than you've been using under a VT.
Why force your video card to render text, when X can do it for you?
Bonus -- no lag from switching modes (or anything else) when flipping between the terminal and a GUI program. Double-bonus -- you can actually copy/paste between the terminal and a GUI program. And just because I like it -- you can always shave off a chunk of the screen for a panel, and put useful indicators there -- a clock in particular.
I think shopping in a nice muscle car or even a ricejob would be a lot more exciting than a Formula 1. A nice muscle car, or a "ricejob" (offensive term?), is going to do more than one thing. It'll have heating, cooling, a radio, turn signals...
A Formula 1 car might not even be street legal, let alone comfortable for anything other than racing.
Right tool for the job, can't be said enough. Even CPUs have things like math coprocessors. It's no surprise that even if a video card technically can run Linux, it might not be very good at it. (A fair example: PS2 Linux.)
the first think I would buy would be a fancy car A few hundred thousand should get you a damned nice car. Anything after that and you're just wanking.
And I would pay off my house too.... and take a vacation... How much is that?
I'm betting an open source video card is going to cost enough that unless you go out of your way to spend money, you can afford pretty much all the stuff you listed, and it'd still be a drop in the bucket compared to development costs.
If I had, say, ten or twenty million dollars, I'd spend a few hundred thousand on myself, and then think about how to benefit the world with the rest of it.
I mean 2.5, precisely because it was the unstable branch. The very notion of a long-term unstable branch spanning years would seem to go against the idea that we could pick any kind of synchronized release schedule that would hold for every project.
Most people just want a functional suite of office applications that works more or less the way they've come to expect such programs. And thanks to Office's Ribbon, and various Vista "improvements", Linux and OpenOffice may actually be closer to the way they've come to expect.
I can't tell if this was sarcasm or not.
I can't count the number of times I've been writing a letter to family and needed the wife to put in a few words. Just "a few words"... so she can't just come to that computer to type it up?
All it took was an extra server, and new licensing seats. Compare that to, say, Google Docs, in which all it takes is an Internet connection. Or a Samba server and OpenOffice -- can't edit simultaneously, but if it's just between you and your wife, how hard is it to cooperate on that?
I'll be recommending people upgrade to Office 2007, if and only if I can configure some reasonable support for ODF as the default file format. Otherwise, I'll recommend they upgrade to OpenOffice or some other alternative.
How do we know they aren't going to try to do what they successfully did to Netscape. Oh please, let them try. Have you forgotten that Mozilla rose from the ashes of Netscape, and Firefox from the ashes of Mozilla?
If however they are really trying to comply with ODF then hats off to MS for being serious about embracing standards. It might be the first time they've done this in good faith...
But then, look at IE. It took some serious competition, in the form of Firefox, but IE finally did shape up and start adding features (tabs) and reasonable standards compliance.
All we really need, then, is an ACID test for ODF, in which we can show that OpenOffice, KOffice, Google Docs, and even isolated projects like AbiWord and Gnumeric do better than Office, thus shaming Microsoft into doing it right. That assumes they don't get it right the first time, although that does seem unlikely.
I doubt much time, if any at all, was put into security considerations for the code. I doubt much should.
Honestly, what is the attack vector here? This isn't a web browser. It's not even a multiplayer game. I very much doubt it needs to talk to the network at all.
I can imagine that parts of it would be "insecure", but only in the sense that someone who already has access to the system could exploit the game to... get access to the system. Horrors!
Imagine someone finding a bug, and somehow through magic there is a whole trusted system of which this patch will get reviewed and distributed back to the schools, and have them actually update all copies. I know! We'll call it a package manager! In fact, it could even manage more than one "package", and one package could depend on another... Man, this is going to revolutionize the world of 1995!
Seriously, open source it, we'll put it in Edibuntu, problem solved.
Obscurity may be bad security Wrong. Obscurity is not security, full stop.
People will find vulnerabilities, if they are exploitable at all -- especially if it's as widely distributed as you say. If it's in a position to be compromised for anything (see above), you'll need a patch distribution system, and you'll need people finding vulnerabilities.
Open sourcing it would be a much better use of taxpayer dollars than hiring a crack team to do all that, and port it to Linux for the schools that want it.
There's always the possibility of the protocol itself being updated, or of either protocol or your users demanding new features.
There's also the possibility of DJB deciding on his own interpretation of the protocol, often going back to the actual RFC, and ignoring how it's implemented. I wouldn't mind this, if there was a "plays nice with others" option to enable, but there isn't -- about all I could do is edit the source myself and recompile, or download someone else's patch and hope it applies properly.
And, of course, no way to get it from a repository -- I need to compile from scratch every time.
That's why it being public domain helps, but there's still the problem of either dealing with DJB or forking it.
Simple example: Most mailservers will accept either \n or \r\n line endings. For example, the end of a message is signaled by a dot on a single line -- in other words, \r\n.\r\n -- and DJB insists on only accepting that, thus making it a bitch to talk to qmail via telnet or netcat. Others, like Postfix, actually are liberal in what they accept, and are actually entertaining and almost fun to talk to over netcat.
I'm probably unique in that I actually like the way DJB software is often configured. I like things like maildir, and the use of a makefile to rebuild a "cdb" file, and a directory full of individual files (often hardlined) to configure which domain goes to which upstream server in dnscache.
So I like some of DJB's design patterns, but I can't deal with his personality problems, and his cowboy approach to standards.
Unbound is a DNS resolver, not a server. PowerDNS will do both. As a server, it's technically offtopic, but...
I love the fact that there are pluggable backends. More than that, I love the pipe backend. I realize this is an "everything looks like a nail" scenario, but I actually wrote a PowerDNS->REST client with that, and then a Rails server behind it.
Slow? Sure, but I can always setup a slave -- either someone like DynDNS, or another PowerDNS server with a faster backend (MySQL, Postgres, maybe even SQLite?)
Overkill? Sure, but I can't get over the fact that I've written a DNS server in Rails.
For whatever reason, people have decided that a holy quest to "destroy Microsoft" and encourage wide-spread adoption of gnu/linux-based operating systems would be totally awesome. Mostly for the reason that if this did happen, we wouldn't have to deal with Microsoft.
Simple example: I have to know how to setup and use Samba, so I can share files with Windows. If Linux was the majority, it would be Microsoft who would have to implement a decent NFS (or ssh) client instead.
Mint is even better with its media support, but completely lacks dev tools if you install from the live image -- seriously, what sort of *nix system thinks you don't need a C compiler by default and makes you go looking for it in the repositories? A smart one?
I remember loving Gentoo because of this kind of thing -- nothing preinstalled except the bare minimum to boot, get a commandline, and run the package manager. After that, it's up to you to grab the rest. I've done the same with minimal Debian installations -- it's awesome.
I would much rather be able to type "sudo apt-get install gcc" than have the machine assume I'm a C programmer. Even better, on Ubuntu, I just type "gcc" or "make", and if it's not installed, the error message tells me exactly what command to type to install it from the package manager.
The systems are targeted at different sets of people with different requirements and philosophies. Holding off on releasing Red Hat until Ubuntu is ready, which requires KDE and GNOME to sync up (more or less) sounds a little ridiculous and over-the-top. That, I agree with.
Slashdot Mirror
Still, I would count it in the category of "search engine and online endeavors." Lean and clean search interface with no ads. Last I checked, Google has ads. Live does too, in pretty much exactly the same position and format.
And not really in the same league, but I don't think anyone would call Serenity worse than Firefly.
Chronicles of Riddick -- it's not as if Pitch Black was a particularly good or well-known movie. It wasn't even promoted as a sequel that way. Not saying Riddick was great, but it was better than Pitch Black. But that defies stereotypes anyway -- there was a kind of ok anime, but the best was the videogame.
One more, while I'm at it: Star Trek. Even numbered movies vs odd. Matrix I actually didn't think the sequels were that bad. In particular, I think what was probably needed was some serious budget cuts and an editor -- the version we saw in the theaters resembles a "Director's Cut".
Trim down the absurdly long action scenes, trim down the rambling dialog, and they could actually be good. Want to see the original be bad? Play the Path of Neo videogame.
Then again, the biggest problem is that it's exactly the same story they told with the original -- The One slowly wakes up, discovers a bigger world, gains new powers, and in the last few minutes of the movie, he has an epiphany and simply solves the problem, Deus Ex Machina style. (The Machine swarm consciousness is even credited as Deus Ex Machina.)
Same with Spaceballs, which brings us back on topic... the kid who played Anakin. Especially the kid. Just superbly terrible acting. Luke was whiny, yeah, but Anakin was dreadful. Yes, it was painful to watch, which doesn't really make for a good movie. But it was also damned realistic.
Oh, and there's a military laser project named after it. Are people willing to give their life, or alter their concept of what life is about in the most sacred way, because of it? I'd say that's what "religion" means, and enough people call themselves Jedi to make it a religion. to say that it has "shaped the course of human history" is a bit over the top. Not going to debate that one, but I don't think that's what GP said: You may not like the movie, but to say it's "just a movie" is like saying "the Bible is just a book"--perhaps in some literal sense it's "a book," but it's one that has shaped the course of human history. So that "shaped the course of human history" is about the Bible, and why it shouldn't just be called "a book". There are other reasons that Star Wars should not just be called "a movie"...
And frankly, it's too early to tell. We've had the Bible for at least a millennium or two. We've only been able to make movies for a little over a century -- and only in color, with sound, for about half that time.
It's simply physically impossible for a movie to have had as much of a chance to become as world-changing (for better or worse) as the Bible is -- it's simply had more time.
My guess is, there wasn't that much of a budget for the first movie. But I don't know, I'm not reading the book either. Mark Hammil and Carrie Fisher weren't exactly any good. Alec Guinness was pretty damned good, though. As for continuity? Please! One minute Luke and Leia are about to get hot and heavy, and the next we're told they're brother and sister. ...which they didn't know. Retconning? Maybe. But absolutely consistent, if a bit gross. Vader as Luke's father was unlikely though plausible, that is until the pathetic explanation that was Episode 3. And what's pathetic about that, exactly?
Jar Jar has a high, whiny, irritating voice. He appears to be based on an incredibly offensive stereotype. He looks goofy at best. He's clumsy -- he may try to help, but if he actually does any good, it's only because of pure dumb luck. That's just off the top of my head.
Comparing him to R2 -- R2 is cute. He's got personality, despite being a machine (almost because of it), and initiative. He usually knows what's going on (moreso than 3PO), and is actually helpful.
I actually liked most of the prequels alright -- saw the first when I was young enough to enjoy it (even Jar Jar), and didn't have high hopes for the second and third (by then I was old enough to hate Jar Jar). There were a few really horrible moments, and also a few moments worth watching.
But it does say something when Ryan vs Dorkman is more fun to watch than most of the lightsaber duels in the actual movies.
I did, however, find a "set selection end" option on the right-click menu in konsole. That makes it two clicks instead of one, which is annoying, but the feature is there. Changing the mouse pointer from a pixel stepping "I-bar" to an inverse block of the whole character cell, would be nice (that's what I have in text mode). I don't really have an answer to that, although it is really the same thing -- after all, one pixel you're on one character, one pixel you're on another, either way. In a GUI, it makes sense to simply click somewhere between the character you want and the character you don't.
Of course, that's not a solution, just amazed it was a problem. I'm also not entirely sure I see advantages to your way, other than that you're used to it -- give it a few weeks, maybe?
Another trick I like, though it isn't always applicable, is double-click -- selects a word -- and double-click+drag, to select multiple words. I also need an instant change from one terminal screen to another terminal screen directly. That's why I mentioned a few that support tabs. That means no fancy moving or sliding of windows or desktops in and out and no switching screens in rotation one by one. Maybe this is best done by desktops rather than windows. Depending on the window manager, yes, you could do that. I always do it one by one, but I do see shortcuts for all of them which can be set.
And checking again with Konsole, yes, I can switch to an arbitrary "session" (read: tab) with a configurable keystroke. And it's possible to hide the tab bar. Unfortunately, it seems to be limited to 12 such keystrokes/sessions out of the box, but I suspect it would be possible to change that with a config file, and if not, it should be a trivial hack, even if you know nothing about GUI programming.
I can't see any reason why there would be effects you can't turn off. And this needs to work even if the terminal emulator is in full screen mode Can't see why it wouldn't, though of course the proper solution there is to find a window manager that will let you throw apps into a fullscreen mode.
Also, especially when I'm in a terminal, I barely touch the mouse at all. Not entirely on topic, but there is one feature I desperately miss from my Mac: It distinguished pretty thoroughly between "windows" and "applications". I would usually open four 80x25 Terminals, and I'd have a keystroke to cycle through them -- and still room on the screen for some GUI apps.
I know about alt+tab, but that switches between all windows, or if you're lucky, all apps. I want a way to switch between apps, and then another way to switch between windows belonging to an app (or some arbitrary value of "group"). My programming area is networks and servers. I don't know anything about GUI programming. Do you know anything about FPGA programming? Seems like it might become as big a project either way...
So it looks like, at the end of the day, your biggest problem is the GUI treatment of a cursor, vs good old gpm.
Just so long as you at least verify fingerprints via the phone. Fingerprints aren't any more secret than the public key, but at least on the phone, a MITM insertion attack is much more difficult -- they would sound different.
VPN access is per-machine. FTP access is per-user. Making it accessible to anyone on the VPN is equivalent to chmod'ing 777.
It's amazing how many people make this mistake. NEVER implement an unauthenticated protocol, unless you can completely guard access to it -- and by that, I mean use it over pipelines, UNIX sockets, or in wrappers that include authentication.
Oh, and FTP sucks. I can't think of a good reason to use it at all, ever. Use Samba if it's convenient, otherwise things like scp/sftp, rsync, or actual database replication.
It must have been a very, very long time since you've tried terminal emulation under X. It's trivial now to get something like Konsole or gnome-terminal, or even a ton of aterms under Fluxbox, to run in a tabbed mode. Given the right combination of terminal emulation and window manager, you could probably get it as good or better than you've been using under a VT.
Why force your video card to render text, when X can do it for you?
Bonus -- no lag from switching modes (or anything else) when flipping between the terminal and a GUI program. Double-bonus -- you can actually copy/paste between the terminal and a GUI program. And just because I like it -- you can always shave off a chunk of the screen for a panel, and put useful indicators there -- a clock in particular.
A Formula 1 car might not even be street legal, let alone comfortable for anything other than racing.
Right tool for the job, can't be said enough. Even CPUs have things like math coprocessors. It's no surprise that even if a video card technically can run Linux, it might not be very good at it. (A fair example: PS2 Linux.)
I'm betting an open source video card is going to cost enough that unless you go out of your way to spend money, you can afford pretty much all the stuff you listed, and it'd still be a drop in the bucket compared to development costs.
If I had, say, ten or twenty million dollars, I'd spend a few hundred thousand on myself, and then think about how to benefit the world with the rest of it.
I mean 2.5, precisely because it was the unstable branch. The very notion of a long-term unstable branch spanning years would seem to go against the idea that we could pick any kind of synchronized release schedule that would hold for every project.
I'll be recommending people upgrade to Office 2007, if and only if I can configure some reasonable support for ODF as the default file format. Otherwise, I'll recommend they upgrade to OpenOffice or some other alternative.
But then, look at IE. It took some serious competition, in the form of Firefox, but IE finally did shape up and start adding features (tabs) and reasonable standards compliance.
All we really need, then, is an ACID test for ODF, in which we can show that OpenOffice, KOffice, Google Docs, and even isolated projects like AbiWord and Gnumeric do better than Office, thus shaming Microsoft into doing it right. That assumes they don't get it right the first time, although that does seem unlikely.
Honestly, what is the attack vector here? This isn't a web browser. It's not even a multiplayer game. I very much doubt it needs to talk to the network at all.
I can imagine that parts of it would be "insecure", but only in the sense that someone who already has access to the system could exploit the game to... get access to the system. Horrors! Imagine someone finding a bug, and somehow through magic there is a whole trusted system of which this patch will get reviewed and distributed back to the schools, and have them actually update all copies. I know! We'll call it a package manager! In fact, it could even manage more than one "package", and one package could depend on another... Man, this is going to revolutionize the world of 1995!
Seriously, open source it, we'll put it in Edibuntu, problem solved. Obscurity may be bad security Wrong. Obscurity is not security, full stop.
People will find vulnerabilities, if they are exploitable at all -- especially if it's as widely distributed as you say. If it's in a position to be compromised for anything (see above), you'll need a patch distribution system, and you'll need people finding vulnerabilities.
Open sourcing it would be a much better use of taxpayer dollars than hiring a crack team to do all that, and port it to Linux for the schools that want it.
Development was sponsored by public funds. We all paid for it.
Give me one good reason the source code shouldn't be released.
There's always the possibility of the protocol itself being updated, or of either protocol or your users demanding new features.
There's also the possibility of DJB deciding on his own interpretation of the protocol, often going back to the actual RFC, and ignoring how it's implemented. I wouldn't mind this, if there was a "plays nice with others" option to enable, but there isn't -- about all I could do is edit the source myself and recompile, or download someone else's patch and hope it applies properly.
And, of course, no way to get it from a repository -- I need to compile from scratch every time.
That's why it being public domain helps, but there's still the problem of either dealing with DJB or forking it.
Simple example: Most mailservers will accept either \n or \r\n line endings. For example, the end of a message is signaled by a dot on a single line -- in other words, \r\n.\r\n -- and DJB insists on only accepting that, thus making it a bitch to talk to qmail via telnet or netcat. Others, like Postfix, actually are liberal in what they accept, and are actually entertaining and almost fun to talk to over netcat.
I'm probably unique in that I actually like the way DJB software is often configured. I like things like maildir, and the use of a makefile to rebuild a "cdb" file, and a directory full of individual files (often hardlined) to configure which domain goes to which upstream server in dnscache.
So I like some of DJB's design patterns, but I can't deal with his personality problems, and his cowboy approach to standards.
Unbound is a DNS resolver, not a server. PowerDNS will do both. As a server, it's technically offtopic, but...
I love the fact that there are pluggable backends. More than that, I love the pipe backend. I realize this is an "everything looks like a nail" scenario, but I actually wrote a PowerDNS->REST client with that, and then a Rails server behind it.
Slow? Sure, but I can always setup a slave -- either someone like DynDNS, or another PowerDNS server with a faster backend (MySQL, Postgres, maybe even SQLite?)
Overkill? Sure, but I can't get over the fact that I've written a DNS server in Rails.
Simple example: I have to know how to setup and use Samba, so I can share files with Windows. If Linux was the majority, it would be Microsoft who would have to implement a decent NFS (or ssh) client instead. Mint is even better with its media support, but completely lacks dev tools if you install from the live image -- seriously, what sort of *nix system thinks you don't need a C compiler by default and makes you go looking for it in the repositories? A smart one?
I remember loving Gentoo because of this kind of thing -- nothing preinstalled except the bare minimum to boot, get a commandline, and run the package manager. After that, it's up to you to grab the rest. I've done the same with minimal Debian installations -- it's awesome.
I would much rather be able to type "sudo apt-get install gcc" than have the machine assume I'm a C programmer. Even better, on Ubuntu, I just type "gcc" or "make", and if it's not installed, the error message tells me exactly what command to type to install it from the package manager. The systems are targeted at different sets of people with different requirements and philosophies. Holding off on releasing Red Hat until Ubuntu is ready, which requires KDE and GNOME to sync up (more or less) sounds a little ridiculous and over-the-top. That, I agree with.
So this is where we got the horror that is the KDE4 Panel.