Interesting thought -- for a while at Boston College there were "Computers in Management" (as in "how to sit at your desk and write presentations) classes run in ClarisWorks as the prevailing environment. Didn't last though.
See, MIT's CS majors (I don't know if they still do this, since I never went there) used to learn to program in Scheme. Not a bad idea, given that it makes you think about how to do things.
What I'd love to know is how many "CS" classes amount basically to "how to use Visual C++ on Windows". My attitude is that if you're learning to program, your code had better not be platform dependent.
Truly. Plan 9/Brazil ('cos that's what they seem to be calling it internally these days) is quite cool (picture it: a completely distributed operating environment where your desktop screen is pretty much the same whether you're running on a PC or a mainframe or something in between); it's Unix and NFS taken to extremes.
I would love to have played with Plan 9 myself, but hardware costs enough to begin with; I don't want to shell out $300 either for something I'd consider a learning experience (like the reason I once bought a secondhand Newton; I'm a Palm junkie but I wanted to learn). Their best bet if they want to make a splash and break this percieved deadlock is to open it up. Put it under something like a BSD or APSL-type license and give it away at media cost, then see what happens...
Nononononono. The Ethernet is used only on the fighters. Anyone can clearly see that the fighter umbilical is FDDI. (And of course the mother ship runs Windows -- don't you know that the reason Jeff Goldblum was using a Mac (apart from the fact that they own him anyway) was that the aliens couldn't backcrack him? Go ahead, moderate me down; my karma is a bit swollen anyway, and this bit wasn't half as funny as I'd hoped.../Brian
Well, read it. The funniest part apart from that is the part where they say that copy protection is important for creativity. The fact is that the entertainment industry as we've known it since Thomas Edison invented the phonograph and (can't quite remember how to spell Edward in Anglo-Saxon) Muybridge invented the motion picture camera is dead and was from the moment MP3 hit the streets. I've recently reluctantly partially come around to the pirates' side, not because I think that what they do is right (I don't; if nothing else it shows a tremendous lack of respect for the creators of the product) but because I think the whole issue is moot. Look around us. The tech is there for everybody to see. For six grand I can get an iMac, a digital camcorder, and a GoVideo double-decker VCR and start my own movie studio. For ten grand, I can make that a PowerMac G4 or a high-end Sony VAIO and a nice big RAID array. For the price of a new sport-utility vehicle, I could build a small Linux-based rendering cluster. You get the idea. I predict that we will never see a monster-budget blockbuster like Titanic again. Within two years, the entire industry will be in anarchy as execs realize that there's no hope of holding the line on the status quo. Star salaries will plummet and production budgets will tailspin as the studios try to hedge their bets against collapse. What will this lead to? Your guess is as good as mine./Brian
And the thing was that when PPC came down the pike, all of us Mac junkies (or ex-Mac junkies as the case may be; I still like Macs alot but I'm a Linux loyalist these days) were saying, "Cool! Decent game hardware!"
It was; the hardware was never the problem, and a lot of game developers were going to sign on pretty early. But nothing ever really came of it until well after the advent of the iMac because a) Copland crashed and burned when Apple engineers couldn't quite figure out when to stop adding things (and a bloody shame too; I rather wish Apple would release the source code for posterity because what I saw of Copland was pretty damn cool) and b) Apple *almost* crashed and burned under the listless leadership of Michael Spindler cost them their first shot at a unified MacOS/Unix system (the long-forgotten failed Sun merger).
To be honest, I don't really blame John Carmack for being anti-Mac. To this day the MacOS is best thought of as a cluttered attic -- it may in fact be organized, but it's not the best situation and it's long past time to junk it and rebuild from scratch. That's Rhapsody and Carbon.
Even better -- Quake (or at least Q1) (as well as Doom, now that I think of it) is open-source. Carmack's got them coming and going -- if they lose him, not only do they shoot themselves in the foot, Quake forks.
And then, we find out if trademarks still matter in the Open Source world.
And I suppose you have the same opinion about Dilbert or Rhymes With Orange, two of (IMHO) the most brilliantly funny mainstream comics out there right now. And neither Scott Adams nor Hillary Price is really all that much of an artist.
If anything, User Friendly is much funnier than either of the above. The thing is that it's very in-jokey as well; for those who understand it, it raises the humor value, but for those who don't, it's a turnoff. I'm thinking you mostly just don't get the jokes.
Anyone gonna get this one right? BSD is older than Linux by a decade and a half, and there's a lot of story there that it doesn't seem like anyone still remembers...
Okay, flames off. Grab some cocoa and rally 'round, and ol' Uncle Brian gonna tell you about some other guy named Bill and his friends...
Unix didn't really become Unix until John Lions wrote the Lions book (and any Slashdotter who doesn't have a copy should be shaken mercilessly until enough money to either buy or photocopy a copy falls out; about US$30 should do it for a legit copy, and a trip to Kinko's would be much cheaper, not to mention totally appropriate if you know the story:-) ). At that point, people started banging on the Unix kernel all over, and the Open Source culture as we know it began.
It's easy enough with all the Linux hype these days to lose sight of where it all began, but here's the story: back around '76, when AT&T shipped Unix V6, the University of California at Berkeley got its hands on a source license and its students started hacking away at it. Somewhere along the way, BSD supplanted TENEX and its proprietary brethren as the OS running the backbone of the then-ARPANET; as a result, it became THE wide-area network OS for the eighties. Software such as the BSD TCP/IP stack (still the standard outside the Linux world) and sendmail are still at the core of much of the net, and most of it came out of Berkeley.
Along the line, many of these students struck out on their own (most notably Bill Joy at Sun), and when they went off to build their workstation systems, they took BSD with them, thus bringing about systems such as SunOS and HP/UX, as well as several that are now long forgotten; not only was the BSD license in effect a donation to the public domain, the code itself was something that the engineers themselves had developed, which was more than could be said about the multiple permutations of Unix coming out of Murray Hill, NJ. BSD even came to the PC, in the form of Bill and Lynne Jolitz' 386BSD, a precursor to the modern BSD designs.
The modern BSD era began in 1994, when UC-Berkeley's Computer Science Research Group, disbanded a few years earlier, regrouped to finish the BSD story at Berkeley. The release they made was called 4.4BSDlite, and was actually quite widely available in bookstores at the time. It consisted of a CD-ROM containing a mostly-complete AT&T-free source tree (with a few unreplaced but critical AT&T files removed) and four books containing man pages and other docs.
This event is what created two of the four modern BSD variants (FreeBSD and netBSD); OpenBSD was a fork off of netBSD that stressed security concerns, while the newest major variant, Darwin/MacOS X, came out of Apple's post-acquisition merging of NeXT's Mach/BSD hybrid OS and the 4.4lite codebase. BSD was also the base for OSF/1, the only major variant of which is Compaq's Tru64 Unix.
Today, I think BSD is viewed as a one-off, a newcomer in a world dominated by Linux. I suspect most of the history I've just related simply isn't all that well known anymore, and it certainly isn't media-friendly (what, no underdog-makes-good? No flashy technology? No David-vs.-Goliath? HACKERS?! Not interested, sorry...). This is a shame; that's why I'm posting this.
I think the reason that BSD persists today (apart from having about the best TCP/IP stack available) is because it's a link to the past that most commercial Unices haven't preserved. By running a BSD, you're expressing solidarity with some random hacker poring over a third-generation photocopy of the Lions book on a brand-new VAX twenty years ago. And hey, I prefer penguins to daemons myself, but that's a connection that we Linux users can't make.
Saying that BSD is related to Linux is wrong on many levels. Apart from both being open-source Unices, there is no real connection. Linux is its own beast, and I think over the long term it will eventually evolve away from its Unixness somewhat. Minix, now freed from its old copyright restrictions, will probably morph into an embedded-systems OS; I'd love to see it on a WinCE (MinCE?) unit, for example, since it's just about the perfect size. But BSD will still be carrying the Free Unix torch years from now. Long may it.
An interesting, and not entirely uncalled-for comment...
I've seen the movie myself. I look upon Truffaut's version of F451 as being a mixed bag: generally a tremendous waste of time and a massive perversion of Bradbury's rather gritty, all-too-real plot with a few strokes of absolute genius thrown in to keep it from being a total failure.
The bad: Wow. Too much to mention. Men in jetpacks instead of the six-legged Hound. Name changes. Wooden acting. Absurd direction. This beast should have been burned along with the books. The closing scene, where the human books gather, is disgustingly maudlin. The plot was streamlined and stylized into unrecognizability (including the removal of Guy Montag's first name). The production as a whole was a soporific, entirely missing the lurching, gut-wrenching action-lull pace that Bradbury established in the book.
The good: Surprisingly, a few little touches that would have perked up the book had Bradbury thought of them. The opening credits were done as a voice-over (appropriate for such an aggressively post-literate setting). A few others that don't come to mind.
Oh, well. Maybe someday someone will come along who actually understands SF and will do F451 the way Bradbury actually wrote it.
/Brian
Re:Its proponents would of course be called...
on
AtheOS
·
· Score: 1
Interesting point. I had been trying to figure out what the name meant ever since I saw it a few days ago. I still don't particularly like it (it would be almost as difficult to take seriously as Linus's coinage Freax (fre-ax?)), but as a development name, if that's really the rationale behind it, it works pretty well.
As for the OS itself, I like the idea. POSIX-compliance is pretty much a necessity, but more power to the author if he can create a mainstream non-Unix OS by the same model Linus did. I don't know if I'll ever get around to running AtheOS, but I'm thinking I'll be following it for a while...
Those who do not remember the past are doomed to forget that Donkey Kong was originally the bad guy...
The AC I'm responding to has neglected one of the key aspects of gaming: gameplay beats flashy tech any day. Granted, Q3 and the like are visually amazing. Granted, the closest anyone ever got to Lara Croft in the days of sidescrollers was Samus in Metroid (after you beat the game and figured out she was a she after all).
But you know, all those old games... they were pretty damn fun to play. It's true, the quality isn't that great -- 8 bit chips aren't known for their 3D graphics performance -- but that doesn't mean that they can't keep up with today's best. Frankly, I've seen Super Mario 64, and I don't think it really measures up to the original. Why? Because the original was a side-scroller, whereas SM64 is a third-person 3D game. It actually sort of spoils the effect, not to mention making it much harder to adapt to the new controls.
As for demanding much higher quality games... see what I said above about play vs. tech. There's a company called Ambrosia that has been making some of the best shareware games available for the MacOS for years now, starting with an amazing Asteroids clone called Maelstrom. The best example of what I'm talking about: they redid Centipede/Millipede as Apeiron. The graphics and sound, it is true, are drastically enhanced, and Apeiron has features such as powerups that neither original ever had... but it's the same game. It plays very close to the original, enough to satisfy anyone's Centipede jones. Compare that to the modern Centipede, which is 3-dimensional, graphically gorgeous... and just plain wrong.
You can sit around playing your amped-up Quake games all you want. I'll appreciate it as much as you do. But don't tell me that just because a game is played on 20-year-old hardware it doesn't meet modern quality standards. If Ms. PacMan is still fun to me after fifteen or more years, than I can't legitimately say it's not cool, or tricky, or enjoyable no matter what "quality standards" it does or does not meet.
Actually, that's exactly what happened. Rhapsody/MacOS X Server is NextStep with a Mac interface and the Blue Box (Classic environment) added.
What was going on when Apple bought out NeXT was that Apple needed a new OS and Copland had developed a case of terminal second-system effect that ultimately killed it (and Gershwin, its follow-on). After the Apple/Sun merger collapsed (most likely thanks to Michael Spindler's incompetence), Gil Amelio took over and a bit later made the NeXT buy. Rhapsody hit the radar screen almost immediately.
Rhapsody was originally just the Macified NextStep; new apps were to be written in a Javafied (though you could still use ObjC if you wanted) version of OPENSTEP (now called Cocoa). The developer community, however, by and large rejected Rhapsody in that form, and Apple was forced to backtrack and create Carbon, all the while making subtle changes in what would become MacOS X Server (Apple's third version of Mac Unix, after A/UX and MkLinux, for those of you keeping track at home).
For the record, I find the comment of BSD and Mach being inseparable to be rather funny -- you can't do anything useful by separating them, certainly, but it's a gross oversimplification (like, has anyone considered trying to hack Darwin to run MkLinux on the same microkernel?).
Just out of curiosity, what is a "commercial document" anyway? If I had Office Academic and I start writing a book, does that mean that I had to pony up to publish it legally?
I'm sorry, licenses that control content creation don't wash with me. I believe one of Microsoft's better-known legal abuses in the Chicago days involved a clause in the license for MFC that made it rather complicated for anyone to write something like a word processor or spreadsheet with MFC because it competed directly with MS. I can't possibly see how this can be considered legit -- it's impossible to enforce and it seems to be in rather bad faith. The funny thing is that MS did this in a market where there were plenty of VC++ alternatives out there...
(For convenience -- and since I'm a longtime Mac man who followed this whole scene pretty closely -- I will refer to OS X Server as Rhapsody.)
It's still a fair comparison -- OS X Server/Rhapsody is sort of a transitional form, maybe a Homo ergaster version of the Homo sapiens MacOS X final.
With Apple flying as high as it has been for the last year and a half or so, it's easy to forget the dark days of 1996-7, when the Spindler regime botched the Sun merger, the Amelio regime made like an ER doc that stabilized the patient and left without continuing the treatment, and Jobs pulled off a palace coup to bring the company back from near-death. The OS strategy that came out of the Next acquisition came on the coattails of the Copland disaster, and had to change a few times for a few reasons. Rhapsody in its original form was essentially rejected by the Mac community, as Java had little credibility and nobody wanted to bother learning Objective-C; this was why Carbon had to be created. Therefore, the Rhapsody that shipped as MacOS X Server was different from the Rhapsody that was originally planned.
And it's flat-out unfair to say that OS X is that far from Rhapsody. Mach is Mach; your statement is equivalent to saying that Linux 1.x isn't really Linux because it has a different version number. What you are looking at are two different incarnations of the same Mac-flavored Unix; OS X has a drastically different interface from Rhapsody, and the driver architecture has been massively rewritten (probably the reason the Intel port is nontrivial, btw), but they're two different flavors of the same product.
Now you want to gut the BSD layer and reimplement a complete Posix-compatible operating system from scratch, then you've got a case.
Rather, look at Rhapsody as a milestone: trust us, you'll get what you wanted, here's the important stuff right up front.
Saying that MS did nothing wrong here is entirely dependent on your point of view. If you are one of those people who accepts the findings of fact in the DOJ case, than what they've done is entirely in line with the usual crap that got them in trouble in the first place.
The thing is that for any other company, it probably *would* be entirely a justifiable business position, but it would get them blown off the map for being pointlessly nonconformist. This ludicrous (and (note: IANAL) most likely indefensible in any reasonable court of law; my copy of the OpenStep spec has a similar if less restrictive license on it and I don't think that stopped the GNUStep people) attempt at information restriction is probably no more than a scare tactic; they figure they have the money to bury the Samba people no matter what the legalities are.
Doesn't Slashdot have any IP lawyers that read this? ACLU people? Anyone with a law education and a half a shot glass of common sense? Comments?
Whoa, boy. No need to hit the igniter switch there.
Darwin is actually considered a fourth major BSD variant by some. What you've said is essentially correct, but it oversimplifies. What Darwin is is the BSD core that NextStep was built on way the hell back in 1986 or so, basically BSD over the Mach microkernel with a built-in Objective-C runtime. In that respect, its closest relative is probably Lites. When NextStep became Rhapsody (i.e. MacOS X Server) Apple rebuilt the BSD server using 4.4BSD and a newer Mach kernel; Darwin came somewhat later as Apple chose to release the low-level parts of the system under an Open Source license.
As for playing with it... Darwin supports pretty much all non-PPC603 PCI PowerMacs (why it doesn't support the 64/6500 series is a mystery to me) and it would be great source material for anyone looking to hack the G4 (are you listening, Jean-Louis Gassee?). The APSL is a bit strange (generally Open Source, but in practice forbids private-only development since Apple claims a right to all derived works), but it's not that restrictive, just a bit overinquisitive. The most useful thing about it is that Darwin 1.0 is essentially the same kernel as MacOS X will use.
The most interesting thing about it is that John Carmack from id Software has done a lot of playing with it and seems to rather like it, even if it's been a moving target since it first became available. Right now it's a very bare-bones system (wait for OS X if you want flash and friendliness), but it's got enough services and apparently it's stable enough that you could use it for quite a few real-world server applications.
Question, though: anyone know where to buy a CD with the Darwin installer on it?
One could argue that this is Intel just trying to show Microsoft just how painless it is to open up your APIs:-)
Anyway, I have to agree with those who have pointed out that Intel is being Open Source friendly. It's a rare company in Intel's position that actually gets around to figuring out on its own that trade secrets aren't all they're cracked up to be. Truthfully, one can't look at Intel as doing anything more than responding to market demands. They know that the installed base for Linux is huge, and it's going to grow. In order to maintain their position in the Linux market so everyone doesn't switch over to PowerPC or Alpha, they have to play the game the same as IBM or SGI are doing. That means swallowing your pride and opening up.
It's becoming a common thing, actually -- Lego brought out their Mindstorms sets for Windows only, but they at least don't crack down on third-party development environments. Texas Instruments found out that so many people were hacking their calculators and programming them in assembly code that when they shipped the TI-83 and TI-86 they broke down and told people how to do it their way. Why? Not because Stallman propaganda is finally getting to them. They just realized that it's suicidal to blow off potential customers just because they aren't doing it through channels.
Now if only the MPAA would figure that out and back off on DVDs...
You know, here's the thing about HTML fonts: they're a bad idea. They screw up a user's control over how a site is supposed to be viewed.
IMHO vanilla HTML is probably the best cross-platform doc standard, followed by PDF if some kind of formatting is necessary (for example, an online version of a published book). PostScript is somewhat outdated for that purpose and should really be lumped in with troff and LaTeX as intermediate formats, not really suitable for public consumption.
Someone moderate that one up as Funny or Insightful or something.
I love the irony here. It's not unlike using Linux to keep old hardware in service, but it's much more insidious because an open-source Divx codec runs precisely counter to the entire intent of the system.
In any case, I don't know that it can be stopped: DMCA allows reverse engineering for interoperability, and this has got to be (IANAL) a crystal-clear case of same. It's not piracy; people who got burned on Divx just want to be able to watch their movies.
May I concur, at the risk of being moderated as Redundant?
The thing is, something like this should probably be made public knowledge -- the main idea being to embarrass the vendor into fixing the bug. Seems to me that it's almost obligatory for the poster of the original comment to bring said exploit to everyone's attention; otherwise you're leaving other users in the lurch wondering if their MTA is the one you've just cracked.
You know, the truth is that I simply don't buy their licensing terms. "Treat it as a trade secret" my USB port -- that's like telling a little kid not to touch the cookie jar when they didn't even know there was something in there.
It's legal FUD, no more. I'd just love to see someone work up the nerve to challenge it (anywhere except Virginia, of course).
The botched link refers to an entry in the Jargon File. The story is that Ken Thompson built a login program for an early version of Unix that enabled a backdoor login under the username "kt", and created a C compiler that would insert that code whenever it recognized that login was being recompiled; in order to insure that the crack would go undetected, there was other code built into the compiler that knew when it was recompiling itself that silently slipped the crack into the new version of the compiler itself. Apparently, it worked pretty well.
The crack only partially made it out of Bell Labs; the login utility escaped, but apparently the compiler never left Murray Hill.
First off, you're stretching it including Squeak and Oberon as OS/Language systems. But be that as it may...
Isn't that what Multics was trying to do with PL/I? Even then, I don't know that it necessarily *has* to be that way. It's sort of an ivory-towerish view: one must eat one's own dogfood, all the time, every time.
I had a conversation recently with someone with some very strong opinions about programming languages (I'm doing a bit of a language project myself at the moment trying to create an applications language in Klingon) and he made the point that a lot of Lisp Machine enthusiasts feel that they got screwed by the Unix crowd. His thing was that Lisp was purely engineered away from actual user experience and therefore is too pure for its own good. Putting aside for a moment whether this is actually a legitimate criticism (which it is in the case of Scheme, but may not be for Common Lisp), he has a good point in the abstract. Ideals exist for a reason, but they're not always very practical.
Unix is the way it is because it works. EVERYTHING is a component, and the plumbing is designed mostly to be human-readable. Unix is not intentionally an elegant system; it is hacks laid on hacks in such a way as to get the job done as efficiently as possible. Why short, cryptic names for commands? Not everyone's a good typist. Why pipes? Interface metaphors before anyone did any research on them.
If you like pretty, there's plenty of pretty out there. I don't knock Smalltalk-80 or Genera or any of those other environments, but in the end you can't shove a technique down someone's throat. The Unix philosophy is "by any means necessary" -- that's why there's so many Perl junkies out there, for example. Pretty is great. It allows someone to take pride in their work and to show off what they're really capable off. The problem with pretty is that it forces your users to think like you. If it's a lousy fit your users won't touch it.
I mentioned my Klingon programming language earlier. I'm writing it in Perl. Why? Because Perl is very openended about what it can do. I get tokenizing almost for free using split() (there are some issues regarding string parsing that I'm ignoring at the moment). There are no perl style police, so I can write my code as slack as I want (and it's pretty slack, believe me) and nobody will care unless I need to speed it up or cut down on memory usage. Someone with a more purist view would probably want me to go with lex'n'yacc, or even write a recursive descent parser, or even do the whole damn thing in assembler (it's rather forthy in style). But Perl works for me. It's not pure (if you've ever read the Llama book even Larry Wall admits in the intro that it's a crufty mess), it's not pretty, but it works.
So as for architectural simplicity, yes. Increased performance is entirely dependent on the implementor, but I suspect it has nothing to do with tying OS to language. But total freedom... not in the least. Look at it this way: you can write a haiku about anything, but if it doesn't have seventeen syllables it's not really a haiku, is it?
Slashdot Mirror
Interesting thought -- for a while at Boston College there were "Computers in Management" (as in "how to sit at your desk and write presentations) classes run in ClarisWorks as the prevailing environment. Didn't last though.
See, MIT's CS majors (I don't know if they still do this, since I never went there) used to learn to program in Scheme. Not a bad idea, given that it makes you think about how to do things.
What I'd love to know is how many "CS" classes amount basically to "how to use Visual C++ on Windows". My attitude is that if you're learning to program, your code had better not be platform dependent.
/Brian
Truly. Plan 9/Brazil ('cos that's what they seem to be calling it internally these days) is quite cool (picture it: a completely distributed operating environment where your desktop screen is pretty much the same whether you're running on a PC or a mainframe or something in between); it's Unix and NFS taken to extremes.
I would love to have played with Plan 9 myself, but hardware costs enough to begin with; I don't want to shell out $300 either for something I'd consider a learning experience (like the reason I once bought a secondhand Newton; I'm a Palm junkie but I wanted to learn). Their best bet if they want to make a splash and break this percieved deadlock is to open it up. Put it under something like a BSD or APSL-type license and give it away at media cost, then see what happens...
/Brian
Nononononono. The Ethernet is used only on the fighters. Anyone can clearly see that the fighter umbilical is FDDI. (And of course the mother ship runs Windows -- don't you know that the reason Jeff Goldblum was using a Mac (apart from the fact that they own him anyway) was that the aliens couldn't backcrack him? Go ahead, moderate me down; my karma is a bit swollen anyway, and this bit wasn't half as funny as I'd hoped... /Brian
Well, read it. The funniest part apart from that is the part where they say that copy protection is important for creativity. The fact is that the entertainment industry as we've known it since Thomas Edison invented the phonograph and (can't quite remember how to spell Edward in Anglo-Saxon) Muybridge invented the motion picture camera is dead and was from the moment MP3 hit the streets. I've recently reluctantly partially come around to the pirates' side, not because I think that what they do is right (I don't; if nothing else it shows a tremendous lack of respect for the creators of the product) but because I think the whole issue is moot. Look around us. The tech is there for everybody to see. For six grand I can get an iMac, a digital camcorder, and a GoVideo double-decker VCR and start my own movie studio. For ten grand, I can make that a PowerMac G4 or a high-end Sony VAIO and a nice big RAID array. For the price of a new sport-utility vehicle, I could build a small Linux-based rendering cluster. You get the idea. I predict that we will never see a monster-budget blockbuster like Titanic again. Within two years, the entire industry will be in anarchy as execs realize that there's no hope of holding the line on the status quo. Star salaries will plummet and production budgets will tailspin as the studios try to hedge their bets against collapse. What will this lead to? Your guess is as good as mine. /Brian
And the thing was that when PPC came down the pike, all of us Mac junkies (or ex-Mac junkies as the case may be; I still like Macs alot but I'm a Linux loyalist these days) were saying, "Cool! Decent game hardware!"
It was; the hardware was never the problem, and a lot of game developers were going to sign on pretty early. But nothing ever really came of it until well after the advent of the iMac because a) Copland crashed and burned when Apple engineers couldn't quite figure out when to stop adding things (and a bloody shame too; I rather wish Apple would release the source code for posterity because what I saw of Copland was pretty damn cool) and b) Apple *almost* crashed and burned under the listless leadership of Michael Spindler cost them their first shot at a unified MacOS/Unix system (the long-forgotten failed Sun merger).
To be honest, I don't really blame John Carmack for being anti-Mac. To this day the MacOS is best thought of as a cluttered attic -- it may in fact be organized, but it's not the best situation and it's long past time to junk it and rebuild from scratch. That's Rhapsody and Carbon.
/Brian
Even better -- Quake (or at least Q1) (as well as Doom, now that I think of it) is open-source. Carmack's got them coming and going -- if they lose him, not only do they shoot themselves in the foot, Quake forks.
And then, we find out if trademarks still matter in the Open Source world.
/Brian
And I suppose you have the same opinion about Dilbert or Rhymes With Orange, two of (IMHO) the most brilliantly funny mainstream comics out there right now. And neither Scott Adams nor Hillary Price is really all that much of an artist.
If anything, User Friendly is much funnier than either of the above. The thing is that it's very in-jokey as well; for those who understand it, it raises the humor value, but for those who don't, it's a turnoff. I'm thinking you mostly just don't get the jokes.
/Brian
Greg was originally blonde; somewhere between the Techie Detox episode and the Microsoft raid he stopped bleaching his hair.
(Or at least that's what I assume happened. It was never actually mentioned in the strip.)
/Brian
Anyone gonna get this one right? BSD is older than Linux by a decade and a half, and there's a lot of story there that it doesn't seem like anyone still remembers...
:-) ). At that point, people started banging on the Unix kernel all over, and the Open Source culture as we know it began.
Okay, flames off. Grab some cocoa and rally 'round, and ol' Uncle Brian gonna tell you about some other guy named Bill and his friends...
Unix didn't really become Unix until John Lions wrote the Lions book (and any Slashdotter who doesn't have a copy should be shaken mercilessly until enough money to either buy or photocopy a copy falls out; about US$30 should do it for a legit copy, and a trip to Kinko's would be much cheaper, not to mention totally appropriate if you know the story
It's easy enough with all the Linux hype these days to lose sight of where it all began, but here's the story: back around '76, when AT&T shipped Unix V6, the University of California at Berkeley got its hands on a source license and its students started hacking away at it. Somewhere along the way, BSD supplanted TENEX and its proprietary brethren as the OS running the backbone of the then-ARPANET; as a result, it became THE wide-area network OS for the eighties. Software such as the BSD TCP/IP stack (still the standard outside the Linux world) and sendmail are still at the core of much of the net, and most of it came out of Berkeley.
Along the line, many of these students struck out on their own (most notably Bill Joy at Sun), and when they went off to build their workstation systems, they took BSD with them, thus bringing about systems such as SunOS and HP/UX, as well as several that are now long forgotten; not only was the BSD license in effect a donation to the public domain, the code itself was something that the engineers themselves had developed, which was more than could be said about the multiple permutations of Unix coming out of Murray Hill, NJ. BSD even came to the PC, in the form of Bill and Lynne Jolitz' 386BSD, a precursor to the modern BSD designs.
The modern BSD era began in 1994, when UC-Berkeley's Computer Science Research Group, disbanded a few years earlier, regrouped to finish the BSD story at Berkeley. The release they made was called 4.4BSDlite, and was actually quite widely available in bookstores at the time. It consisted of a CD-ROM containing a mostly-complete AT&T-free source tree (with a few unreplaced but critical AT&T files removed) and four books containing man pages and other docs.
This event is what created two of the four modern BSD variants (FreeBSD and netBSD); OpenBSD was a fork off of netBSD that stressed security concerns, while the newest major variant, Darwin/MacOS X, came out of Apple's post-acquisition merging of NeXT's Mach/BSD hybrid OS and the 4.4lite codebase. BSD was also the base for OSF/1, the only major variant of which is Compaq's Tru64 Unix.
Today, I think BSD is viewed as a one-off, a newcomer in a world dominated by Linux. I suspect most of the history I've just related simply isn't all that well known anymore, and it certainly isn't media-friendly (what, no underdog-makes-good? No flashy technology? No David-vs.-Goliath? HACKERS?! Not interested, sorry...). This is a shame; that's why I'm posting this.
I think the reason that BSD persists today (apart from having about the best TCP/IP stack available) is because it's a link to the past that most commercial Unices haven't preserved. By running a BSD, you're expressing solidarity with some random hacker poring over a third-generation photocopy of the Lions book on a brand-new VAX twenty years ago. And hey, I prefer penguins to daemons myself, but that's a connection that we Linux users can't make.
Saying that BSD is related to Linux is wrong on many levels. Apart from both being open-source Unices, there is no real connection. Linux is its own beast, and I think over the long term it will eventually evolve away from its Unixness somewhat. Minix, now freed from its old copyright restrictions, will probably morph into an embedded-systems OS; I'd love to see it on a WinCE (MinCE?) unit, for example, since it's just about the perfect size. But BSD will still be carrying the Free Unix torch years from now. Long may it.
/Brian
An interesting, and not entirely uncalled-for comment...
I've seen the movie myself. I look upon Truffaut's version of F451 as being a mixed bag: generally a tremendous waste of time and a massive perversion of Bradbury's rather gritty, all-too-real plot with a few strokes of absolute genius thrown in to keep it from being a total failure.
The bad: Wow. Too much to mention. Men in jetpacks instead of the six-legged Hound. Name changes. Wooden acting. Absurd direction. This beast should have been burned along with the books. The closing scene, where the human books gather, is disgustingly maudlin. The plot was streamlined and stylized into unrecognizability (including the removal of Guy Montag's first name). The production as a whole was a soporific, entirely missing the lurching, gut-wrenching action-lull pace that Bradbury established in the book.
The good: Surprisingly, a few little touches that would have perked up the book had Bradbury thought of them. The opening credits were done as a voice-over (appropriate for such an aggressively post-literate setting). A few others that don't come to mind.
Oh, well. Maybe someday someone will come along who actually understands SF and will do F451 the way Bradbury actually wrote it.
/Brian
Interesting point. I had been trying to figure out what the name meant ever since I saw it a few days ago. I still don't particularly like it (it would be almost as difficult to take seriously as Linus's coinage Freax (fre-ax?)), but as a development name, if that's really the rationale behind it, it works pretty well.
As for the OS itself, I like the idea. POSIX-compliance is pretty much a necessity, but more power to the author if he can create a mainstream non-Unix OS by the same model Linus did. I don't know if I'll ever get around to running AtheOS, but I'm thinking I'll be following it for a while...
/Brian
Those who do not remember the past are doomed to forget that Donkey Kong was originally the bad guy...
The AC I'm responding to has neglected one of the key aspects of gaming: gameplay beats flashy tech any day. Granted, Q3 and the like are visually amazing. Granted, the closest anyone ever got to Lara Croft in the days of sidescrollers was Samus in Metroid (after you beat the game and figured out she was a she after all).
But you know, all those old games... they were pretty damn fun to play. It's true, the quality isn't that great -- 8 bit chips aren't known for their 3D graphics performance -- but that doesn't mean that they can't keep up with today's best. Frankly, I've seen Super Mario 64, and I don't think it really measures up to the original. Why? Because the original was a side-scroller, whereas SM64 is a third-person 3D game. It actually sort of spoils the effect, not to mention making it much harder to adapt to the new controls.
As for demanding much higher quality games... see what I said above about play vs. tech. There's a company called Ambrosia that has been making some of the best shareware games available for the MacOS for years now, starting with an amazing Asteroids clone called Maelstrom. The best example of what I'm talking about: they redid Centipede/Millipede as Apeiron. The graphics and sound, it is true, are drastically enhanced, and Apeiron has features such as powerups that neither original ever had... but it's the same game. It plays very close to the original, enough to satisfy anyone's Centipede jones. Compare that to the modern Centipede, which is 3-dimensional, graphically gorgeous... and just plain wrong.
You can sit around playing your amped-up Quake games all you want. I'll appreciate it as much as you do. But don't tell me that just because a game is played on 20-year-old hardware it doesn't meet modern quality standards. If Ms. PacMan is still fun to me after fifteen or more years, than I can't legitimately say it's not cool, or tricky, or enjoyable no matter what "quality standards" it does or does not meet.
/Brian
Actually, that's exactly what happened. Rhapsody/MacOS X Server is NextStep with a Mac interface and the Blue Box (Classic environment) added.
What was going on when Apple bought out NeXT was that Apple needed a new OS and Copland had developed a case of terminal second-system effect that ultimately killed it (and Gershwin, its follow-on). After the Apple/Sun merger collapsed (most likely thanks to Michael Spindler's incompetence), Gil Amelio took over and a bit later made the NeXT buy. Rhapsody hit the radar screen almost immediately.
Rhapsody was originally just the Macified NextStep; new apps were to be written in a Javafied (though you could still use ObjC if you wanted) version of OPENSTEP (now called Cocoa). The developer community, however, by and large rejected Rhapsody in that form, and Apple was forced to backtrack and create Carbon, all the while making subtle changes in what would become MacOS X Server (Apple's third version of Mac Unix, after A/UX and MkLinux, for those of you keeping track at home).
For the record, I find the comment of BSD and Mach being inseparable to be rather funny -- you can't do anything useful by separating them, certainly, but it's a gross oversimplification (like, has anyone considered trying to hack Darwin to run MkLinux on the same microkernel?).
/Brian
Sort of. I think that Rhapsody was originally built from 4.4BSD-Lite over Mach.
/Brian
Whoa. Mindblowing.
Just out of curiosity, what is a "commercial document" anyway? If I had Office Academic and I start writing a book, does that mean that I had to pony up to publish it legally?
I'm sorry, licenses that control content creation don't wash with me. I believe one of Microsoft's better-known legal abuses in the Chicago days involved a clause in the license for MFC that made it rather complicated for anyone to write something like a word processor or spreadsheet with MFC because it competed directly with MS. I can't possibly see how this can be considered legit -- it's impossible to enforce and it seems to be in rather bad faith. The funny thing is that MS did this in a market where there were plenty of VC++ alternatives out there...
/Brian
(For convenience -- and since I'm a longtime Mac man who followed this whole scene pretty closely -- I will refer to OS X Server as Rhapsody.)
It's still a fair comparison -- OS X Server/Rhapsody is sort of a transitional form, maybe a Homo ergaster version of the Homo sapiens MacOS X final.
With Apple flying as high as it has been for the last year and a half or so, it's easy to forget the dark days of 1996-7, when the Spindler regime botched the Sun merger, the Amelio regime made like an ER doc that stabilized the patient and left without continuing the treatment, and Jobs pulled off a palace coup to bring the company back from near-death. The OS strategy that came out of the Next acquisition came on the coattails of the Copland disaster, and had to change a few times for a few reasons. Rhapsody in its original form was essentially rejected by the Mac community, as Java had little credibility and nobody wanted to bother learning Objective-C; this was why Carbon had to be created. Therefore, the Rhapsody that shipped as MacOS X Server was different from the Rhapsody that was originally planned.
And it's flat-out unfair to say that OS X is that far from Rhapsody. Mach is Mach; your statement is equivalent to saying that Linux 1.x isn't really Linux because it has a different version number. What you are looking at are two different incarnations of the same Mac-flavored Unix; OS X has a drastically different interface from Rhapsody, and the driver architecture has been massively rewritten (probably the reason the Intel port is nontrivial, btw), but they're two different flavors of the same product.
Now you want to gut the BSD layer and reimplement a complete Posix-compatible operating system from scratch, then you've got a case.
Rather, look at Rhapsody as a milestone: trust us, you'll get what you wanted, here's the important stuff right up front.
/Brian
Saying that MS did nothing wrong here is entirely dependent on your point of view. If you are one of those people who accepts the findings of fact in the DOJ case, than what they've done is entirely in line with the usual crap that got them in trouble in the first place.
The thing is that for any other company, it probably *would* be entirely a justifiable business position, but it would get them blown off the map for being pointlessly nonconformist. This ludicrous (and (note: IANAL) most likely indefensible in any reasonable court of law; my copy of the OpenStep spec has a similar if less restrictive license on it and I don't think that stopped the GNUStep people) attempt at information restriction is probably no more than a scare tactic; they figure they have the money to bury the Samba people no matter what the legalities are.
Doesn't Slashdot have any IP lawyers that read this? ACLU people? Anyone with a law education and a half a shot glass of common sense? Comments?
/Brian
Whoa, boy. No need to hit the igniter switch there.
Darwin is actually considered a fourth major BSD variant by some. What you've said is essentially correct, but it oversimplifies. What Darwin is is the BSD core that NextStep was built on way the hell back in 1986 or so, basically BSD over the Mach microkernel with a built-in Objective-C runtime. In that respect, its closest relative is probably Lites. When NextStep became Rhapsody (i.e. MacOS X Server) Apple rebuilt the BSD server using 4.4BSD and a newer Mach kernel; Darwin came somewhat later as Apple chose to release the low-level parts of the system under an Open Source license.
As for playing with it... Darwin supports pretty much all non-PPC603 PCI PowerMacs (why it doesn't support the 64/6500 series is a mystery to me) and it would be great source material for anyone looking to hack the G4 (are you listening, Jean-Louis Gassee?). The APSL is a bit strange (generally Open Source, but in practice forbids private-only development since Apple claims a right to all derived works), but it's not that restrictive, just a bit overinquisitive. The most useful thing about it is that Darwin 1.0 is essentially the same kernel as MacOS X will use.
The most interesting thing about it is that John Carmack from id Software has done a lot of playing with it and seems to rather like it, even if it's been a moving target since it first became available. Right now it's a very bare-bones system (wait for OS X if you want flash and friendliness), but it's got enough services and apparently it's stable enough that you could use it for quite a few real-world server applications.
Question, though: anyone know where to buy a CD with the Darwin installer on it?
/Brian
One could argue that this is Intel just trying to show Microsoft just how painless it is to open up your APIs :-)
Anyway, I have to agree with those who have pointed out that Intel is being Open Source friendly. It's a rare company in Intel's position that actually gets around to figuring out on its own that trade secrets aren't all they're cracked up to be. Truthfully, one can't look at Intel as doing anything more than responding to market demands. They know that the installed base for Linux is huge, and it's going to grow. In order to maintain their position in the Linux market so everyone doesn't switch over to PowerPC or Alpha, they have to play the game the same as IBM or SGI are doing. That means swallowing your pride and opening up.
It's becoming a common thing, actually -- Lego brought out their Mindstorms sets for Windows only, but they at least don't crack down on third-party development environments. Texas Instruments found out that so many people were hacking their calculators and programming them in assembly code that when they shipped the TI-83 and TI-86 they broke down and told people how to do it their way. Why? Not because Stallman propaganda is finally getting to them. They just realized that it's suicidal to blow off potential customers just because they aren't doing it through channels.
Now if only the MPAA would figure that out and back off on DVDs...
/Brian
You know, here's the thing about HTML fonts: they're a bad idea. They screw up a user's control over how a site is supposed to be viewed.
IMHO vanilla HTML is probably the best cross-platform doc standard, followed by PDF if some kind of formatting is necessary (for example, an online version of a published book). PostScript is somewhat outdated for that purpose and should really be lumped in with troff and LaTeX as intermediate formats, not really suitable for public consumption.
/Brian
Someone moderate that one up as Funny or Insightful or something.
I love the irony here. It's not unlike using Linux to keep old hardware in service, but it's much more insidious because an open-source Divx codec runs precisely counter to the entire intent of the system.
In any case, I don't know that it can be stopped: DMCA allows reverse engineering for interoperability, and this has got to be (IANAL) a crystal-clear case of same. It's not piracy; people who got burned on Divx just want to be able to watch their movies.
Go with this, folks...
/Brian
May I concur, at the risk of being moderated as Redundant?
The thing is, something like this should probably be made public knowledge -- the main idea being to embarrass the vendor into fixing the bug. Seems to me that it's almost obligatory for the poster of the original comment to bring said exploit to everyone's attention; otherwise you're leaving other users in the lurch wondering if their MTA is the one you've just cracked.
/Brian
You know, the truth is that I simply don't buy their licensing terms. "Treat it as a trade secret" my USB port -- that's like telling a little kid not to touch the cookie jar when they didn't even know there was something in there.
It's legal FUD, no more. I'd just love to see someone work up the nerve to challenge it (anywhere except Virginia, of course).
/Brian
The botched link refers to an entry in the Jargon File. The story is that Ken Thompson built a login program for an early version of Unix that enabled a backdoor login under the username "kt", and created a C compiler that would insert that code whenever it recognized that login was being recompiled; in order to insure that the crack would go undetected, there was other code built into the compiler that knew when it was recompiling itself that silently slipped the crack into the new version of the compiler itself. Apparently, it worked pretty well.
The crack only partially made it out of Bell Labs; the login utility escaped, but apparently the compiler never left Murray Hill.
/Brian
First off, you're stretching it including Squeak and Oberon as OS/Language systems. But be that as it may...
Isn't that what Multics was trying to do with PL/I? Even then, I don't know that it necessarily *has* to be that way. It's sort of an ivory-towerish view: one must eat one's own dogfood, all the time, every time.
I had a conversation recently with someone with some very strong opinions about programming languages (I'm doing a bit of a language project myself at the moment trying to create an applications language in Klingon) and he made the point that a lot of Lisp Machine enthusiasts feel that they got screwed by the Unix crowd. His thing was that Lisp was purely engineered away from actual user experience and therefore is too pure for its own good. Putting aside for a moment whether this is actually a legitimate criticism (which it is in the case of Scheme, but may not be for Common Lisp), he has a good point in the abstract. Ideals exist for a reason, but they're not always very practical.
Unix is the way it is because it works. EVERYTHING is a component, and the plumbing is designed mostly to be human-readable. Unix is not intentionally an elegant system; it is hacks laid on hacks in such a way as to get the job done as efficiently as possible. Why short, cryptic names for commands? Not everyone's a good typist. Why pipes? Interface metaphors before anyone did any research on them.
If you like pretty, there's plenty of pretty out there. I don't knock Smalltalk-80 or Genera or any of those other environments, but in the end you can't shove a technique down someone's throat. The Unix philosophy is "by any means necessary" -- that's why there's so many Perl junkies out there, for example. Pretty is great. It allows someone to take pride in their work and to show off what they're really capable off. The problem with pretty is that it forces your users to think like you. If it's a lousy fit your users won't touch it.
I mentioned my Klingon programming language earlier. I'm writing it in Perl. Why? Because Perl is very openended about what it can do. I get tokenizing almost for free using split() (there are some issues regarding string parsing that I'm ignoring at the moment). There are no perl style police, so I can write my code as slack as I want (and it's pretty slack, believe me) and nobody will care unless I need to speed it up or cut down on memory usage. Someone with a more purist view would probably want me to go with lex'n'yacc, or even write a recursive descent parser, or even do the whole damn thing in assembler (it's rather forthy in style). But Perl works for me. It's not pure (if you've ever read the Llama book even Larry Wall admits in the intro that it's a crufty mess), it's not pretty, but it works.
So as for architectural simplicity, yes. Increased performance is entirely dependent on the implementor, but I suspect it has nothing to do with tying OS to language. But total freedom... not in the least. Look at it this way: you can write a haiku about anything, but if it doesn't have seventeen syllables it's not really a haiku, is it?
/Brian