I saw this and I knew someone would bring this up.
Actually, this is precisely why I created var'aq in the first place -- to investigate what a computer language created by a non-English speaker might look like.
An awful lot like PostScript, it turns out, if you're a Klingon...
One of my coauthors, Chris Pressey (see www.catseye.mb.ca if you're not familiar with his work) has even suggested trying to create a language with an inflected grammar like Latin (I've even thought of basing one on Latin; its working name is SPQL, but I don't know if I'll ever get around to creating it). Imagine what that parser would look like -- you sure as hell couldn't do it easily in yacc.
But I'm glad someone brought it up -- I was going to, but hey, if someone else is tooting my horn...
So do being able to answer those questions make you a CCCP (Caldera Certified CP/M Person):-)
Actually, it depends on what you consider CS. At Boston College, there are two CS degrees. The A&S degree handles a lot of heavy-duty theoretical stuff (including Algorithms, the one class I couldn't handle:-( ), while the CSOM (Caroll School of Management) degree is based more on real-world environments. They overlap, but they're not identical.
I once tried to begin an open-source project with only an idea, no code in hand. It died from lack of direction with out line one of code ever being written.
Don't do that.
Instead what you do is do what Linus did -- get something out there first, and *then* open it up. You'll get a lot further that way.
What I find rather strange is that even though Mosaic has been sitting around basically embalmed for the last four years the licensing restrictions still haven't changed. I don't really get this. Does this have something to do with Microsoft's licensing/theft of Spyglass code, or is it simply inertia?
Actually, I have to say that/. looks much better in old-style HTML; the only real problem is that the primitive HTML formatting tends to make it look a bit garbled up top.
Actually, I might just add a graphic to my own web pages... they're almost pure text and won't break on anything at all (though the odd graphic is a jpeg). Mosaic-enhanced? I wish.../Brian
Now the big question -- anyone got MacOS X Server or GNUStep to bring this into the 21st century? I got a website that would love to have a functional copy of the very first browser.
How about this? Let's NOT close down the crack house down the block. Okay, so the hookers and guns are dragging down the neighborhood and keeping everyone inside, but the dealers have a right to express themselves by selling crack.
The fact is that it was not a free speech issue! The ab.warez.* groups have no legitimate use -- they are explicitly for the illegal copying of software created by people who do not wish to have their software copied in such a manner. Anything that can be legally transmitted by such a forum is by definition off topic. Anything that is on topic is illegal.
IANAL, but it seems that this is not a moral issue, it's a legal one.
I'd sort of sum it up like this: Because it's there.
Essentially, by using X you get the same thing Apple got when they bought NeXT or all those workstation companies got when they decided to use BSD instead of writing their own OSes -- a stable, tested product. X has been around damn near forever. Yes, perhaps it's overkill to an extent, but X works.
Now the real question is whether X is the right tool for the job, but I don't think that's much of an issue either. X is a GUI toolkit, not a GUI per se. The real issue is what *lcewm (hypothetical LinuxCE window manager) will look like when it's done.
Apparently because Adobe controls the standard; GNU considers PostScript and PDF proprietary, despite the existence of freeware like Ghostscript.
I have mixed feelings about that; on the one hand, it is an Adobe trademark (at least Acrobat is). On the other hand, Adobe is quite explicit about allowing non-Adobe implementations of both technologies. That, IMHO, makes it as open a standard as any out there.
A few years back I tried to convince Boston College to stop carrying the alt.binaries.warez.* groups on their Usenet server because there is simply no legitimate purpose for those groups. I still think I was in the right, even though some criticized me for it and the BC legal department wimped out because of an alleged free-speech issue.
I don't know if I'd do the same thing today, since Gnutella and FreeNet have rendered the piracy issue moot. But if RomNet is really Napster for video game ROMs, it's illegal as all hell.
Good or bad, I can't say. There is a case to be made for a myMP3.com-type model, but the console companies will never let it happen (just ask Sony and Nintendo, who truly despise the emulator market even though their hardware is probably loss-leadered).
You know, a few people have pointed at Galeon as a smaller Mozilla. I looked at the website (http://galeon.sourceforge.com). Just wanted to say I thought the project's heart was in the right place, but creating a GPL-MozPL crossbreed is going to make damn sure that the project never makes a commercial distribution -- what's the point of offering a smaller browser if you need the bigger browser installed to run it?
Here's the thing about Mozilla, see. It's open source. You don't like it, you download the freakin' source code and change it.
I don't know about anyone else, but I'm sick and tired of people bitching and whining that Mozilla sucks. Okay, it sucks. But there's a lot of goodies in there -- a Javascript engine, an HTML renderer, lots of stuff like that. You don't like Mozilla M16? Y'all do yourselves a favor, get the Mozilla code, and WRITE YOUR OWN BROWSER!
I have a problem with people who complain about situations like this but don't do anything. Me, I don't use Mozilla so I have nothing to complain about. Okay, a lot of you are busy people with lives and paychecks to worry about. You can't do anything. And not everyone with a vested interest in this project is going to be able to contribute anything more than the occasional skin (not that I think that particular feature is a terribly good idea); we're not all hackers here. But dammit, if you don't like it and you want to see some real free browser competition, you have absolutely no right to complain if you have the power to do something about it. That is why we have open source code, and that's why Netscape did this with Mozilla in the first place.
You non-programmers, you people with lives, you can't do any more than talk about the problems. The source package is huge, and a lot to wade through. But I find it outrageous that people walk around and bitch about how unstable and slow the code is and yet make no effort to even provide moral support. Mozilla is the ONLY significant standalone open-source browser out there. kfm is minimal, Opera is purely commercial, and everything else (Amaya, Arena, HotJava, Mosaic) is either dead or still experimental.
WE NEED MOZILLA TO SUCCEED. Reminding people of things like the vaporware issue (a nonsequitur since Open Source vaporware is an oxymoron), the declining market share, and the sluggishness of the *current implementation* is bullshit. If you can't contribute, tell the Mozilla people your problems. If you can, shut your trap and start coding.
Whew. Feel better now.
After that little jag, I'll put my money where my mouth is if the world wants me to. That Lions Book for Mozilla -- if someone's willing to pay me to write it (or at least coordinate it), I'd be happy to (pay because I can't afford to volunteer my time). As for the AOL issue, if there's a real feeling that AOL isn't doing right by its own software, why haven't we seen a Mozilla fork yet?
/Brian
Re:Another Industrial Design Coup for Apple
on
Apple Cube Confirmed
·
· Score: 1
I think you're rather jumping the gun...
Dylan first. Dylan was a pretty nice language, sure, but it's very much in the long tradition of companies trying to get people to use their languages instead of commodity products (witness C# from Microsoft... hell, even Java...). The fact is that Dylan was never necessary and was just too different from what most people used.
The Newton. Well, I have a Newton MP120 (v1.2 OS), and I think I can tell you pretty much everything that went wrong with it. Newton broke too far out of the computer paradigm -- it tried to be an electronic notebook and wound up creating an interface that was at best confusing. The trusted-compiler model that NewtonScript used was no help; until very late in the Newton's life cycle it was more or less impossible to do a direct port. Newton as an architecture was locked up tight, just like WinCE is now.
Newton descended into self-parody towards the end anyway -- as Palm was showing the world that PDAs didn't need to be able to do everything to do their job and reducing hardware size and cost accordingly, Newtons were getting more expensive and remaining clumsy. The eMate was workable, even innovative. The $1000 MessagePad 2100, still the size of a large paperback and now packaged with a keyboard the same size, was not.
The Newton was innovative, and despite the crap it's gotten had the best handwriting recognition in the business. But it simply wasn't the tool it could have been. I don't blame Apple for that; they had to write the rules as they went along when they invented the PDA. But I do blame them for not keeping pace.
Quickdraw GX... well, that IMHO was just poorly organized. It wasn't cross-platform, and it was really a Copland feature that was brought out far too early. I rather miss it myself, but Apple simply didn't push hard enough.
The real tragedies in Apple's story are OpenDoc (why can't they opensource the wreckage and let the market decide whether it can use it?) and Copland (same deal). To those who don't remember, Apple was creating great hardware in those days (even in these days where G4s are going dual-processor standard I'd still love to have a 6500 in the house), but their software was lagging badly. It nearly killed them.
The simple fact is that all the projects above were draining Apple where it counted. Killing Newton was probably a bad idea, but Newton had failed to evolve. OpenDoc was simply getting buried because developers weren't quite committing to it. QDGX was too narrow. Dylan was cool, but new languages are a toy -- I'm creating a programming language right now and I can honestly say I don't expect it to be all that serious...
This may be a bit redundant, but I'll try anyway...
1. Maybe, maybe not. You're assuming that commercial open-source == abandonment; if that's so why did Apple go with Darwin?
They're striking out into new territory either way. If they are copping out as you say, than they're doing the responsible thing and making sure that Star Office stays in the public record, to make a reappearance later. Otherwise, they're doing what Netscape tried to do with Mozilla and indeed leveraging outside talent to make the project better.
2. This is simply flat-out wrong. If anything, as long as StarOffice provides something that people want, open-sourcing it will ensure that it goes wherever it's wanted.
And in any case, you talk about magnitude of cross-platform software... oy. Linux. *BSD. Apache. Perl. gcc. Maybe you're talking end-user apps -- well, no one has really tried yet. Mozilla is still not yet out of the starting gates. StarOffice will be the real litmus test.
3. If things are as gloomy as you say, forking StarOffice is the best thing that could possibly happen. If StarOffice does fork, that means two things:
-Someone considers it worth keeping around. -Someone thinks they can do it better than Sun.
IMHO that would actually be a credibility boost -- someone thinks it's worth the trouble to improve. The key would be finding a solid corporate backer for said fork.
4. That's because we don't have enough Mac geeks in the Open Source world, IMHO. The Mac world (the Palm community as well) is very much a pay-for-play world; shareware is the dominant form of redistributable software, and the really successful shareware authors are very, very good (try an Ambrosia game sometime). More than anyone else, Mac people are very sensitive to crappy interface design because it's such a major part of the experience. Windows people aren't quite so picky because Windows had no consistent design philosophy until Win95 came out, and Linux (geek OS that we all know it is) is the Unix world's first real splash in the consumer market in its thirty-some-odd years of existence. But the Mac hackers know very much what they're doing, and most of them want to get paid.
Figure this. The hardware in an N64 is basically a baby Indy, circa 1996 or so. That's still pretty respectable horsepower even today, at least outside the Wintel world. It's been cracked pretty wide open, it seems.
Two ideas, one hardware, one software. Hardware, an N64 cartridge with an Ethernet port and a USB adapter for the controller port. Software, Linux N64 (could in theory be based of the existing MIPS/SGI port, no?).
Okay, now, everyone repeat after me...
"I'd love to see a Beowulf of those things!"
Think about it... heavy-duty processing power for US$100 a box, and you could fit about eight of them on one shelf on a standard rack. You'd spend more money on the hubs than the CPUs...
My personal experience, before having read any of the four hundred or so responses already under this thread...
I learned to program on a Commodore 128, in MSBASIC 7.0, one manifestation of the prototypical policy=implementation environment. It wasn't all that bad (structured programming constructs had been backhacked into the language, but that didn't mean much to a twelve-year-old trying to figure out how to write an operating system in BASIC), but to learn to code on such an environment is very limiting. There is no hand-holding outside the manual, but you can't get lost easily (short of a rogue POKE) so it doesn't matter. It wasn't for a very long time that I understood the benefits of keeping the implementation separate from the environment. This is what has made Unix more successful than many other environments of its generation.
The beauty of Unix is that implementation doesn't matter -- you have your choice of tools, and you're free to use whatever you're comfortable with. You don't have to write every last piece of code in bash or jump through hoops to do otherwise. Linux goes traditional Unix one better by traditionally offering a far richer development environment than any other system on the market. Not only do you have a solid, time-tested C compiler (gcc), you have any number of other languages like Perl (my personal language of choice), Python, Fortran, etc...
Another aspect of GUI systems is the IDE. I used to think IDEs were the greatest idea until I switched my Mac environment from Think to CodeWarrior. Now CodeWarrior is pretty damn good, and AppleScripting gives you every bit as much flexibility as shell scripting and then some, but to me it's not particularly obvious how to construct a tool chain in a project build -- there is a learning curve that simply doesn't exist in a traditional shell environment.
I don't like make(1) very much, I'll admit, but there's a certain simplicity to using componentized tools that just isn't there when you're using modulomonoliths like CodeWarrior or Visual Studio. Fast turnaround tools are much more easily managed under Unix, as well; I don't think I could meet the demands of the constant fine-tuning my current language project requires if I wasn't writing it in Perl (or maybe Python), and as good a tool as MacPerl might be it's still a GUI app.
So what's so great about Linux? The flexibility. Yes, there's a learning curve. But it's worth it just to not have to worry about the complications involved in managing a full-up project file for a simple program.
Good idea, except for one thing: isn't Marathon (1 and 2, anyway) open-source?
Think about it: Microsoft gets the talent, but Bungie offloads the hot games. Why doesn't it worry about Marathon? Because technically it no longer controls it.
Keeping in mind, of course, that right or wrong there was good reason for Jackson to be a bit slanted. After all, he was the judge for the '95 consent decree that MS urinated on so freely, and it's tough to keep a straight face when the most powerful company in the world is blatantly, obviously lying.
Now for that right or wrong part: maybe Jackson should have kept a straight face, and that could hurt. IMHO MS won't do all that well in Appeals, but they might wind up doing better than they would have because of that...
Just out of curiosity, what does the Linux body count look like these days anyway? Solaris x86 seems to be universally considered a joke, SCO doesn't seem to be all that far from throwing in the towel, and Irix is circling the drain...
I do agree that Solaris/SPARC is in it for the long haul, and NextStep/MacOS X is just getting out of the gates. SCO is in an odd situation, though; keepers of the flame, yet their bread-and-butter is possibly the most marginal of the major unices out there.
That's a good point, about Unix. Saying it was "designed" for anything is giving Ken and Dennis entirely the wrong type of credit.
Actually, a friend of mine had an interesting insight when I discussed with him a project I was doing on my own regarding programming languages like Lisp. His point was that the reason elegantly-designed solutions like these don't necessarily take off is that they're purely engineered. The reason I like hacking in Perl is that it's designed (loosely speaking:-) ) around how natural language works. It's evolved by being used. Lisp, particularly Scheme, didn't do that; Scheme is sort of a hothouse flower, very pretty, but something of a toy language outside its intended purpose (teaching computer science students how to hack).
Read the book. It's a rather interesting ramble about user interface; Linux Journal slammed it for not really sticking to its message, but I'm not entirely certain it was supposed to have one in the first place.
Slashdot Mirror
I saw this and I knew someone would bring this up.
Actually, this is precisely why I created var'aq in the first place -- to investigate what a computer language created by a non-English speaker might look like.
An awful lot like PostScript, it turns out, if you're a Klingon...
One of my coauthors, Chris Pressey (see www.catseye.mb.ca if you're not familiar with his work) has even suggested trying to create a language with an inflected grammar like Latin (I've even thought of basing one on Latin; its working name is SPQL, but I don't know if I'll ever get around to creating it). Imagine what that parser would look like -- you sure as hell couldn't do it easily in yacc.
But I'm glad someone brought it up -- I was going to, but hey, if someone else is tooting my horn...
/Brian
So do being able to answer those questions make you a CCCP (Caldera Certified CP/M Person) :-)
:-( ), while the CSOM (Caroll School of Management) degree is based more on real-world environments. They overlap, but they're not identical.
Actually, it depends on what you consider CS. At Boston College, there are two CS degrees. The A&S degree handles a lot of heavy-duty theoretical stuff (including Algorithms, the one class I couldn't handle
/Brian
I once tried to begin an open-source project with only an idea, no code in hand. It died from lack of direction with out line one of code ever being written.
Don't do that.
Instead what you do is do what Linus did -- get something out there first, and *then* open it up. You'll get a lot further that way.
/Brian
What I find rather strange is that even though Mosaic has been sitting around basically embalmed for the last four years the licensing restrictions still haven't changed. I don't really get this. Does this have something to do with Microsoft's licensing/theft of Spyglass code, or is it simply inertia?
/Brian
Actually, I have to say that /. looks much better in old-style HTML; the only real problem is that the primitive HTML formatting tends to make it look a bit garbled up top.
Actually, I might just add a graphic to my own web pages... they're almost pure text and won't break on anything at all (though the odd graphic is a jpeg). Mosaic-enhanced? I wish... /Brian
Now the big question -- anyone got MacOS X Server or GNUStep to bring this into the 21st century? I got a website that would love to have a functional copy of the very first browser.
http://www.geocities.com/connorbd/dustydeck
/Brian
This ain't flamebait; it isn't even a good troll.
Let's see -- unnecessary political yammering, guilt by association, uninformed claptrap...
Oy.
/Brian
(Hey, at least I signed my name, grits boy...)
Not a bad idea -- can you do it with those oddball IR links I see on the top of lots of cell phones these days? :-) /Brian
Maybe. But if he's going to do that he ought to think about using all the statistics on the page to prove it.
The truth is, Fred's story seems like nothing more or less than a mass-media-scale troll. And not a very good one -- he played it far too straight.
Actually, it seems that's all I have to say on that point...
/Brian
mustn't say it... mustn't provoke my inner troll... GAAAH! HOW ABOUT A BEOWULF OF THOSE?!?!?!
Okay, got that out of my system.
Actually, a Beowulf wouldn't be a terribly good idea because of budget issues, but if you could find a grant you could get away with it.
How about something like a hardware emulator? You could try Knuth's MIX or MMIX systems, or maybe design your own or copy an existing chip.
How about hacking into Minix? The code is under a BSD license now and anyone can do anything with it.
How about getting SCO's Ancient Unix license and porting Unix V6/V7 to modern hardware?
How about some Lego robotics?
Just a few ideas...
/Brian
How about this? Let's NOT close down the crack house down the block. Okay, so the hookers and guns are dragging down the neighborhood and keeping everyone inside, but the dealers have a right to express themselves by selling crack.
The fact is that it was not a free speech issue! The ab.warez.* groups have no legitimate use -- they are explicitly for the illegal copying of software created by people who do not wish to have their software copied in such a manner. Anything that can be legally transmitted by such a forum is by definition off topic. Anything that is on topic is illegal.
IANAL, but it seems that this is not a moral issue, it's a legal one.
/Brian
I'd sort of sum it up like this: Because it's there.
Essentially, by using X you get the same thing Apple got when they bought NeXT or all those workstation companies got when they decided to use BSD instead of writing their own OSes -- a stable, tested product. X has been around damn near forever. Yes, perhaps it's overkill to an extent, but X works.
Now the real question is whether X is the right tool for the job, but I don't think that's much of an issue either. X is a GUI toolkit, not a GUI per se. The real issue is what *lcewm (hypothetical LinuxCE window manager) will look like when it's done.
/Brian
Apparently because Adobe controls the standard; GNU considers PostScript and PDF proprietary, despite the existence of freeware like Ghostscript.
I have mixed feelings about that; on the one hand, it is an Adobe trademark (at least Acrobat is). On the other hand, Adobe is quite explicit about allowing non-Adobe implementations of both technologies. That, IMHO, makes it as open a standard as any out there.
/Brian
RomNet. Oy.
A few years back I tried to convince Boston College to stop carrying the alt.binaries.warez.* groups on their Usenet server because there is simply no legitimate purpose for those groups. I still think I was in the right, even though some criticized me for it and the BC legal department wimped out because of an alleged free-speech issue.
I don't know if I'd do the same thing today, since Gnutella and FreeNet have rendered the piracy issue moot. But if RomNet is really Napster for video game ROMs, it's illegal as all hell.
Good or bad, I can't say. There is a case to be made for a myMP3.com-type model, but the console companies will never let it happen (just ask Sony and Nintendo, who truly despise the emulator market even though their hardware is probably loss-leadered).
/Brian
You know, a few people have pointed at Galeon as a smaller Mozilla. I looked at the website (http://galeon.sourceforge.com). Just wanted to say I thought the project's heart was in the right place, but creating a GPL-MozPL crossbreed is going to make damn sure that the project never makes a commercial distribution -- what's the point of offering a smaller browser if you need the bigger browser installed to run it?
/Brian
Folks, call me a karma whore, but...
Here's the thing about Mozilla, see. It's open source. You don't like it, you download the freakin' source code and change it.
I don't know about anyone else, but I'm sick and tired of people bitching and whining that Mozilla sucks. Okay, it sucks. But there's a lot of goodies in there -- a Javascript engine, an HTML renderer, lots of stuff like that. You don't like Mozilla M16? Y'all do yourselves a favor, get the Mozilla code, and WRITE YOUR OWN BROWSER!
I have a problem with people who complain about situations like this but don't do anything. Me, I don't use Mozilla so I have nothing to complain about. Okay, a lot of you are busy people with lives and paychecks to worry about. You can't do anything. And not everyone with a vested interest in this project is going to be able to contribute anything more than the occasional skin (not that I think that particular feature is a terribly good idea); we're not all hackers here. But dammit, if you don't like it and you want to see some real free browser competition, you have absolutely no right to complain if you have the power to do something about it. That is why we have open source code, and that's why Netscape did this with Mozilla in the first place.
You non-programmers, you people with lives, you can't do any more than talk about the problems. The source package is huge, and a lot to wade through. But I find it outrageous that people walk around and bitch about how unstable and slow the code is and yet make no effort to even provide moral support. Mozilla is the ONLY significant standalone open-source browser out there. kfm is minimal, Opera is purely commercial, and everything else (Amaya, Arena, HotJava, Mosaic) is either dead or still experimental.
WE NEED MOZILLA TO SUCCEED. Reminding people of things like the vaporware issue (a nonsequitur since Open Source vaporware is an oxymoron), the declining market share, and the sluggishness of the *current implementation* is bullshit. If you can't contribute, tell the Mozilla people your problems. If you can, shut your trap and start coding.
Whew. Feel better now.
After that little jag, I'll put my money where my mouth is if the world wants me to. That Lions Book for Mozilla -- if someone's willing to pay me to write it (or at least coordinate it), I'd be happy to (pay because I can't afford to volunteer my time). As for the AOL issue, if there's a real feeling that AOL isn't doing right by its own software, why haven't we seen a Mozilla fork yet?
/Brian
I think you're rather jumping the gun...
Dylan first. Dylan was a pretty nice language, sure, but it's very much in the long tradition of companies trying to get people to use their languages instead of commodity products (witness C# from Microsoft... hell, even Java...). The fact is that Dylan was never necessary and was just too different from what most people used.
The Newton. Well, I have a Newton MP120 (v1.2 OS), and I think I can tell you pretty much everything that went wrong with it. Newton broke too far out of the computer paradigm -- it tried to be an electronic notebook and wound up creating an interface that was at best confusing. The trusted-compiler model that NewtonScript used was no help; until very late in the Newton's life cycle it was more or less impossible to do a direct port. Newton as an architecture was locked up tight, just like WinCE is now.
Newton descended into self-parody towards the end anyway -- as Palm was showing the world that PDAs didn't need to be able to do everything to do their job and reducing hardware size and cost accordingly, Newtons were getting more expensive and remaining clumsy. The eMate was workable, even innovative. The $1000 MessagePad 2100, still the size of a large paperback and now packaged with a keyboard the same size, was not.
The Newton was innovative, and despite the crap it's gotten had the best handwriting recognition in the business. But it simply wasn't the tool it could have been. I don't blame Apple for that; they had to write the rules as they went along when they invented the PDA. But I do blame them for not keeping pace.
Quickdraw GX... well, that IMHO was just poorly organized. It wasn't cross-platform, and it was really a Copland feature that was brought out far too early. I rather miss it myself, but Apple simply didn't push hard enough.
The real tragedies in Apple's story are OpenDoc (why can't they opensource the wreckage and let the market decide whether it can use it?) and Copland (same deal). To those who don't remember, Apple was creating great hardware in those days (even in these days where G4s are going dual-processor standard I'd still love to have a 6500 in the house), but their software was lagging badly. It nearly killed them.
The simple fact is that all the projects above were draining Apple where it counted. Killing Newton was probably a bad idea, but Newton had failed to evolve. OpenDoc was simply getting buried because developers weren't quite committing to it. QDGX was too narrow. Dylan was cool, but new languages are a toy -- I'm creating a programming language right now and I can honestly say I don't expect it to be all that serious...
/Brian
This may be a bit redundant, but I'll try anyway...
1. Maybe, maybe not. You're assuming that commercial open-source == abandonment; if that's so why did Apple go with Darwin?
They're striking out into new territory either way. If they are copping out as you say, than they're doing the responsible thing and making sure that Star Office stays in the public record, to make a reappearance later. Otherwise, they're doing what Netscape tried to do with Mozilla and indeed leveraging outside talent to make the project better.
2. This is simply flat-out wrong. If anything, as long as StarOffice provides something that people want, open-sourcing it will ensure that it goes wherever it's wanted.
And in any case, you talk about magnitude of cross-platform software... oy. Linux. *BSD. Apache. Perl. gcc. Maybe you're talking end-user apps -- well, no one has really tried yet. Mozilla is still not yet out of the starting gates. StarOffice will be the real litmus test.
3. If things are as gloomy as you say, forking StarOffice is the best thing that could possibly happen. If StarOffice does fork, that means two things:
-Someone considers it worth keeping around.
-Someone thinks they can do it better than Sun.
IMHO that would actually be a credibility boost -- someone thinks it's worth the trouble to improve. The key would be finding a solid corporate backer for said fork.
4. That's because we don't have enough Mac geeks in the Open Source world, IMHO. The Mac world (the Palm community as well) is very much a pay-for-play world; shareware is the dominant form of redistributable software, and the really successful shareware authors are very, very good (try an Ambrosia game sometime). More than anyone else, Mac people are very sensitive to crappy interface design because it's such a major part of the experience. Windows people aren't quite so picky because Windows had no consistent design philosophy until Win95 came out, and Linux (geek OS that we all know it is) is the Unix world's first real splash in the consumer market in its thirty-some-odd years of existence. But the Mac hackers know very much what they're doing, and most of them want to get paid.
/Brian
Figure this. The hardware in an N64 is basically a baby Indy, circa 1996 or so. That's still pretty respectable horsepower even today, at least outside the Wintel world. It's been cracked pretty wide open, it seems.
Two ideas, one hardware, one software. Hardware, an N64 cartridge with an Ethernet port and a USB adapter for the controller port. Software, Linux N64 (could in theory be based of the existing MIPS/SGI port, no?).
Okay, now, everyone repeat after me...
"I'd love to see a Beowulf of those things!"
Think about it... heavy-duty processing power for US$100 a box, and you could fit about eight of them on one shelf on a standard rack. You'd spend more money on the hubs than the CPUs...
/Brian
My personal experience, before having read any of the four hundred or so responses already under this thread...
I learned to program on a Commodore 128, in MSBASIC 7.0, one manifestation of the prototypical policy=implementation environment. It wasn't all that bad (structured programming constructs had been backhacked into the language, but that didn't mean much to a twelve-year-old trying to figure out how to write an operating system in BASIC), but to learn to code on such an environment is very limiting. There is no hand-holding outside the manual, but you can't get lost easily (short of a rogue POKE) so it doesn't matter. It wasn't for a very long time that I understood the benefits of keeping the implementation separate from the environment. This is what has made Unix more successful than many other environments of its generation.
The beauty of Unix is that implementation doesn't matter -- you have your choice of tools, and you're free to use whatever you're comfortable with. You don't have to write every last piece of code in bash or jump through hoops to do otherwise. Linux goes traditional Unix one better by traditionally offering a far richer development environment than any other system on the market. Not only do you have a solid, time-tested C compiler (gcc), you have any number of other languages like Perl (my personal language of choice), Python, Fortran, etc...
Another aspect of GUI systems is the IDE. I used to think IDEs were the greatest idea until I switched my Mac environment from Think to CodeWarrior. Now CodeWarrior is pretty damn good, and AppleScripting gives you every bit as much flexibility as shell scripting and then some, but to me it's not particularly obvious how to construct a tool chain in a project build -- there is a learning curve that simply doesn't exist in a traditional shell environment.
I don't like make(1) very much, I'll admit, but there's a certain simplicity to using componentized tools that just isn't there when you're using modulomonoliths like CodeWarrior or Visual Studio. Fast turnaround tools are much more easily managed under Unix, as well; I don't think I could meet the demands of the constant fine-tuning my current language project requires if I wasn't writing it in Perl (or maybe Python), and as good a tool as MacPerl might be it's still a GUI app.
So what's so great about Linux? The flexibility. Yes, there's a learning curve. But it's worth it just to not have to worry about the complications involved in managing a full-up project file for a simple program.
/Brian
Good idea, except for one thing: isn't Marathon (1 and 2, anyway) open-source?
Think about it: Microsoft gets the talent, but Bungie offloads the hot games. Why doesn't it worry about Marathon? Because technically it no longer controls it.
Fire up the code mirrors, boys ('n girls)...
/Brian
Keeping in mind, of course, that right or wrong there was good reason for Jackson to be a bit slanted. After all, he was the judge for the '95 consent decree that MS urinated on so freely, and it's tough to keep a straight face when the most powerful company in the world is blatantly, obviously lying.
Now for that right or wrong part: maybe Jackson should have kept a straight face, and that could hurt. IMHO MS won't do all that well in Appeals, but they might wind up doing better than they would have because of that...
/Brian
Just out of curiosity, what does the Linux body count look like these days anyway? Solaris x86 seems to be universally considered a joke, SCO doesn't seem to be all that far from throwing in the towel, and Irix is circling the drain...
I do agree that Solaris/SPARC is in it for the long haul, and NextStep/MacOS X is just getting out of the gates. SCO is in an odd situation, though; keepers of the flame, yet their bread-and-butter is possibly the most marginal of the major unices out there.
/Brian
That's a good point, about Unix. Saying it was "designed" for anything is giving Ken and Dennis entirely the wrong type of credit.
:-) ) around how natural language works. It's evolved by being used. Lisp, particularly Scheme, didn't do that; Scheme is sort of a hothouse flower, very pretty, but something of a toy language outside its intended purpose (teaching computer science students how to hack).
Actually, a friend of mine had an interesting insight when I discussed with him a project I was doing on my own regarding programming languages like Lisp. His point was that the reason elegantly-designed solutions like these don't necessarily take off is that they're purely engineered. The reason I like hacking in Perl is that it's designed (loosely speaking
So bully for Lucent -- I'd love to play...
/Brian
Read the book. It's a rather interesting ramble about user interface; Linux Journal slammed it for not really sticking to its message, but I'm not entirely certain it was supposed to have one in the first place.
/Brian