the clipboard was originally intended to create a short-term copy of selected data. since that's what it does with the text of selected filenames, i'm afraid calling it 'totally inconsistent' is putting things a bit strongly. cut-and-paste for filenames is totally consistent with cutting and pasting text in any other context. the fact that you're pleased to define the metaphor a different way doesn't mean that the existing product has none.
on top of that, your own metaphor isn't as consistent as you might think. what would happen, for instance, if you cut one file, then cut another before pasting the first one into a new folder? under the standard clipboard metaphor, that would be a silent and irrevocable deletion of the file, which violates the principle of clarity.. a biggie in the mac os. for that matter, what should happen if you cut one file, select another, and paste? by default, the thing selected is replaced by the thing in the clipboard, so does that mean we should delete the selected file and replace it with the one just cut?
assuming we did manage to work around the difficulties, there's still another problem to consider: overloading the interface. if cut-and-paste does one thing when you've selected the text of a filename, and something else when you've selected the icon, the interface contains a modality that's likely to breed mistakes. the two types of selection are visually similar, and i don't think anyone believes that the average user would always get the distinction right on the fly. interface designers don't have as much license to blame their problems on stupid users who couldn't find a clue with both hands and a flashlight as other programmers, because the whole point of the game is to find something that makes sense to those very users.
BTW - your assertion that cut-and-paste normally copies everything, not some specific object property, is incorrect. the clipboard can actually carry several parallel versions of the copied information, and is designed to paste the version most compatible with the context of the target environment. if you cut a piece of text that's in 12-point Times New Roman, right justified, etc, all that style information is a property of what was copied. you can still paste that selection into a window that doesn't support all your style properties, though. the clipboard just strips off any information that isn't appropriate to the new context.
by that light, the fact that only the filename appears in the new context when you cut-and-paste from the Finder is *entirely* consistent with the overall metaphor.
> In my opinion, listing grips and other people in movie credits > is ridiculous. Their influence is insignificant, and doesn't > take any 'art', they could be easily replace by anyone else > trained in the field and the work wouldn't suffer.
apparently you're not a stagehand. i am. you probably don't know how few people there are who are actually well trained for each job. you probably don't know what a best boy is, let alone that there's no practical way to become one wihtout the patronage of someone already placed in the industry. you probably don't know that _Casablanca_ is rated as one of the greatest movies of all time because of the miracles the focus puller was able to bring off. you probably bitched that the Star Trek movies looked like TV episiodes without ever knowing why (answer: zero involvement by the focus puller). you probably don't know that _Titanic_, that CG tour-de-force, had a scene that couldn't be shot with the latest & greatest programmable cameras, and that the production company had to fly a 70-odd year old, retired focus puller to Mexico to do half a day (for him) of work that *no one* else could do.
you probably wouldn't even be able to say what shot it was if you watched the movie.
you probably have no idea how much skill is being lost in the industry because the people who know how to do all these things are retiring without passing their lore on to a new generation.
movies are hard work, and everyone.. right down the the gofers.. has to do their job right, or it shows. you may not be able to see the effects of their work, but that's strictly because you don't know what to look for. granted, i can't look at a 30-second clip and tell you whether the catering was good, but i can tell you a lot about the production environment. shitty catering, disorganized bookkeeping, techinical glitches that mean one more take at the end of a long day.. all of those are visible in the way the performers move. there's a lot of information about the making of a movie, right there in the product, as long as you know how to find it.
if you don't happen to like sitting through the credits, that's fine. it's your call, and it would be stupid for me to say you're 'robbing the workers' of their credit. just don't take the extra step into pointy-haired-boss logic and assume that anything you don't understand has to be easy. *that's* a disservice to skilled workers in any field.
as programmers, we owe our colleagues in other trades the same respect we want for ourselves. the best way to get it is to show some of our own.
> this may be just what we need to increase the number of female Linux users...
oh my..
balls-out Linux advocacy meets Archie-Bunker-style, patronizing chauvenism. just what the community needs.
condescention on this order doesn't do anything good for the public perception of Linux/OSS. it just reinforces the idea that we're children in a sandbox, playing at being grown-up.
the first rule of respect is that to get it, you have to be willing give it. if people associate Linux advocacy with a continual barrage of "you-suck-we're-better" dribbling, they won't respect Linux. what they'll do is decide Linux *needs* to measure itself against weak and pathetic opponents, and can't stand up to anything else.
the sentiment above is pro-Linux FUD. pure and simple. its goal is to kick the legs out from under something that steals limelight from Linux.
FUD is bad. its goal is to suppress one type of information by drowning it out with another. FUD is the antithesis of what OSS/Linux is all about. OSS/Linux doesn't need FUD. FUD hurts OSS/Linux, no matter who it's directed at, or who benefits from it. a Linux that needs FUD is a Linux that will follow in the footsteps of DOS/Windows, relying on hype and disdain rather than working to make itself better.
stop it.
.. and you owe an apology to every she-geek out there whose priorities *don't* include finding a laptop that coordinates with today's nail polish.
* In real life you could not make such choices, since you are * constrained to splitting a gold bar along gold atoms, which are * discrete units.
piffle..;-)
there's a perfectly sensible demonstration of the Banach-Tarski paradox which can be done in the normal universe:
the solid matter in your gold bar only occupies an infinitesimal fraction of the bar's apparent volume. there are also a finite, albeit large, number of gold atoms in that volume, so, as you said, any discussion of infinitely subdividing them is semantically flawed. empty space is continuous, though, so that *can* be subdivided infinitely. all you have to do is stop thinking about the bar as a structure composed of atoms, and think of the atoms as a sort of 3D-spray-paint to mark off a particular volume of empty space.
if you take a 1kg gold bar, you can infinitesimally subdivide the empty space which exists between the atoms, and use that to create a second, equivalent volume of empty space right next to it. of course, nobody can see it, so we need to mark the boundaries of the space with some atoms. take 1kg of neutron star matter, fluff half of it out so the protons and electrons are no longer in physical contact, and count out the resulting particles until you've built yourself 1kg of new gold atoms. place those gold atoms in the volume of empty space you extracted from the original gold bar so people can see where the edges are, and there you go: two gold bars of equivalent volume, extracted from the volume of a single gold bar.
yes, we had to add extra mass so people could actually see & feel the results, but that's a secondary concern. it was the _volume_ that we split. sure, the engineering is a bit tough at our current level of technology, but it's possible. i think you'll also find that any attempt to disprove the infinitesimal subdivision of empty space will lead you to a proof of existence for the Luminiferous Ether, which has been pretty well kicked in the head by now. the Uncertainty principle says we can do whatever we damn well please, as long as it's too small to see, or if we put things back the way they were before anyone looks.
sorry for the language, but let's take a reality check.. a fair chunk of the work being done in business today wouldn't even be *possible* without computers.
case in point: trot down to your local office supply store and buy a pack of ledger paper. now take it to your accounting department and ask them to run the numbers for next quarter's budget on it.. no spreadsheets allowed. if you're lucky, they'll just laugh in your face.
the same is true for all those databases and custom front-end applications which allow people sitting at a phone to take orders. just *try* checking the availability of every item manually (and don't forget, you not only have to check the shelves, you also have to cross reference against the other pending orders which haven't been packed yet). let's see what happens to the concept of same-day shipping if we go back to the good old paper-and-pencil method.
next up: desktop publishing.. sure, The Computer Is Just A Glorified Typewriter.. one which prints in any font, any size, isn't locked to a character-mapped raster, can generate graphics on the fly, allows you to make incremental changes to a document without losing all your previous work in the process, and can print as many identical copies of that document as you want. go haul out your old Smith Corona and see how long it takes to knock out a dozen employment applications, or twenty copies of the yearly sales report (*with* graphs) for the upcoming board meeting.. and remember, close doesn't count: they all have to be *identical*. how many of us can truly say we're nostalgic for the days of carbon paper and white-out?
computers have allowed us to become blase' about the prospect of an individual doing, in a few minutes or hours, what used to take entire departments days or weeks (hyperbole? go add up all the day's transactions in a bank by hand). if anyone wants to measure the effect of computers on productivity, i want to see a test where they do *exactly* the same work without recourse to computers at all.
Re:which infinity are we talking about?
on
Infinite Space
·
· Score: 1
> the continuum is precisely 2^(A_0), ie the cardinality of the power set of the natural > numbers.
wow.. you're braver than i am. i ducked out with a handwave rather than facing the task of explaining cardinalities.;-)
thank you, as well, for the correct identification of C. it's been a while, and i'd forgotten the exact relationship.
> also, in response to the first post, the rationals are not "discrete." they get very > close to one another.
loose use of terminology on my part, sorry..
i should've known better than to toss out a word with a specific mathematical definition and hope to get away with the conventional usage. it just sounded better than 'unique' or 'distinct' at the time. heat of the moment, thrill of the chase, and all that, y'know..;-)
which infinity are we talking about?
on
Infinite Space
·
· Score: 2
first off: yes, i have studied transfinite numbers, so i'm not just making this up as i go.
the fact of the matter is that infinities come in different sizes. the number of integers in the number line is designated with the symbol aleph-nought.. that's the first letter of the hebrew alphabet, sub zero. aleph-nought is also the number of rational fractions, even numbers, odd numbers, integers greater (or less) than zero, integer multiples of 37, or any other value which can be expressed strictly as a combination of integers. as an interesting side effect, there are exactly as many discrete, rational points in a one-dimensional line of length 1 as there are in an infinite three-dimensional space.
the number of irrational fractions.. numbers which cannot be expressed as a combination of any finite integers, like the square root of five.. is also infinite. it's just that the infinite number of irrationals is provably larger than aleph-nought. that number is designated as 'continuum', which is written as a gothic C, and falls somewhere between aleph-one and aleph-two if memory serves me correctly.
as for what aleph-one and aleph-two are, trust me, it's not worth the explanation it would take.. they're just more orders of infinity, each being a provably different size than any of the others. there are actually an infinite number of possible infinities, each of which is infinitely larger (or smaller) than any other.
meanwhile, numerical analysis.. the discipline which studies the approximation of values to an arbitrary level of precision.. offers a completely different spin on the issue.
the basic tenet of analysis is that in the real world, you eventually have to stop calculating decimal places and get on with life. it admits freely that the numbers it produces are wrong, but compensates by being able to prove that any inaccuracy in the numbers it produces will be small enough not to matter. numerical analysis effectively redefines zero as "any number so small we can ignore it".
that's signifcantly more useful in everyday life than the absolute precision of pure mathematics. anyone who wishes to argue otherwise will please refrain from doing so until they've first calculated the complete expansion of Pi.
the redefinition of zero has a serious impact on the concept of infinity, because infinity can be expressed as the value of any number divided by zero. at a more practical level, analysis only deals with one kind of infinity, which is defined as 'any number so large that making it bigger doesn't matter'.
by the second standard, yes, the web can be described as infinite. of course, so can a shoebox. the decision not to care is purely subjective, on the part of the person doing the measuring. as Hamlet said: "i could be bounded in a nut shell, and count myself a king of infinite space, were it not that i have bad dreams."
it's infinite for you of you say it is, and not infinite for me if i say it's not, and there's nothing in the laws of mathematics or logic which says those two positions have to be mutually exclusive unless we've both agreed on some kind of standard for 'big enough'. three grams is one hell of an error if you're measuring the mass of a proton, but irrelevant if you're measuring the mass of a galaxy. there's no such standard here, so there's no meaningful argument either way.
and don't even get me started on the subjectivity inherent in the word 'valuable'.
pro-Apple or anti, good license or bad, Open Source or not, when was the last time you recall public feedback having any effect on the terms of a Large Software Corporation's licensing agreements? the trend throughout the industry is rather on the "you'll take what our lawyers say you'll take and if you don't like it tough shit" side, if you'll recall.
if nothing else comes out of this whole issue, i'd be willing to bet that the APSL has been read and considered by more people than the "screw reading this, just click the Accept button and get on with the install" licenses of every web browser ever released.
for that matter, how many people would you say have downloaded and read the GPL, the OSD, and all the other 'pure' open source licenses for the first time in their lives, strictly because they wanted to cite the authoritative documents in this battle?;-)
as a matter of fact, yes..
on
MacMafia
·
· Score: 1
i happen to have a Mac SE, vintage '89, sitting on my desk at the moment with that exact hardware configuration. it runs OS 7.1, a word processor, and a telnet client quite comfortably, thank you. i have a couple more SEs with dual floppies and no HD at all. i don't consider them workhorses by any means, but they're good for taking notes while i step through code with a debugger on my main machine, and as sattelite testing units when i'm working on distributed code.
to expand on my earlier point, i believe it's entirely possible to port Linux to that kind of platform. i also know, however, that anyone who does such a port will have to deal with exactly the same kind of resource management issues as the developers of the Mac OS did 10-15 years ago.
nothing in software is free, and there are all sorts of mutually exclusive goals in software design.. that's part of any programmer's working reality. if you want to do do one thing well you have to pay for it by giving up something else.
Carmac is a game designer, which means his main goal is fairly well defined: devote every possible CPU cycle to drawing the next frame. that's a perfectly reasonable goal, and the game engine he's written is a speed demon. OTOH, he pays for that speed by using the resources of the system to the max. if his code *isn't* a resource hog, he isn't getting the kind of power that he could be.
the Mac OS, by contrast, was designed for an era when a 4MB SIMM was a major investment, not a keychain (enthusiastic Linux advocates sometimes forget that the revolution didn't have much momentum back when 32MB cost a few $K). it was also designed for a market that was relatively forgiving when it came to sheer, display-crunching, multimedia throughput. given that context, it only makes sense for the Mac OS to have been developed with an eye towards resource optimization.
deveopment of a coherent resource management schema is the single biggest PITA of programming. it includes memory allocation, error checking, and is tightly linked to your underlying domain model. *nobody* likes to go in and retune their resource management, because it falls just this side of a complete overhaul.
Carmac's code (though i haven't seen the source) is almost certainly designed around a resource model of "i want it, and i want it *NOW*", which is normal when you optimize for speed. the Mac OS is designed around a model which encourages a more careful (or stodgy, depending on your POV) approach to resource allocation.
to say that Carmac's code sucks because it didn't run on the first try under the Mac OS would be just as irrelevant as saying the OS sucks because it didn't support his resource model. the only meaningful statement is that the two aren't entirely compatible. either one could doubtless be tweaked to fit the other, but in either case, that would involve a lot of work.
Carmac probably doesn't want to redesign his whole resource management schema just to boost the cross-platform stability of his game engine. fortunately, he doesn't have to.. Apple has already recognized the value of retooling its OS to use the kind of resource luxury we take for granted today. in a few seasons, the whole memory manager/cooperative multitasking thing will be history, and people will have to find something new about the Mac OS to hate.
okay.. 'clueless' is a harsh word.. the guy does know a *lot* about game programming.
still, take a look at what he's saying.. by his own admission, he dived into low-level programming in an OS where he didn't know the memory model, and found himself having to reboot.. what a surprise. in effect, he's blaming the OS for his own mistake.
i know he's using that example as a lead-in to pointing out that the Mac OS memory model isn't designed around protected memory. okay, valid point.. the memory model is old, and is definitely due for some upgrading. OTOH, take a look at the conditions in which the thing was originally designed to run.
when Linux can run the kernel, a GUI, and a single application with 256K of RAM, another 512K of library routines in ROM, and nothing but a floppy drive, i'll listen to arguments that there's a better way.
as for the issue of preemptive versus cooperative multitasking, i'm willing to grant that OS technology has pretty much shifted to being preemptive. process scheduling is tightly linked to the memory model, though. preemptive multitasking is closely associated with a protected memory model, and a cooperative scheme works well with some of the assumptions built into the Mac OS memory model.
i'm surprised Carmac didn't mention the problems which result from direct pointer references into memory.. the OS repacks memory periodically to reduce heap fragmentation, so you can't assume that pointers will remain valid across certain system calls. the OS supports a double-indirection scheme which balances ease of use (**mp as opposed to *mp) against internal cleaning. OTOH, maybe he just locked all his pointers to keep them from being moved, and caught the second-stage frustrations of heap fragmentation.
again.. write a port of Linux that runs a GUI preemptively in 1MB of RAM with a 10MB hard drive, and i'll listen carefully to any arguments about alternative designs.
i'm not saying the Mac OS is fundamentally better.. in fact, i think protected memory and preemptive multitasking are better solutions now that the computing resources to support them are affordable. coherent memory and cooperative multitasking are legacy issues, and they're being dealt with.
okay, for starters, i'm a member of the pro-Apple camp. i'll skip over the reasons why, because to the best of my knowledge nobody has ever been convinced one way or the other by that kind of discussion.
now..
i remember having gotten a chuckle or two out of some of the near-miss Microsoft domain names which have been registered over the past couple of years, so i think it's only fair to laugh along when people post stories, even erroneous ones, about Apple doing something dumb as regards a domain name. the hallmark of a good sense of humor is the ability to laugh when you're on the receiving end of the joke.
OTOH, i also remember what *happened* to all those near-miss Microsoft domain names which have been registered over the past couple of years.. the shadow of the MS legal machine fell across them, and they were never again seen by the eyes of mortal man.
i don't see any reason why Apple should have any less right to protect its trademarks.
if the kid had somehow registered APPLEIMAC.COM *before* Apple announced the product to the public, i might have some sympathy for him. as it is, he's sitting on a name composed of two terms which were trademarked long before he got the bright idea to call dibs on the combination.
the thing i find surprising about the whole issue is the idea that anyone thinks the kid is anything *but* an opportunist.. well-meaning or not. even more surprising is the prospect that anyone thinks there's any doubt about the outcome of this scenario. there's a saying which describes his chance of winning if he's silly enough to go to court over the matter.. it involves the words 'snowflake' and 'hell'.
another old saying is, "never argue with a man who buys ink by the barrel".. in this case, the kid's parents are going to get a valuable lesson in what happens when you support of precocious youngster who violates the intellectual property rights of a company with its own legal division. once they see the price tag attached to their son's creativity, i personally think his folks will find this whim of his a lot less endearing.
close, but not quite. NT pulls some of its basic design from VMS, which was running mainframes while unix was invading them.
i make no statement about which was better.. long and vicious flamewars were crackling on that subject years before i got near either one.. i will mention, however, that VMS was much larger, and stuffed almost to bursting with features.
Slashdot Mirror
the clipboard was originally intended to create a short-term copy of selected data. since that's what it does with the text of selected filenames, i'm afraid calling it 'totally inconsistent' is putting things a bit strongly. cut-and-paste for filenames is totally consistent with cutting and pasting text in any other context. the fact that you're pleased to define the metaphor a different way doesn't mean that the existing product has none.
on top of that, your own metaphor isn't as consistent as you might think. what would happen, for instance, if you cut one file, then cut another before pasting the first one into a new folder? under the standard clipboard metaphor, that would be a silent and irrevocable deletion of the file, which violates the principle of clarity.. a biggie in the mac os. for that matter, what should happen if you cut one file, select another, and paste? by default, the thing selected is replaced by the thing in the clipboard, so does that mean we should delete the selected file and replace it with the one just cut?
assuming we did manage to work around the difficulties, there's still another problem to consider: overloading the interface. if cut-and-paste does one thing when you've selected the text of a filename, and something else when you've selected the icon, the interface contains a modality that's likely to breed mistakes. the two types of selection are visually similar, and i don't think anyone believes that the average user would always get the distinction right on the fly. interface designers don't have as much license to blame their problems on stupid users who couldn't find a clue with both hands and a flashlight as other programmers, because the whole point of the game is to find something that makes sense to those very users.
BTW - your assertion that cut-and-paste normally copies everything, not some specific object property, is incorrect. the clipboard can actually carry several parallel versions of the copied information, and is designed to paste the version most compatible with the context of the target environment. if you cut a piece of text that's in 12-point Times New Roman, right justified, etc, all that style information is a property of what was copied. you can still paste that selection into a window that doesn't support all your style properties, though. the clipboard just strips off any information that isn't appropriate to the new context.
by that light, the fact that only the filename appears in the new context when you cut-and-paste from the Finder is *entirely* consistent with the overall metaphor.
> In my opinion, listing grips and other people in movie credits
> is ridiculous. Their influence is insignificant, and doesn't
> take any 'art', they could be easily replace by anyone else
> trained in the field and the work wouldn't suffer.
apparently you're not a stagehand. i am. you probably don't know how few people there are who are actually well trained for each job. you probably don't know what a best boy is, let alone that there's no practical way to become one wihtout the patronage of someone already placed in the industry. you probably don't know that _Casablanca_ is rated as one of the greatest movies of all time because of the miracles the focus puller was able to bring off. you probably bitched that the Star Trek movies looked like TV episiodes without ever knowing why (answer: zero involvement by the focus puller). you probably don't know that _Titanic_, that CG tour-de-force, had a scene that couldn't be shot with the latest & greatest programmable cameras, and that the production company had to fly a 70-odd year old, retired focus puller to Mexico to do half a day (for him) of work that *no one* else could do.
you probably wouldn't even be able to say what shot it was if you watched the movie.
you probably have no idea how much skill is being lost in the industry because the people who know how to do all these things are retiring without passing their lore on to a new generation.
movies are hard work, and everyone.. right down the the gofers.. has to do their job right, or it shows. you may not be able to see the effects of their work, but that's strictly because you don't know what to look for. granted, i can't look at a 30-second clip and tell you whether the catering was good, but i can tell you a lot about the production environment. shitty catering, disorganized bookkeeping, techinical glitches that mean one more take at the end of a long day.. all of those are visible in the way the performers move. there's a lot of information about the making of a movie, right there in the product, as long as you know how to find it.
if you don't happen to like sitting through the credits, that's fine. it's your call, and it would be stupid for me to say you're 'robbing the workers' of their credit. just don't take the extra step into pointy-haired-boss logic and assume that anything you don't understand has to be easy. *that's* a disservice to skilled workers in any field.
as programmers, we owe our colleagues in other trades the same respect we want for ourselves. the best way to get it is to show some of our own.
oh my..
balls-out Linux advocacy meets Archie-Bunker-style, patronizing chauvenism. just what the community needs.
condescention on this order doesn't do anything good for the public perception of Linux/OSS. it just reinforces the idea that we're children in a sandbox, playing at being grown-up.
the first rule of respect is that to get it, you have to be willing give it. if people associate Linux advocacy with a continual barrage of "you-suck-we're-better" dribbling, they won't respect Linux. what they'll do is decide Linux *needs* to measure itself against weak and pathetic opponents, and can't stand up to anything else.
the sentiment above is pro-Linux FUD. pure and simple. its goal is to kick the legs out from under something that steals limelight from Linux.
FUD is bad. its goal is to suppress one type of information by drowning it out with another. FUD is the antithesis of what OSS/Linux is all about. OSS/Linux doesn't need FUD. FUD hurts OSS/Linux, no matter who it's directed at, or who benefits from it. a Linux that needs FUD is a Linux that will follow in the footsteps of DOS/Windows, relying on hype and disdain rather than working to make itself better.
stop it.
* constrained to splitting a gold bar along gold atoms, which are
* discrete units.
piffle.. ;-)
there's a perfectly sensible demonstration of the Banach-Tarski paradox which can be done in the normal universe:
the solid matter in your gold bar only occupies an infinitesimal fraction of the bar's apparent volume. there are also a finite, albeit large, number of gold atoms in that volume, so, as you said, any discussion of infinitely subdividing them is semantically flawed. empty space is continuous, though, so that *can* be subdivided infinitely. all you have to do is stop thinking about the bar as a structure composed of atoms, and think of the atoms as a sort of 3D-spray-paint to mark off a particular volume of empty space.
if you take a 1kg gold bar, you can infinitesimally subdivide the empty space which exists between the atoms, and use that to create a second, equivalent volume of empty space right next to it. of course, nobody can see it, so we need to mark the boundaries of the space with some atoms. take 1kg of neutron star matter, fluff half of it out so the protons and electrons are no longer in physical contact, and count out the resulting particles until you've built yourself 1kg of new gold atoms. place those gold atoms in the volume of empty space you extracted from the original gold bar so people can see where the edges are, and there you go: two gold bars of equivalent volume, extracted from the volume of a single gold bar.
yes, we had to add extra mass so people could actually see & feel the results, but that's a secondary concern. it was the _volume_ that we split. sure, the engineering is a bit tough at our current level of technology, but it's possible. i think you'll also find that any attempt to disprove the infinitesimal subdivision of empty space will lead you to a proof of existence for the Luminiferous Ether, which has been pretty well kicked in the head by now. the Uncertainty principle says we can do whatever we damn well please, as long as it's too small to see, or if we put things back the way they were before anyone looks.
sorry for the language, but let's take a reality check.. a fair chunk of the work being done in business today wouldn't even be *possible* without computers.
case in point: trot down to your local office supply store and buy a pack of ledger paper. now take it to your accounting department and ask them to run the numbers for next quarter's budget on it.. no spreadsheets allowed. if you're lucky, they'll just laugh in your face.
the same is true for all those databases and custom front-end applications which allow people sitting at a phone to take orders. just *try* checking the availability of every item manually (and don't forget, you not only have to check the shelves, you also have to cross reference against the other pending orders which haven't been packed yet). let's see what happens to the concept of same-day shipping if we go back to the good old paper-and-pencil method.
next up: desktop publishing.. sure, The Computer Is Just A Glorified Typewriter.. one which prints in any font, any size, isn't locked to a character-mapped raster, can generate graphics on the fly, allows you to make incremental changes to a document without losing all your previous work in the process, and can print as many identical copies of that document as you want. go haul out your old Smith Corona and see how long it takes to knock out a dozen employment applications, or twenty copies of the yearly sales report (*with* graphs) for the upcoming board meeting.. and remember, close doesn't count: they all have to be *identical*. how many of us can truly say we're nostalgic for the days of carbon paper and white-out?
computers have allowed us to become blase' about the prospect of an individual doing, in a few minutes or hours, what used to take entire departments days or weeks (hyperbole? go add up all the day's transactions in a bank by hand). if anyone wants to measure the effect of computers on productivity, i want to see a test where they do *exactly* the same work without recourse to computers at all.
> the continuum is precisely 2^(A_0), ie the cardinality of the power set of the natural
;-)
;-)
> numbers.
wow.. you're braver than i am. i ducked out with a handwave rather than facing the task of explaining cardinalities.
thank you, as well, for the correct identification of C. it's been a while, and i'd forgotten the exact relationship.
> also, in response to the first post, the rationals are not "discrete." they get very
> close to one another.
loose use of terminology on my part, sorry..
i should've known better than to toss out a word with a specific mathematical definition and hope to get away with the conventional usage. it just sounded better than 'unique' or 'distinct' at the time. heat of the moment, thrill of the chase, and all that, y'know..
first off: yes, i have studied transfinite numbers, so i'm not just making this up as i go.
the fact of the matter is that infinities come in different sizes. the number of integers in the number line is designated with the symbol aleph-nought.. that's the first letter of the hebrew alphabet, sub zero. aleph-nought is also the number of rational fractions, even numbers, odd numbers, integers greater (or less) than zero, integer multiples of 37, or any other value which can be expressed strictly as a combination of integers. as an interesting side effect, there are exactly as many discrete, rational points in a one-dimensional line of length 1 as there are in an infinite three-dimensional space.
the number of irrational fractions.. numbers which cannot be expressed as a combination of any finite integers, like the square root of five.. is also infinite. it's just that the infinite number of irrationals is provably larger than aleph-nought. that number is designated as 'continuum', which is written as a gothic C, and falls somewhere between aleph-one and aleph-two if memory serves me correctly.
as for what aleph-one and aleph-two are, trust me, it's not worth the explanation it would take.. they're just more orders of infinity, each being a provably different size than any of the others. there are actually an infinite number of possible infinities, each of which is infinitely larger (or smaller) than any other.
that's why mathemeticians call transfinite mathematics 'counterintuitive'.
meanwhile, numerical analysis.. the discipline which studies the approximation of values to an arbitrary level of precision.. offers a completely different spin on the issue.
the basic tenet of analysis is that in the real world, you eventually have to stop calculating decimal places and get on with life. it admits freely that the numbers it produces are wrong, but compensates by being able to prove that any inaccuracy in the numbers it produces will be small enough not to matter. numerical analysis effectively redefines zero as "any number so small we can ignore it".
that's signifcantly more useful in everyday life than the absolute precision of pure mathematics. anyone who wishes to argue otherwise will please refrain from doing so until they've first calculated the complete expansion of Pi.
the redefinition of zero has a serious impact on the concept of infinity, because infinity can be expressed as the value of any number divided by zero. at a more practical level, analysis only deals with one kind of infinity, which is defined as 'any number so large that making it bigger doesn't matter'.
by the second standard, yes, the web can be described as infinite. of course, so can a shoebox. the decision not to care is purely subjective, on the part of the person doing the measuring. as Hamlet said: "i could be bounded in a nut shell, and count myself a king of infinite space, were it not that i have bad dreams."
it's infinite for you of you say it is, and not infinite for me if i say it's not, and there's nothing in the laws of mathematics or logic which says those two positions have to be mutually exclusive unless we've both agreed on some kind of standard for 'big enough'. three grams is one hell of an error if you're measuring the mass of a proton, but irrelevant if you're measuring the mass of a galaxy. there's no such standard here, so there's no meaningful argument either way.
and don't even get me started on the subjectivity inherent in the word 'valuable'.
pro-Apple or anti, good license or bad, Open Source or not, when was the last time you recall public feedback having any effect on the terms of a Large Software Corporation's licensing agreements? the trend throughout the industry is rather on the "you'll take what our lawyers say you'll take and if you don't like it tough shit" side, if you'll recall.
;-)
if nothing else comes out of this whole issue, i'd be willing to bet that the APSL has been read and considered by more people than the "screw reading this, just click the Accept button and get on with the install" licenses of every web browser ever released.
for that matter, how many people would you say have downloaded and read the GPL, the OSD, and all the other 'pure' open source licenses for the first time in their lives, strictly because they wanted to cite the authoritative documents in this battle?
i happen to have a Mac SE, vintage '89, sitting on my desk at the moment with that exact hardware configuration. it runs OS 7.1, a word processor, and a telnet client quite comfortably, thank you. i have a couple more SEs with dual floppies and no HD at all. i don't consider them workhorses by any means, but they're good for taking notes while i step through code with a debugger on my main machine, and as sattelite testing units when i'm working on distributed code.
to expand on my earlier point, i believe it's entirely possible to port Linux to that kind of platform. i also know, however, that anyone who does such a port will have to deal with exactly the same kind of resource management issues as the developers of the Mac OS did 10-15 years ago.
nothing in software is free, and there are all sorts of mutually exclusive goals in software design.. that's part of any programmer's working reality. if you want to do do one thing well you have to pay for it by giving up something else.
Carmac is a game designer, which means his main goal is fairly well defined: devote every possible CPU cycle to drawing the next frame. that's a perfectly reasonable goal, and the game engine he's written is a speed demon. OTOH, he pays for that speed by using the resources of the system to the max. if his code *isn't* a resource hog, he isn't getting the kind of power that he could be.
the Mac OS, by contrast, was designed for an era when a 4MB SIMM was a major investment, not a keychain (enthusiastic Linux advocates sometimes forget that the revolution didn't have much momentum back when 32MB cost a few $K). it was also designed for a market that was relatively forgiving when it came to sheer, display-crunching, multimedia throughput. given that context, it only makes sense for the Mac OS to have been developed with an eye towards resource optimization.
deveopment of a coherent resource management schema is the single biggest PITA of programming. it includes memory allocation, error checking, and is tightly linked to your underlying domain model. *nobody* likes to go in and retune their resource management, because it falls just this side of a complete overhaul.
Carmac's code (though i haven't seen the source) is almost certainly designed around a resource model of "i want it, and i want it *NOW*", which is normal when you optimize for speed. the Mac OS is designed around a model which encourages a more careful (or stodgy, depending on your POV) approach to resource allocation.
to say that Carmac's code sucks because it didn't run on the first try under the Mac OS would be just as irrelevant as saying the OS sucks because it didn't support his resource model. the only meaningful statement is that the two aren't entirely compatible. either one could doubtless be tweaked to fit the other, but in either case, that would involve a lot of work.
Carmac probably doesn't want to redesign his whole resource management schema just to boost the cross-platform stability of his game engine. fortunately, he doesn't have to.. Apple has already recognized the value of retooling its OS to use the kind of resource luxury we take for granted today. in a few seasons, the whole memory manager/cooperative multitasking thing will be history, and people will have to find something new about the Mac OS to hate.
okay.. 'clueless' is a harsh word.. the guy does know a *lot* about game programming.
still, take a look at what he's saying.. by his own admission, he dived into low-level programming in an OS where he didn't know the memory model, and found himself having to reboot.. what a surprise. in effect, he's blaming the OS for his own mistake.
i know he's using that example as a lead-in to pointing out that the Mac OS memory model isn't designed around protected memory. okay, valid point.. the memory model is old, and is definitely due for some upgrading. OTOH, take a look at the conditions in which the thing was originally designed to run.
when Linux can run the kernel, a GUI, and a single application with 256K of RAM, another 512K of library routines in ROM, and nothing but a floppy drive, i'll listen to arguments that there's a better way.
as for the issue of preemptive versus cooperative multitasking, i'm willing to grant that OS technology has pretty much shifted to being preemptive. process scheduling is tightly linked to the memory model, though. preemptive multitasking is closely associated with a protected memory model, and a cooperative scheme works well with some of the assumptions built into the Mac OS memory model.
i'm surprised Carmac didn't mention the problems which result from direct pointer references into memory.. the OS repacks memory periodically to reduce heap fragmentation, so you can't assume that pointers will remain valid across certain system calls. the OS supports a double-indirection scheme which balances ease of use (**mp as opposed to *mp) against internal cleaning. OTOH, maybe he just locked all his pointers to keep them from being moved, and caught the second-stage frustrations of heap fragmentation.
again.. write a port of Linux that runs a GUI preemptively in 1MB of RAM with a 10MB hard drive, and i'll listen carefully to any arguments about alternative designs.
i'm not saying the Mac OS is fundamentally better.. in fact, i think protected memory and preemptive multitasking are better solutions now that the computing resources to support them are affordable. coherent memory and cooperative multitasking are legacy issues, and they're being dealt with.
in the meantime, RTFM.
okay, for starters, i'm a member of the pro-Apple camp. i'll skip over the reasons why, because to the best of my knowledge nobody has ever been convinced one way or the other by that kind of discussion.
now..
i remember having gotten a chuckle or two out of some of the near-miss Microsoft domain names which have been registered over the past couple of years, so i think it's only fair to laugh along when people post stories, even erroneous ones, about Apple doing something dumb as regards a domain name. the hallmark of a good sense of humor is the ability to laugh when you're on the receiving end of the joke.
OTOH, i also remember what *happened* to all those near-miss Microsoft domain names which have been registered over the past couple of years.. the shadow of the MS legal machine fell across them, and they were never again seen by the eyes of mortal man.
i don't see any reason why Apple should have any less right to protect its trademarks.
if the kid had somehow registered APPLEIMAC.COM *before* Apple announced the product to the public, i might have some sympathy for him. as it is, he's sitting on a name composed of two terms which were trademarked long before he got the bright idea to call dibs on the combination.
the thing i find surprising about the whole issue is the idea that anyone thinks the kid is anything *but* an opportunist.. well-meaning or not. even more surprising is the prospect that anyone thinks there's any doubt about the outcome of this scenario. there's a saying which describes his chance of winning if he's silly enough to go to court over the matter.. it involves the words 'snowflake' and 'hell'.
another old saying is, "never argue with a man who buys ink by the barrel".. in this case, the kid's parents are going to get a valuable lesson in what happens when you support of precocious youngster who violates the intellectual property rights of a company with its own legal division. once they see the price tag attached to their son's creativity, i personally think his folks will find this whim of his a lot less endearing.
for those who don't know the name, you can find it in _The Fountainhead_.
close, but not quite. NT pulls some of its basic design from VMS, which was running mainframes while unix was invading them.
i make no statement about which was better.. long and vicious flamewars were crackling on that subject years before i got near either one.. i will mention, however, that VMS was much larger, and stuffed almost to bursting with features.