That's precisely how it is now. \Device\HardDisk0\Partition0, (yes, backslashes, the path parser can do either forward or backslashes, though the object manager uses backslashes canonically). Colon is a reserved character in filenames, but it's used to separate the filename from the fork name in NTFS. Ironically, the only system I know of that can use forked files from the shell is cygwin -- again, because of a braindead parser in the cmd.exe and explorer shells that rejects them.
Symbolic links don't exist per se in windows, but something more powerful does, and that's reparse points. reparse points are understood in the kernel and do not require application support, and they act exactly like translators in hurd. Unlike hurd, reparse points are stackable, whereas hurd translators are not. Reparse points are also used to implement mount points as well.
Then why am I able to use it to copy, rename, move, and delete files on my local system? Konquerer knows damn little about web browsing -- that's KHTML's job, which is simply embedded as a component in konquerer. Konquerer is a component-based generic browser, not just a web browser. It's what MS tried to do with its explorer integration, and actually didn't pull off -- the problem with IE on windows is that it's not integrated enough. If it were seamlessly integrated, it wouldn't rearrange menus ("hey, View:Folder Options just got moved to Tools:Internet Options!"), it would let me drag and drop multiple files from ftp onto the desktop, and developers would be able to write shell extensions that worked on URL's instead of targeting a completely different API. Now that XP has the Win2k kernel, and thus reparse points, maybe this will start to happen... Doubt it, they're still foisting drive letters on us even though they've outlived their usefulness.
Back to the point, if you think konquerer is "just a web browser" when you're using it as such, then it's accomplished its goal of fooling you into believing that;)
When IPV6 is deployed, how do I prevent the machines on the inside of my firewall from being routable?
Tell your firewall to not route it. The only reason 10.0.0.0 and 192.168.0.0 (I don't remember the class C one) are non-routable is because every single hop has wired into it the knowledge that those aren't routable.
Plus, I have to imagine there are nonroutable IP6 blocks as well...
I have a few problems with Nielsen... like his process, or seeming lack thereof. Take a gander at http://www.useit.com/alertbox/20020203.html where he says that cleartype will "probably" save 10-15% in reading speed, then taking that same "probably", goes on to suggest thousands of dollars in savings because of increased efficiency. It's a big leap. See, Visual Studio.NET supports cleartype also, so just for kicks, I turned it on, since I use a laptop. Color fringies everywhere unless I looked at the screen just so, where they were only tolerable. My head hurt just trying to read the text on the screen.
I would assert (with just as much to back it up, but what's good for the goose is good for... another goose?) that a well designed typeface that takes aliasing into account is just as "efficient" to read. I certainly don't see Bitstream and company trumping the cost savings of their fonts, because the claim is ludicrous on its face (no pun intended).
> A function should always throw out data that doesn't match its parameters.
No, it should signal an exceptional condition. Checking the return value of strlcpy or strncpy for "actual bytes written" means checking it against strlen... scanning the string just to get the length. If I could simply get a return value that indicated success or failure, that would be infinitely preferable in my codew. Not that C has strings anyway, it just has some array hacks to deal with moving around what amounts to void* chunks, and not even efficiently at that.
It's SPAM. Not advertising, SPAM. Just because it is "thriving" does not give them the right to spam us.
It certainly does seem to give you and the rest of the slashdoterati the right to jump to wild and hysterical conclusions though. Doubleclick acquired MessageMedia. MessageMedia makes mailing list software with html mail click-through tracking abilities (same way those email valentines cards do) so they can tell who was interested in the mailings. Their main business is in doing systems integration, to massage the marketing data from various databases into mailing list categories. Sounds like a lot more work than address harvesting for spam, don't it?
But it's easier to jump on the scapegoat, isn't it? They should have to prove their innocence and justify whatever they do isn't wrong, because it's just too darn hard to find facts, isn't it?
Re:GNOME vs KDE for the newbie
on
GNOME 2.0 Beta
·
· Score: 2
Funny thing, it's Linux I just can't stand as a desktop OS, because it's always so damn *slow* whenever I want to do desktop-ish work on it. I can use emacs to write code on either platform, but when I want to run off some handouts, I just get the job done faster in Word. I used to run windows for IE, now I run windows for Mozilla. I discovered it's just Mozilla on Linux that gives the project a bad name...
I don't want to be "forced to understand" what I'm doing. I already know what it's doing behind the scenes. I'd rather it stay behind the damn scenes where it belongs.
Actually, mono never really goes away. Some believe it's one possible cause of chronic fatigue syndrome. Must be, I for one am really tired all the time of the trolls that masquerade as.NET articles on slashdot...
Erm... BeOS might be nice and all, but a) no ISV's are touching it now, and b) neither are hardware makers.
To be everything windows is, you have to bury your developers in mountains of documentation and SDK's. Linux, for all its hodgepodge character, actually does work there, in perhaps not providing a great depth of documentation but certainly a breadth of SDK's to choose from. Be managed to screw its developers every time it turned around (it's on PPC, oh nevermind now it's not, oh now it's an appliance!) and the documentation, except for a few notable subsystems like the filesystem, was grossly sub-par.
*cough*.. erm, glibc requires kernel headers to compile. It works on linux, HURD, and that is it. I'd really love to see it compile and work on win32 or a mac, let alone OpenVMS or S/390
> I have a copy of Microsoft Windows 1.0 on 5.25" Floppies! It's fully functional, with manual!
Wasn't it called "Presentation Manager for DOS" back then? I seem to remember it was PMD then Windows 286, then Windows 3.0, after which everyone knows the numbering.
There's a quote a radio station I used to listen to used: "You get what you pay for. And you're listening to us . . . for free."
I don't pay, so slashdot is free to be worthless? As for your radio analogy, turn the dial sometime, and except for some hold-out college stations, you get crap, crap, crap, endless regurgitation of audio pablum, More Of The Same everywhere you go. Is that all you expect of slashdot?
I post because I actually give a damn about this this so-called community, I identify with it despite all the crap. Perhaps I should simply vote with my feet and find a more reasonable atmosphere, but I prefer a bustling community to an empty one. It just saddens me to see that the focus of discussion is getting hijacked by the articles themselves, away from meaningful news and toward hot-button flamebait.
Where is the moderation for articles? How about cumulative votes for "objective" vs "opinionated" or "accurate" vs "inaccurate"? How about some reputation points for submitters and editors along these lines?
Slashdot is a pro-Linux site.
No, it seems more to be an anti-MS site. Perhaps I wouldn't mind so much if it focused on actual news instead of continually grinding its editorial axe. Removing the snide commentary from the article bodies and moving it to the posts to stand and be moderated along with the rest of us plebians would be a great start.
Basically, scheme and haskell AREN'T available for the CLR; the intersection of C# and haskell, or scheme, are available, with haskell or scheme style syntax, but it isn't the same as real scheme.
I was about to post a reply here, and to my chagrin, a fairly flameful one... something I'm noticing myself do too often, but that's another story.
Anyway, I'll stand by my assertion that MS is putting real effort, and not just marketing, into making MSIL a target for other languages other than C#. The tailcall instruction for example, very useful for functional languages. In any case, C#, MSIL, the JVM, and every native CPU instruction set are turing-complete, so their intersection is "that which can be computed" (with varying degrees of efficiency, no doubt). Scheme and Haskell for.NET are perhaps in an embryonic phase (I discovered to my horror that Mondrian doesn't perform compile-time type-checking, the fundamental strength of Haskell!) but they are ultimately no less constrained by the VM than they are by a real CPU (other than the fact that the simple stack-based VM is a poor match for CPU architectures that stress usage of registers instead)... And at least MS is acknowledging that other paradigms (functional, logic) exist and aren't beating the OOP drum to the exclusion of everything else.
Re:Simple solution to a growing complexity
on
Read the Fine Print
·
· Score: 2
> I realize that Slashdot is being composed of more and more pro MS supporters.
No, just more and more people have grown sick and tired of reading juvenile MS-bashing rants on slashdot in lieu of actual news, and perhaps even resent the fact that in this economy, these people still get paid for this slipshod level of journalism.
Slashdot's editorial position tends to be. To the point of producing snide commentary in the body of the articles submitted by someone else. Minor version of software for Linux get announced with great fanfare, but nary a peep whenever there's a new development in the Windows development world, because after all, "real geeks don't do windows".
Thankfully the backlash of respondents against the knee-jerk bill-bashing is getting stronger...
Re:When will the real evolution of RTS arive?
on
HIstory of RTS Games
·
· Score: 2
The mindless action of highlighting a large group of cheap single functionality units and pointing them in the direction to roll over anything they come across.
Worked for Russia in WW2. It helped that they were mopping up an overextended army that depended on armor that no longer had any that was usable...
Anyway, it's one reason I prefer games that do *not* involve resource gathering and unit production, such as the Myth games and the Close Combat series. In CC, you can expend infantry on rushing a position, but you'll find yourself short for the whole campaign. TA is one of the few settings I can think of where rush tactics wouldn't be ultimately suicidal in terms of undermining morale and logistics -- they're just robots that are literally sprayed out by the dozens. Otherwise, after a couple rushes, you might find a few unit commanders that are a touch reluctant to go on fire missions with 90% casualties...
"CLR lacks continuations"... so does JVM
"CLR requires static single inheritance"... so does JVM
"CLR lacks multiple dispatch"... etc
I certainly think that the CLR could stand the criticism and have its hype deflated, but I'm not finding a lot to recommend the JVM. Sun doesn't even acknowledge, much less support languages other than Java on the JVM, with the exception of GJ, which it would rather absorb than support.
I might also note that there are languages very much not like C# available for.NET, including haskell and scheme.
.NET is the big umbrella marketing term of Microsoft's to describe their systems integration framework. Not necessarily with foreign OS's (though it should be easier to do so), but with MS's own tools. By itself, it's kind of meaningless except to say that it's Microsoft and it's new, so if you're "doing something with.NET", chances are you're not fiddling with DDE for example, but otherwise it could be a lot of things.
So what.NET covers is mostly two things: first, a common language runtime. This is that "clone of java" everyone goes ape over, as if Java invented bytecode compilation and virtual machines. This means Visual Basic, Visual C++, and their sorta-objective-C language C# will compile down to the same bytecodes, so they can call each other's routines, use each other's objects, and so forth.
That takes care of integration on one machine, so the next step is to resurrect remote procedure calls with an OO flavor, and that's what SOAP (Simple Object Access Protocol) is (the OO is actually an extension called, no kidding, ROPE). It's basically RPC over XML (the main creator of XML RPC helped design SOAP), which MS figures, probably rightly so, will be a hell of a lot easier to work with than DCE RPC, the hugely arcane protocol that all of NT's current network services as well as DCOM are based on. SOAP is already enjoying some success in existing products, and since it's so easy to implement, has rather complete implementations in languages like perl and python already.
So boiled down, it's a bytecode VM and a kindler gentler RPC. Not hugely exciting but for the fact that it makes working with MS tech a whole lot simpler, and MS is really putting a lot more effort into it than it did with its half-hearted push of DCOM and COM+ (aka DNA).
I for one will lump anything that uses.NET in with Microsoft products, even if it's "open source". Why take the chance? I'm surprised that so many/. folks are calling.NET "progress" or "a standard". It's just a Microsoft technology
Score 5.
There you have it, the intellectual honesty of slashdot.
> Remember,.NET requires interaction with a server somewhere
Do tell me which parts of the Common Language Runtime require connectivity to a remote host? Did the fact that you haven't the remotest clue what.NET is even dissuade you for a moment from opening your mouth and opining about the wisdom of adopting the technology? Miguel has actually used.NET -- have you even looked at it?
I'm sorry, thanks for playing, but the Object Manager covers devices. And it's dynamically built (you ever seen MAKEDEV on NT?) Yes, I'm not talking about the DOS-derived branch of windows any more than you're talking about v7 unix. It's just the lousy shell that keeps you from being able to browse it like a filesystem, but it has path syntax (and can address the filesystem as well) like any other filesystem... backslashes and all.
Xv... I'm assuming you're not referring to the image viewer.
NTFS CDs: Even windows 9x has modular filesystem drivers. There's one floating around out there that does ext2. I don't particularly see that as an architectural feature of unix that I asked for, it looks like a feature point you pulled out of your hind end to fill out the list.
> Any yet, as a force limiting our basic biological capabilities gravity no longer affects us.
Ask any obese person or someone with gigantism whether this is true. Gravity is the very thing that determines our body shape. It's why we're not shaped like octopi. You're saying that our shape itself is not a limiting factor to our physical capabilities? The fact that we have machinery to get over physical limitations is proof that we have these limitations in the first place.
The point I was trying to make was that people buy Oracle to get Oracle. The database, the dev tools, the support. None of which Redhat has established any kind of reputation for. I agree that Oracle is often way overkill for many of its applications, but I just don't see myself running payroll on one of Redhat's experiments in new markets. (I see myself running it on DB2 actually, but that's another story)
Actually those predictions are not really opposites. That IBM big iron can run a virtual cluster of those linux boxen, it's not just one big linux box. I always figured if I ever started a hosting company, I'd just get a single mainframe and just run every server off that one (that's why I don't run a hosting company I suppose)
Slashdot Mirror
That's precisely how it is now. \Device\HardDisk0\Partition0, (yes, backslashes, the path parser can do either forward or backslashes, though the object manager uses backslashes canonically). Colon is a reserved character in filenames, but it's used to separate the filename from the fork name in NTFS. Ironically, the only system I know of that can use forked files from the shell is cygwin -- again, because of a braindead parser in the cmd.exe and explorer shells that rejects them.
Symbolic links don't exist per se in windows, but something more powerful does, and that's reparse points. reparse points are understood in the kernel and do not require application support, and they act exactly like translators in hurd. Unlike hurd, reparse points are stackable, whereas hurd translators are not. Reparse points are also used to implement mount points as well.
Best of all, Konqueror is *just* a web browser
;)
Then why am I able to use it to copy, rename, move, and delete files on my local system? Konquerer knows damn little about web browsing -- that's KHTML's job, which is simply embedded as a component in konquerer. Konquerer is a component-based generic browser, not just a web browser. It's what MS tried to do with its explorer integration, and actually didn't pull off -- the problem with IE on windows is that it's not integrated enough. If it were seamlessly integrated, it wouldn't rearrange menus ("hey, View:Folder Options just got moved to Tools:Internet Options!"), it would let me drag and drop multiple files from ftp onto the desktop, and developers would be able to write shell extensions that worked on URL's instead of targeting a completely different API. Now that XP has the Win2k kernel, and thus reparse points, maybe this will start to happen... Doubt it, they're still foisting drive letters on us even though they've outlived their usefulness.
Back to the point, if you think konquerer is "just a web browser" when you're using it as such, then it's accomplished its goal of fooling you into believing that
When IPV6 is deployed, how do I prevent the machines on the inside of my firewall from being routable?
Tell your firewall to not route it. The only reason 10.0.0.0 and 192.168.0.0 (I don't remember the class C one) are non-routable is because every single hop has wired into it the knowledge that those aren't routable.
Plus, I have to imagine there are nonroutable IP6 blocks as well...
I have a few problems with Nielsen ... like his process, or seeming lack thereof. Take a gander at http://www.useit.com/alertbox/20020203.html where he says that cleartype will "probably" save 10-15% in reading speed, then taking that same "probably", goes on to suggest thousands of dollars in savings because of increased efficiency. It's a big leap. See, Visual Studio .NET supports cleartype also, so just for kicks, I turned it on, since I use a laptop. Color fringies everywhere unless I looked at the screen just so, where they were only tolerable. My head hurt just trying to read the text on the screen.
... another goose?) that a well designed typeface that takes aliasing into account is just as "efficient" to read. I certainly don't see Bitstream and company trumping the cost savings of their fonts, because the claim is ludicrous on its face (no pun intended).
I would assert (with just as much to back it up, but what's good for the goose is good for
> A function should always throw out data that doesn't match its parameters.
... scanning the string just to get the length. If I could simply get a return value that indicated success or failure, that would be infinitely preferable in my codew. Not that C has strings anyway, it just has some array hacks to deal with moving around what amounts to void* chunks, and not even efficiently at that.
No, it should signal an exceptional condition. Checking the return value of strlcpy or strncpy for "actual bytes written" means checking it against strlen
It's SPAM. Not advertising, SPAM. Just because it is "thriving" does not give them the right to spam us.
It certainly does seem to give you and the rest of the slashdoterati the right to jump to wild and hysterical conclusions though. Doubleclick acquired MessageMedia. MessageMedia makes mailing list software with html mail click-through tracking abilities (same way those email valentines cards do) so they can tell who was interested in the mailings. Their main business is in doing systems integration, to massage the marketing data from various databases into mailing list categories. Sounds like a lot more work than address harvesting for spam, don't it?
But it's easier to jump on the scapegoat, isn't it? They should have to prove their innocence and justify whatever they do isn't wrong, because it's just too darn hard to find facts, isn't it?
Funny thing, it's Linux I just can't stand as a desktop OS, because it's always so damn *slow* whenever I want to do desktop-ish work on it. I can use emacs to write code on either platform, but when I want to run off some handouts, I just get the job done faster in Word. I used to run windows for IE, now I run windows for Mozilla. I discovered it's just Mozilla on Linux that gives the project a bad name...
I don't want to be "forced to understand" what I'm doing. I already know what it's doing behind the scenes. I'd rather it stay behind the damn scenes where it belongs.
Actually, mono never really goes away. Some believe it's one possible cause of chronic fatigue syndrome. Must be, I for one am really tired all the time of the trolls that masquerade as .NET articles on slashdot...
Erm ... BeOS might be nice and all, but a) no ISV's are touching it now, and b) neither are hardware makers.
To be everything windows is, you have to bury your developers in mountains of documentation and SDK's. Linux, for all its hodgepodge character, actually does work there, in perhaps not providing a great depth of documentation but certainly a breadth of SDK's to choose from. Be managed to screw its developers every time it turned around (it's on PPC, oh nevermind now it's not, oh now it's an appliance!) and the documentation, except for a few notable subsystems like the filesystem, was grossly sub-par.
> Glibc is cross platform.
.. erm, glibc requires kernel headers to compile. It works on linux, HURD, and that is it. I'd really love to see it compile and work on win32 or a mac, let alone OpenVMS or S/390
*cough*
> I have a copy of Microsoft Windows 1.0 on 5.25" Floppies! It's fully functional, with manual!
Wasn't it called "Presentation Manager for DOS" back then? I seem to remember it was PMD then Windows 286, then Windows 3.0, after which everyone knows the numbering.
There's a quote a radio station I used to listen to used: "You get what you pay for. And you're listening to us . . . for free."
I don't pay, so slashdot is free to be worthless? As for your radio analogy, turn the dial sometime, and except for some hold-out college stations, you get crap, crap, crap, endless regurgitation of audio pablum, More Of The Same everywhere you go. Is that all you expect of slashdot?
I post because I actually give a damn about this this so-called community, I identify with it despite all the crap. Perhaps I should simply vote with my feet and find a more reasonable atmosphere, but I prefer a bustling community to an empty one. It just saddens me to see that the focus of discussion is getting hijacked by the articles themselves, away from meaningful news and toward hot-button flamebait.
Where is the moderation for articles? How about cumulative votes for "objective" vs "opinionated" or "accurate" vs "inaccurate"? How about some reputation points for submitters and editors along these lines?
Slashdot is a pro-Linux site.
No, it seems more to be an anti-MS site. Perhaps I wouldn't mind so much if it focused on actual news instead of continually grinding its editorial axe. Removing the snide commentary from the article bodies and moving it to the posts to stand and be moderated along with the rest of us plebians would be a great start.
Basically, scheme and haskell AREN'T available for the CLR; the intersection of C# and haskell, or scheme, are available, with haskell or scheme style syntax, but it isn't the same as real scheme.
... something I'm noticing myself do too often, but that's another story.
.NET are perhaps in an embryonic phase (I discovered to my horror that Mondrian doesn't perform compile-time type-checking, the fundamental strength of Haskell!) but they are ultimately no less constrained by the VM than they are by a real CPU (other than the fact that the simple stack-based VM is a poor match for CPU architectures that stress usage of registers instead) ... And at least MS is acknowledging that other paradigms (functional, logic) exist and aren't beating the OOP drum to the exclusion of everything else.
I was about to post a reply here, and to my chagrin, a fairly flameful one
Anyway, I'll stand by my assertion that MS is putting real effort, and not just marketing, into making MSIL a target for other languages other than C#. The tailcall instruction for example, very useful for functional languages. In any case, C#, MSIL, the JVM, and every native CPU instruction set are turing-complete, so their intersection is "that which can be computed" (with varying degrees of efficiency, no doubt). Scheme and Haskell for
> I realize that Slashdot is being composed of more and more pro MS supporters.
No, just more and more people have grown sick and tired of reading juvenile MS-bashing rants on slashdot in lieu of actual news, and perhaps even resent the fact that in this economy, these people still get paid for this slipshod level of journalism.
> Slashdot isn't a unitary entity
Slashdot's editorial position tends to be. To the point of producing snide commentary in the body of the articles submitted by someone else. Minor version of software for Linux get announced with great fanfare, but nary a peep whenever there's a new development in the Windows development world, because after all, "real geeks don't do windows".
Thankfully the backlash of respondents against the knee-jerk bill-bashing is getting stronger...
The mindless action of highlighting a large group of cheap single functionality units and pointing them in the direction to roll over anything they come across.
Worked for Russia in WW2. It helped that they were mopping up an overextended army that depended on armor that no longer had any that was usable...
Anyway, it's one reason I prefer games that do *not* involve resource gathering and unit production, such as the Myth games and the Close Combat series. In CC, you can expend infantry on rushing a position, but you'll find yourself short for the whole campaign. TA is one of the few settings I can think of where rush tactics wouldn't be ultimately suicidal in terms of undermining morale and logistics -- they're just robots that are literally sprayed out by the dozens. Otherwise, after a couple rushes, you might find a few unit commanders that are a touch reluctant to go on fire missions with 90% casualties...
"CLR lacks continuations" ... so does JVM
... so does JVM
... etc
.NET, including haskell and scheme.
"CLR requires static single inheritance"
"CLR lacks multiple dispatch"
I certainly think that the CLR could stand the criticism and have its hype deflated, but I'm not finding a lot to recommend the JVM. Sun doesn't even acknowledge, much less support languages other than Java on the JVM, with the exception of GJ, which it would rather absorb than support.
I might also note that there are languages very much not like C# available for
.NET is the big umbrella marketing term of Microsoft's to describe their systems integration framework. Not necessarily with foreign OS's (though it should be easier to do so), but with MS's own tools. By itself, it's kind of meaningless except to say that it's Microsoft and it's new, so if you're "doing something with .NET", chances are you're not fiddling with DDE for example, but otherwise it could be a lot of things.
.NET covers is mostly two things: first, a common language runtime. This is that "clone of java" everyone goes ape over, as if Java invented bytecode compilation and virtual machines. This means Visual Basic, Visual C++, and their sorta-objective-C language C# will compile down to the same bytecodes, so they can call each other's routines, use each other's objects, and so forth.
So what
That takes care of integration on one machine, so the next step is to resurrect remote procedure calls with an OO flavor, and that's what SOAP (Simple Object Access Protocol) is (the OO is actually an extension called, no kidding, ROPE). It's basically RPC over XML (the main creator of XML RPC helped design SOAP), which MS figures, probably rightly so, will be a hell of a lot easier to work with than DCE RPC, the hugely arcane protocol that all of NT's current network services as well as DCOM are based on. SOAP is already enjoying some success in existing products, and since it's so easy to implement, has rather complete implementations in languages like perl and python already.
So boiled down, it's a bytecode VM and a kindler gentler RPC. Not hugely exciting but for the fact that it makes working with MS tech a whole lot simpler, and MS is really putting a lot more effort into it than it did with its half-hearted push of DCOM and COM+ (aka DNA).
I for one will lump anything that uses .NET in with Microsoft products, even if it's "open source". Why take the chance? I'm surprised that so many /. folks are calling .NET "progress" or "a standard". It's just a Microsoft technology
Score 5.
There you have it, the intellectual honesty of slashdot.
> Remember, .NET requires interaction with a server somewhere
.NET is even dissuade you for a moment from opening your mouth and opining about the wisdom of adopting the technology? Miguel has actually used .NET -- have you even looked at it?
Do tell me which parts of the Common Language Runtime require connectivity to a remote host? Did the fact that you haven't the remotest clue what
> devfs
... I'm assuming you're not referring to the image viewer.
I'm sorry, thanks for playing, but the Object Manager covers devices. And it's dynamically built (you ever seen MAKEDEV on NT?) Yes, I'm not talking about the DOS-derived branch of windows any more than you're talking about v7 unix. It's just the lousy shell that keeps you from being able to browse it like a filesystem, but it has path syntax (and can address the filesystem as well) like any other filesystem... backslashes and all.
Xv
NTFS CDs: Even windows 9x has modular filesystem drivers. There's one floating around out there that does ext2. I don't particularly see that as an architectural feature of unix that I asked for, it looks like a feature point you pulled out of your hind end to fill out the list.
> Any yet, as a force limiting our basic biological capabilities gravity no longer affects us.
Ask any obese person or someone with gigantism whether this is true. Gravity is the very thing that determines our body shape. It's why we're not shaped like octopi. You're saying that our shape itself is not a limiting factor to our physical capabilities? The fact that we have machinery to get over physical limitations is proof that we have these limitations in the first place.
> Unix, though, was built to be very flexible, so that it is relatively easy to add or change functionality.
Name one such feature Windows doesn't also have.
The point I was trying to make was that people buy Oracle to get Oracle. The database, the dev tools, the support. None of which Redhat has established any kind of reputation for. I agree that Oracle is often way overkill for many of its applications, but I just don't see myself running payroll on one of Redhat's experiments in new markets. (I see myself running it on DB2 actually, but that's another story)
Actually those predictions are not really opposites. That IBM big iron can run a virtual cluster of those linux boxen, it's not just one big linux box. I always figured if I ever started a hosting company, I'd just get a single mainframe and just run every server off that one (that's why I don't run a hosting company I suppose)