Not 5 years ago, Microsoft was pushing an "Enterprise Architecture" called "DNA" which was basically VB or J++ and the MTS/COM+ Application Server. It was their compeitition to J2EE.
(Note: I hate VB, but...) One problem is that VB integrates well with COM applications that are not EOL. For example:
+ If your shop is automating MS Office 2000 or XP, VB6 is usually the easiest way to do it.
+ If you have an ASP-based web applicaiton, it likely uses VB6/COM for backend components. For example, you may have purchased Microsoft Commerce Server for $$$
Yeah it's possible to do COM stuff with.NET, but it's not nearly as integrated and can be rather clumsy. So it is suprising that VB6 would be EOLed before Microsoft has finished "NETifying" everything.
Long ago, Novell entertained the idea of replacing NetWare with Linux.
Novell actually announced they were replacing NetWare with UNIXWare. The product map whoed this UNIX-based thing called "SuperNOS" would appear after NetWare 4.
Turned out to be a terrible decision because the technical limitations of NetWare killed Novell's dominant standing in the market.
In Unix it is assummed that a program is quite able to look in arbitrary places for it's configuration. The user has to set an environment variable that says "look here", but *that* can be set by another script. Even totally broken programs are fooled into running by symbolic links or by making a script to run it do the necessary rsync
OK, you take these script/envvar/symlink tricks, mulitply by hundreds of applicaitons, multiply by thousands of workstations and the whole thing ends up looking like one big unmaintainable hairball hack attack.
Maybe that's just my Winbrain talking -- but when I do work on *nix systems, it always seems like I'm reverse engineering these tricks, even to get programs running that shipped with the distro.
Maybe a better way to think of it is a "server" mentaility -- custom configurations are required and there pretty much has to be a "pull" model, as you put it. However, when you have thousands of mostly identical workstations, it would seem that "push" is vastly superior to "pull".
I got a copy of the "gold master" of Windows 95 right from a Microsoft marketing shumck about a month early. I kinda doubt that this was a very major accomplishment.
Well, I was talking about NT OSes, not DOS-based ones. 98 uses less RAM than XP, the desktop might be snapper, but I think if you measured things like IO performance, you'd find that it was not at all faster.
That's nice, but maybe next time you can get Steve to stop talking about new product plans (probably a hard thing to do) and instead give us a hint when he will EOL old products.
For example, when I was chatting with the other, fatter, balder and sweaty Steve, he pointed me to a page which states Windows 2000 will go out of support in 2010.
OK let's quote your link: "When the kernel starts to perform swapping it's a sign that the physical memory in the machine has been oversubscribed."
This has absolutely nothing to do with cache -- unless you are condending that OS X is swapping out program segments to increase disk cache, which would be silly, which is why I dismissed it.
It's my observation that OS X tends to need more physical memory than an equivilant Win or Linux system. The "perceptual speed" of a stock 256MB OS X system running mail/web/office tasks is very poor -- basically continually thrashing. Whereas on Windows it's acceptable. Even a 512MB OS X system can be borderline. This corresponds to what people report in Mac-oriented technical forums.
And once you go into swap the perceptual speed is primarily based on Disk IO, not the OS or the CPU. And if you have a slow laptop drive, that just compounds the "perceptual speed" problem. Fine. RAM is cheap.
So, I might as well go back to the original point -- the 8MB minimum specs for NT were a joke, the system would swap like crazy just bringing up the desktop. Realistically NT machines had 32-128MB even back in the mid 90s. Once you eliminated all the swapping, you would find that each NT OS was perceptually much faster than the previous version. I know, I had a Pentium-133 128MB into the Win2000 era.
If I was comparing minimum spec 128MB OS X system, I could equally say they were all slow as shit. But that's exactly what you did with Windows systems.
The memory model was never misunderstood. An honest comparision between the evolution of OS X and Windows was apparently too much for you to synthesize, and you reacted with nothing but distortions and footstamping.
No, the real (Mac)Troll here is you. Come back when your reigned in your ideology enough to have an legitimate discussion.
I guess the opinion here is that only Apple knows how to write grossly inefficient code and then sell optimization as paid upgrade. Have it your way. Sheesh.
(NT 3.1 on modern hardware would be "snappy" because of the Progman desktop, but as soon as you tried to run any programs you would find that they would be in fact much slower than on XP. IO and video performance would be terrible even if you somehow could find drivers for modern hardware.)
Put succinctly: You have no idea what are talking about. You are apparently in your 30s or older, but you write and think like a 12 year old AOLer. You are an idiot and a flamer. Go away.
Like the other poster, I have an old G3 PowerBook and on that machine, MacOS X 10.3 is "painfully slow and bloated", but not as slow and bloated as 10.0 was. OTOH, both 10.3 and NT4 ran fine on new machines when they came out.
> stupid memory management and GUI composition and multitasking structure
Meaningless technobabble. I will say that OS X is ridiclously memory hungry and that has to limit percieved speed.
NT 3.1 to 4.0 was 3 years, so the timescale is almost identical. And yes, I ran all 3 versions on similar 486/66 hardware and the later versions were much faster.
Microsoft went through the exact same process a long time ago -- WinNT 3.1 was the OS X 10.0 of the 90s, lots of potential but painfully slow. NT 3.5 was faster, and NT 4.0 was even faster still. Contrary to the convential wisdom here, Win XP is slightly faster than all of them, if more memory hungry.
Eventually the core OS is well optimized and there's no magic fairy dust which will give you speed increases. (Especially because Apple is focusing on video-card based processing which by in large is not supported on older machines.)
One issue with the security updates is that Apple has not made it clear how long the official support window is. The updates to 10.1 just stopped one day.
With 10.4 coming out, it's not clear if Apple will want to EOL 10.2, even though there's apparently a substantial userbase still on it. My hope is that Apple makes a formal statement saying how long 10.2 users can expect to recieve security patches.
Let me say then, in my experience, getting any form of Mac testing done on webapps is usually fairly low priority, and I personally don't want to support 2 versions of Safari, especially because I find the JS/DHTML support in the current versions to be far below the competition.
daveschroeder seems to believe in the utter sanctimony of NDAs, and I'm simply suggesting that these things must be considered on a case-by-case basis. With an EDU address, my guess is that dave believes these agreements are more special than they actually are.
I don't see how that contradicts my point at all -- Safari is still treated as second tier, due to missing functionailty. As Apple gets Safari up to speed with all the doodads in Mozilla and IE, it's critical that they move the entire userbase.
Slashdot Mirror
Most big PC OEMs (Dell, HP) ship with Sun Java installed. Also Apple and as you might guess, Sun.
No, COM objects can be written in VB.
Not 5 years ago, Microsoft was pushing an "Enterprise Architecture" called "DNA" which was basically VB or J++ and the MTS/COM+ Application Server. It was their compeitition to J2EE.
(Note: I hate VB, but...) One problem is that VB integrates well with COM applications that are not EOL. For example:
.NET, but it's not nearly as integrated and can be rather clumsy. So it is suprising that VB6 would be EOLed before Microsoft has finished "NETifying" everything.
+ If your shop is automating MS Office 2000 or XP, VB6 is usually the easiest way to do it.
+ If you have an ASP-based web applicaiton, it likely uses VB6/COM for backend components. For example, you may have purchased Microsoft Commerce Server for $$$
Yeah it's possible to do COM stuff with
Long ago, Novell entertained the idea of replacing NetWare with Linux.
Novell actually announced they were replacing NetWare with UNIXWare. The product map whoed this UNIX-based thing called "SuperNOS" would appear after NetWare 4.
Turned out to be a terrible decision because the technical limitations of NetWare killed Novell's dominant standing in the market.
In Unix it is assummed that a program is quite able to look in arbitrary places for it's configuration. The user has to set an environment variable that says "look here", but *that* can be set by another script. Even totally broken programs are fooled into running by symbolic links or by making a script to run it do the necessary rsync
OK, you take these script/envvar/symlink tricks, mulitply by hundreds of applicaitons, multiply by thousands of workstations and the whole thing ends up looking like one big unmaintainable hairball hack attack.
Maybe that's just my Winbrain talking -- but when I do work on *nix systems, it always seems like I'm reverse engineering these tricks, even to get programs running that shipped with the distro.
Maybe a better way to think of it is a "server" mentaility -- custom configurations are required and there pretty much has to be a "pull" model, as you put it. However, when you have thousands of mostly identical workstations, it would seem that "push" is vastly superior to "pull".
I got a copy of the "gold master" of Windows 95 right from a Microsoft marketing shumck about a month early. I kinda doubt that this was a very major accomplishment.
Well, I was talking about NT OSes, not DOS-based ones. 98 uses less RAM than XP, the desktop might be snapper, but I think if you measured things like IO performance, you'd find that it was not at all faster.
That's nice, but maybe next time you can get Steve to stop talking about new product plans (probably a hard thing to do) and instead give us a hint when he will EOL old products.
For example, when I was chatting with the other, fatter, balder and sweaty Steve, he pointed me to a page which states Windows 2000 will go out of support in 2010.
OK let's quote your link: "When the kernel starts to perform swapping it's a sign that the physical memory in the machine has been oversubscribed."
This has absolutely nothing to do with cache -- unless you are condending that OS X is swapping out program segments to increase disk cache, which would be silly, which is why I dismissed it.
It's my observation that OS X tends to need more physical memory than an equivilant Win or Linux system. The "perceptual speed" of a stock 256MB OS X system running mail/web/office tasks is very poor -- basically continually thrashing. Whereas on Windows it's acceptable. Even a 512MB OS X system can be borderline. This corresponds to what people report in Mac-oriented technical forums.
And once you go into swap the perceptual speed is primarily based on Disk IO, not the OS or the CPU. And if you have a slow laptop drive, that just compounds the "perceptual speed" problem. Fine. RAM is cheap.
So, I might as well go back to the original point -- the 8MB minimum specs for NT were a joke, the system would swap like crazy just bringing up the desktop. Realistically NT machines had 32-128MB even back in the mid 90s. Once you eliminated all the swapping, you would find that each NT OS was perceptually much faster than the previous version. I know, I had a Pentium-133 128MB into the Win2000 era.
If I was comparing minimum spec 128MB OS X system, I could equally say they were all slow as shit. But that's exactly what you did with Windows systems.
Oh I know, I ran NT4 on a P133 until 1999. Just trying to be fair to the Mac users here, not that they appreciate it.
The memory model was never misunderstood. An honest comparision between the evolution of OS X and Windows was apparently too much for you to synthesize, and you reacted with nothing but distortions and footstamping.
No, the real (Mac)Troll here is you. Come back when your reigned in your ideology enough to have an legitimate discussion.
I'm sorry, I assumed that the speed hacks put into NT4.0 were well known here. The only insight you display is ignorant zealotry.
Show some data that OS X is "better at multitasking" or STFU about "GUI composition structures" or whatever else you are pulling out of your ass.
Hey, I was wrong, but let me admonish you for not lowering your threshold before replying and therefore being redundant. :)
Dude, MS Word X can't even keep up with my typing on that Powerbook G3. 10.3 or no, it's an obsolete computer.
:P
I'll inform Dave Cutler about this concept of "cache". I'm sure once he learns about it he'll consider adding it to Windows.
I guess the opinion here is that only Apple knows how to write grossly inefficient code and then sell optimization as paid upgrade. Have it your way. Sheesh.
(NT 3.1 on modern hardware would be "snappy" because of the Progman desktop, but as soon as you tried to run any programs you would find that they would be in fact much slower than on XP. IO and video performance would be terrible even if you somehow could find drivers for modern hardware.)
Put succinctly: You have no idea what are talking about. You are apparently in your 30s or older, but you write and think like a 12 year old AOLer. You are an idiot and a flamer. Go away.
Like the other poster, I have an old G3 PowerBook and on that machine, MacOS X 10.3 is "painfully slow and bloated", but not as slow and bloated as 10.0 was. OTOH, both 10.3 and NT4 ran fine on new machines when they came out.
> stupid memory management and GUI composition and multitasking structure
Meaningless technobabble. I will say that OS X is ridiclously memory hungry and that has to limit percieved speed.
NT 3.1 to 4.0 was 3 years, so the timescale is almost identical. And yes, I ran all 3 versions on similar 486/66 hardware and the later versions were much faster.
Microsoft went through the exact same process a long time ago -- WinNT 3.1 was the OS X 10.0 of the 90s, lots of potential but painfully slow. NT 3.5 was faster, and NT 4.0 was even faster still. Contrary to the convential wisdom here, Win XP is slightly faster than all of them, if more memory hungry.
Eventually the core OS is well optimized and there's no magic fairy dust which will give you speed increases. (Especially because Apple is focusing on video-card based processing which by in large is not supported on older machines.)
One issue with the security updates is that Apple has not made it clear how long the official support window is. The updates to 10.1 just stopped one day.
With 10.4 coming out, it's not clear if Apple will want to EOL 10.2, even though there's apparently a substantial userbase still on it. My hope is that Apple makes a formal statement saying how long 10.2 users can expect to recieve security patches.
I think the issue with Google is the lack of "XMLHTTP" support which allows one to make async server calls from javascript.
Let me say then, in my experience, getting any form of Mac testing done on webapps is usually fairly low priority, and I personally don't want to support 2 versions of Safari, especially because I find the JS/DHTML support in the current versions to be far below the competition.
I had beer for lunch.
daveschroeder seems to believe in the utter sanctimony of NDAs, and I'm simply suggesting that these things must be considered on a case-by-case basis. With an EDU address, my guess is that dave believes these agreements are more special than they actually are.
I don't see how that contradicts my point at all -- Safari is still treated as second tier, due to missing functionailty. As Apple gets Safari up to speed with all the doodads in Mozilla and IE, it's critical that they move the entire userbase.