I don't think the NT kernel is really the issue here. The type of tuning MS will probably do will involve stripping the machine of _any_ excess "fat" (unused/irrelevant services/protocols/drivers/options, etc.).
On good hardware NT server is really quite stable (uptimes > 100 days can be common). So the kernel and executive are not a problem. The issue here is speed, so they will want to get rid of anything not relevant to the tests.
I personally find it interesting that MS will use the FAT file system rather than NTFS (if I have been following things correctly, that's what they've decided to do). FAT does not support security, auditing or other advanced (and performance hitting!) NTFS features, plus it is possible using a good defragmenter to tune FAT very precisely. Yet Microsoft themselves advise all shops to use NTFS only on servers and not touch FAT with a barge-pole. Weird. (Or "cunning", if you prefer:)
In all seriousness: not necessarily. If you can boot into another operating system (e.g. from a boot disk) and then access the data on the machine, then that machine has been compromised. A C2-certified NT box (or indeed any secured operating system) can't be breached in this way.
The problem is that Mersenne's formula does not reliably generate primes, i.e. a portion of the primes returned by (2^n)-1 [where n is itself prime] are _not_ primes. Hence all Mersenne primes generated via (2^n)-1 have to be tested to make sure they're bona fide, hence expensive computations required:)
I think it was maybe Bob Metcalfe or Nick Petreley who first mooted the idea of making the APIs not only open, but under the control of an independent moderator (much like the definition of HTML and HTTP is under the control of W3C). I don't know how this would/could work in practice, but in theory it sounds pretty good to me. It could certainly help to encourage emulation of Windows apps on other platforms, which (whether we like it or not:) will likely be a deciding factor in any business OS purchase decision. It could also help prevent Microsoft's apps getting secret hooks into the OS (_if_ indeed they do: Microsoft's always denied this, just claiming its programmers were better at utilising the documented API calls than other companies), thus putting their app developers on an even footing with competing app developers.
I don't think it's fair to require Microsoft to open source any of their products unless _all_ software companies open source their products. If companies want to publish source code (a la Apple [albeit under a GPL-type license], Netscape) then that's great, but I don't think one software company should be singled out and forced to publish its code.
What frustrates me with NT is that every two months things change. Lets start with DDE, then OLE, OLE2.0, then ActiveX, then COM, then DCOM, then COM+. Sure they are all kinda the same, but enough differences are present to frustrate the f**k out of you.
I'm not in love with Microsoft by any means, but you're really trying to say Linux kernels, GUIs, applications, etc. are allowed to evolve but NT isn't? That doesn't seem very fair.
Install a Linux box using up-to-date kernels, desktops and apps, go to the North Pole for two months, come back and see how much of the stuff you installed two months ago has been revised by the community. I don't see a big conceptual difference between that and Microsoft evolving NT.
The evolution in technologies you mention that culminated in ActiveX has taken place over a period of about six years; saying the ActiveX interface changes every two months is a bit of an exaggeration:)
Microsoft... want everyone using Windows NT instead of of crappy 16 bit products. In Microsofts eyes, if everyone were on NT their support nightmares would be greatly reduced. This gives them a huge incentive to try and push people to NT.
Unfortunately, the market didn't agree. So Microsoft was forced to create Win95 and then Win98 in an effort to make the transition to NT more bearable. This effort has also set MS's OS back since they've been spending so much time and effort on compatibility.
I don't think this is quite correct... Microsoft never intended to replace Windows 3.x with Windows NT 3.x, if for no other reason than full 16-bit compatibility was not a Windows NT design goal. (One of Windows NT's original design goals was to "be compatible with _most_ DOS and Windows 16-bit applications".) Microsoft always envisioned a Windows 4.x to replace Windows 3.x.
Only in the past 12 to 18 months has Microsoft started to market the idea of a single Windows base for all its users. It's possible that Windows NT 6.x will be this system.
This effort has also set MS's OS back since they've been spending so much time and effort on compatibility.
I tend to agree with this (even though I have no inside knowledge of how Microsoft's OSs are designed), but on the other hand I don't think Microsoft had much choice: if, for example, Windows 4.x had been a clean, incompatible break from Windows 3.x, Microsoft would have had a very hard time convincing consumers Windows 4.x was a compelling upgrade.
Internet succeeded because Microsoft and other idiots ignored it, so smart people did the development instead of them.
How do you define "succeeded" in this context?
(If you mean achieving a mass market then I think an argument can be made that Microsoft has made a contribution, but I'm not sure that's what you mean. It's not what I would have "succeeded" mean in this context.)
I don't yet believe the Internet has "succeeded" or "failed" - I feel that it is still a work in progress, with Internet2 an evolutionary rather than revolutionary step.
Cheers Alastair
Re:Europeans, Canadians, and self loathing America
on
Catching a breath...
·
· Score: 1
Just a couple of things...
Bell was obviously an American, as was Edison.
Obviously? I don't think so. Bell was Scottish and lived in London, and while Edison spent much of his adult life in Detroit he was actually born in Canada.
Early computers were just an evolution of a chinese invention.
It sounds like you're referring to the abacus, but modern digital computers aren't really directly descended from the abacus. They are pretty much directly descended from work done by Charles Babbage, an 18th century English mathematician.
... but who stormed the beach at Normandy?
There were lots of countries represented at Normandy, including the USA.
I'm sorry but I fail to see how a monstrosity such as visual basic could be important whatsoever.
A lot of commercial development is done in VB5 and VB6 because you can build apps (particularly DB-based stuff) pretty quickly. VB prior to version 5 really did suck, but with the new OO features (sans true inheritance [they have a hacked implementation:P ]) and native code compilation the language is now a lot closer to C++ than most people realise.
VB also lends itself well to prototyping, a common commercial approach to software engineering these days.
I actually disagree with this point (although not the sentiment behind it!). I agree that the _perception_ is that NT server is easy, but I my experience has been that in order to make the thing fast and stable (i.e. what you want from a server) takes a lot of understanding and a fair bit of work.
In that regard I guess it's like Linux... the more you know about the platform, the more you'll be able to get out of it. But I definitely agree that people have a _perception_ that NT is so much easier than other NOSes... I guess Microsoft stokes that perception so they get more sales.
I'm sorry to hear that. We use three Exchange servers across two continents to host about 50 people, and it's worked very well for us. But I guess everyone's mileage can vary.
I agree Exchange can be tough to configure and to tune. Not knowing much about sendmail I'd probably find that app tough to configure too. I guess it boils down to what you know.
. . . the advanced stuff (e.g. VBA) is in there as well.
...Sendmail/Qmail are free so Exchange is simple...
Just a small nitpick: sendmail and its ilk and MS Exchange aren't really equivalent products. I don't think there are any workgroup products available for Linux, although Lotus has made noises recently about making a Linux Notes R5 _client_ available.
...MS Barney for Linux is free...
What? They ported this? *grin*
...Squid; proxy...
It gets better; MS Proxy blows so hard we had to dump it. We use squid on Linux 2.2 now and have never looked back.
1) They make crappy software. 2) They market to the lowest common denominator.
Now don't get me wrong, I'm not a fan of Microsoft by any stretch of the imagination, but I don't think it's entirely fair to make blanket statements like these. We use a lot of Microsoft stuff at the place I work and some of it we find to be pretty good. Microsoft Exchange Server in particular we've found to be a very effective platform for email and collaboration, and we've found NT workstation to be nice on the desktop (once you lock it down so users can't kill it).
I'm not sure what you mean by "market to the lowest common denominator". For example, in Office (not my favourite piece of software;) a novice user is well catered for, but the advanced stuff (e.g. VBA) is in there as well.
I think Gates said "_simple_ word processing and spreadsheets". The clear implication there is that open source efforts can't build quality complicated software quickly. It's our job to prove him wrong.
Getting off-topic, I know, but it's possible that control was a tri-state control and as such TRUE and FALSE were not the only possible outcomes. For example, the default checkbox control in VB (and presumably VC++) is a tri-state control: instead of TRUE/FALSE outcomes, the outcomes are vbChecked, vbUnchecked and vbGreyed.
Slashdot Mirror
I don't think the NT kernel is really the issue here. The type of tuning MS will probably do will involve stripping the machine of _any_ excess "fat" (unused/irrelevant services/protocols/drivers/options, etc.).
:)
On good hardware NT server is really quite stable (uptimes > 100 days can be common). So the kernel and executive are not a problem. The issue here is speed, so they will want to get rid of anything not relevant to the tests.
I personally find it interesting that MS will use the FAT file system rather than NTFS (if I have been following things correctly, that's what they've decided to do). FAT does not support security, auditing or other advanced (and performance hitting!) NTFS features, plus it is possible using a good defragmenter to tune FAT very precisely. Yet Microsoft themselves advise all shops to use NTFS only on servers and not touch FAT with a barge-pole. Weird. (Or "cunning", if you prefer
Cheers
Alastair
Windows NT Server Enterprise Edition can scale to 16 processors, just not very well :)
Windows NT shines on four processors, which is presumably why Mindcraft prefers to use a four processor machine for its test.
It also means that MS presumably can't use SP5 on their NT box. (Not that SP5 contains anything particularly relevant to this scenario anyway...)
In all seriousness: not necessarily. If you can boot into another operating system (e.g. from a boot disk) and then access the data on the machine, then that machine has been compromised. A C2-certified NT box (or indeed any secured operating system) can't be breached in this way.
The problem is that Mersenne's formula does not reliably generate primes, i.e. a portion of the primes returned by (2^n)-1 [where n is itself prime] are _not_ primes. Hence all Mersenne primes generated via (2^n)-1 have to be tested to make sure they're bona fide, hence expensive computations required :)
Cheers
Alastair
Um, Exchange runs almost entirely on X.400, SMTP, POP, NNTP and ETRN - aren't those all open standards?
I think it was maybe Bob Metcalfe or Nick Petreley who first mooted the idea of making the APIs not only open, but under the control of an independent moderator (much like the definition of HTML and HTTP is under the control of W3C). I don't know how this would/could work in practice, but in theory it sounds pretty good to me. It could certainly help to encourage emulation of Windows apps on other platforms, which (whether we like it or not :) will likely be a deciding factor in any business OS purchase decision. It could also help prevent Microsoft's apps getting secret hooks into the OS (_if_ indeed they do: Microsoft's always denied this, just claiming its programmers were better at utilising the documented API calls than other companies), thus putting their app developers on an even footing with competing app developers.
I don't think it's fair to require Microsoft to open source any of their products unless _all_ software companies open source their products. If companies want to publish source code (a la Apple [albeit under a GPL-type license], Netscape) then that's great, but I don't think one software company should be singled out and forced to publish its code.
Cheers
Alastair
What frustrates me with NT is that every two months things change. Lets start with DDE, then OLE, OLE2.0, then ActiveX, then COM, then DCOM, then COM+. Sure they are all kinda the same, but enough differences are present to frustrate the f**k out of you.
:)
I'm not in love with Microsoft by any means, but you're really trying to say Linux kernels, GUIs, applications, etc. are allowed to evolve but NT isn't? That doesn't seem very fair.
Install a Linux box using up-to-date kernels, desktops and apps, go to the North Pole for two months, come back and see how much of the stuff you installed two months ago has been revised by the community. I don't see a big conceptual difference between that and Microsoft evolving NT.
The evolution in technologies you mention that culminated in ActiveX has taken place over a period of about six years; saying the ActiveX interface changes every two months is a bit of an exaggeration
Cheers
Alastair
Microsoft ... want everyone using Windows NT instead of of crappy 16 bit products. In Microsofts eyes, if everyone were on NT their support nightmares would be greatly reduced. This gives them a huge incentive to try and push people to NT.
Unfortunately, the market didn't agree. So Microsoft was forced to create Win95 and then Win98 in an effort to make the transition to NT more bearable. This effort has also set MS's OS back since they've been spending so much time and effort on compatibility.
I don't think this is quite correct... Microsoft never intended to replace Windows 3.x with Windows NT 3.x, if for no other reason than full 16-bit compatibility was not a Windows NT design goal. (One of Windows NT's original design goals was to "be compatible with _most_ DOS and Windows 16-bit applications".) Microsoft always envisioned a Windows 4.x to replace Windows 3.x.
Only in the past 12 to 18 months has Microsoft started to market the idea of a single Windows base for all its users. It's possible that Windows NT 6.x will be this system.
This effort has also set MS's OS back since they've been spending so much time and effort on compatibility.
I tend to agree with this (even though I have no inside knowledge of how Microsoft's OSs are designed), but on the other hand I don't think Microsoft had much choice: if, for example, Windows 4.x had been a clean, incompatible break from Windows 3.x, Microsoft would have had a very hard time convincing consumers Windows 4.x was a compelling upgrade.
Cheers
Alastair
Internet succeeded because Microsoft and other idiots ignored it, so smart people did the development instead of them.
How do you define "succeeded" in this context?
(If you mean achieving a mass market then I think an argument can be made that Microsoft has made a contribution, but I'm not sure that's what you mean. It's not what I would have "succeeded" mean in this context.)
I don't yet believe the Internet has "succeeded" or "failed" - I feel that it is still a work in progress, with Internet2 an evolutionary rather than revolutionary step.
Cheers
Alastair
Just a couple of things...
... but who stormed the beach at Normandy?
Bell was obviously an American, as was Edison.
Obviously? I don't think so. Bell was Scottish and lived in London, and while Edison spent much of his adult life in Detroit he was actually born in Canada.
Early computers were just an evolution of a chinese invention.
It sounds like you're referring to the abacus, but modern digital computers aren't really directly descended from the abacus. They are pretty much directly descended from work done by Charles Babbage, an 18th century English mathematician.
There were lots of countries represented at Normandy, including the USA.
I'm sorry but I fail to see how a monstrosity such as visual basic could be important whatsoever.
:P ]) and native code compilation the language is now a lot closer to C++ than most people realise.
A lot of commercial development is done in VB5 and VB6 because you can build apps (particularly DB-based stuff) pretty quickly. VB prior to version 5 really did suck, but with the new OO features (sans true inheritance [they have a hacked implementation
VB also lends itself well to prototyping, a common commercial approach to software engineering these days.
Cheers,
Alastair
...any idiot can set it up and run it [NT]...
I actually disagree with this point (although not the sentiment behind it!). I agree that the _perception_ is that NT server is easy, but I my experience has been that in order to make the thing fast and stable (i.e. what you want from a server) takes a lot of understanding and a fair bit of work.
In that regard I guess it's like Linux... the more you know about the platform, the more you'll be able to get out of it. But I definitely agree that people have a _perception_ that NT is so much easier than other NOSes... I guess Microsoft stokes that perception so they get more sales.
- WinNT: No clue. Anyone?
:)
This is slated for inclusion in Windows 2000. But then again, isn't everything?
I'm sorry to hear that. We use three Exchange servers across two continents to host about 50 people, and it's worked very well for us. But I guess everyone's mileage can vary.
:) Perl _would_ be nicer, wouldn't it?
I agree Exchange can be tough to configure and to tune. Not knowing much about sendmail I'd probably find that app tough to configure too. I guess it boils down to what you know.
. . . the advanced stuff (e.g. VBA) is in there as well.
Uhhh, you forgot the smiley.
Touche
Cheers
A.
...Sendmail/Qmail are free so Exchange is simple...
...MS Barney for Linux is free...
...Squid; proxy...
Just a small nitpick: sendmail and its ilk and MS Exchange aren't really equivalent products. I don't think there are any workgroup products available for Linux, although Lotus has made noises recently about making a Linux Notes R5 _client_ available.
What? They ported this? *grin*
It gets better; MS Proxy blows so hard we had to dump it. We use squid on Linux 2.2 now and have never looked back.
Cheers,
A.
Why are you hassling a language that's 3 years old? Perl lacked a lot of its current features (e.g. OO) 3 years ago too.
If you haven't used VB in 3 years how can you assert that it's nothing but "code templates and cute little icons"?
Why would you try to program a video game in VB? Would you try to program a CGI-ODBC database app in Lisp?
I'm no fan of Microsoft but it seems to me that comments like this just don't add anything of value to the debate.
1) They make crappy software.
2) They market to the lowest common denominator.
Now don't get me wrong, I'm not a fan of Microsoft by any stretch of the imagination, but I don't think it's entirely fair to make blanket statements like these. We use a lot of Microsoft stuff at the place I work and some of it we find to be pretty good. Microsoft Exchange Server in particular we've found to be a very effective platform for email and collaboration, and we've found NT workstation to be nice on the desktop (once you lock it down so users can't kill it).
I'm not sure what you mean by "market to the lowest common denominator". For example, in Office (not my favourite piece of software ;) a novice user is well catered for, but the advanced stuff (e.g. VBA) is in there as well.
Cheers,
A.
I think Gates said "_simple_ word processing and spreadsheets". The clear implication there is that open source efforts can't build quality complicated software quickly. It's our job to prove him wrong.
Since when is reporting the news a problem?
Getting off-topic, I know, but it's possible that control was a tri-state control and as such TRUE and FALSE were not the only possible outcomes. For example, the default checkbox control in VB (and presumably VC++) is a tri-state control: instead of TRUE/FALSE outcomes, the outcomes are vbChecked, vbUnchecked and vbGreyed.