He is correct, the idea is very nice, but the time configuration should be user contolable, I also don't think that it should scan the system every time, it can just as easily keep a small database and update it when a patch is being installed.
Bad movie: you lose couple of dolars.
Bad DB: you lose couple of thousands of dolars, minimum.
Since configuring a DB is an art, (How much time & money does it cost to get the knowledge, anyway? To be Oracle, or SQL Server, or DB2 official configurator.) making movies is also an art, but watching them isn't.
So, I can buy an AIX, get *less* perfomance, for *more* money, or buy a couple of 2K servers, get *more* performance, *less* money, and more reliability*?
Gee, what would I choose?
* If the AIX fails, my DB is gone, if a single 2K box fails, I only lose some performance, not the whole thing.
Can *any* open source DB do half what SQL Server, Oracle, or DB2 do?
Would you put your life on an OS DB? You bank account? I'm sure that there are many people that can give examples of at least Oracle being used in places where failure cause loss of life and/or lots of moeny.
DB2 and SQL too, for that matter, but Oracle is the most common for huge, faultless, database systems.
Absolutely.
You can come to MS and say, I don't like the terms of the EULA, I want other terms, and I'm willing to pay for it.
It's fully within their rights to say no, though.
If you offer enough money, I don't think that you would find refusal (unless you want 2k GPLed, in which case you would probably need more money than BG)
DB2 has this clause, and there is a *very* good reason why this is neccecary.
It's *easy* to misconfigure a database.
And it can be *very* hard to catch this.
How would you feel if there was another mindcraft test, that shows NT 100 times better than linux, and everything looks just fine in the configuration, software & hardware of the linux box?
What would that kind of benchmark do to people deciding between linux & NT?
That is exactly the reason why there is such a clause in every large DB maker EULA.
I don't know about IBM or Oracle, but MS has in their site free demo versions (fully functional, but stop working after 4 months) of most of their 2K line.
Win2K, SQL2K, Exchange2K.
Probably others, but those were the ones I d/l and tested.
A 4 year re-write of the OS is only a proof of cocept?
Wow, I sure would like to see what they do when they actually implements the concept.
XP is not *that* different from 2000, there is much greater difference between NT4 & 2K than 2K & XP.
Most of the differences are above the hood, new interface, skinable UI, new IE,OE. Expansion of CLI tools, so you no longer need GUI for anything, able to turn of GUI, better TS, IIS, etc.
I would say that they worked 4 years on the behind the scene stuff, released an OS superior to NT, and then went on to work on all the rest.
They only forced them not to remove the IE icon on the desktop, not to avoid installing NS as well.
There was nothing that prevented OEMs from installing NS and giving the consumer the choice.
It isn't that hard to turn NT into a unix look alike.
It's already POSIX compatible, and Cgywin is free.
Services for Unix is 99$, and should work quite well, as far as I hear.
> They could easily require that the CD key be recorded in a databse over the net and find out who is using illegal copies.
Um, I think that this is what XP copy protection is all about.
One could argue that MS has reached large enough market so that now it's MS consider it acceptable to make copying hard.
MS has *always* has been very harsh on mass copiers, but rarely has done anything to the casual copiers.
It may be the reason why they claim that casual copiers cost them so much money.
Of course, one could also argue that XP copy protection scheme will be easy enough to crack. Most products are, if you know where to look. (Google comes to mind, frex).
You mean that you *need* a DVD player in an OS?
Hmm, it would seem to me that OSX went to feature freeze and now being debugged.
This is a good thing.
OTOH, I think I like MS recent attidue better, not announce anything spesific about the dates, and just work on the product until it's as near perfect as times allows.
> I don't know of a single product install under w2k that asks for a reboot
IE (but I can sort of understand that, if not approve), WMP (Why, it's a media player, nothing more), Visual Studio (sometimes not, though), some world atlas (from MS, I think) wouldn't work without rebooting.
All the hotfixes that you can think of.
Why do I have to reboot for a patch in IIS?
Installing 128-bit encryption in Win2K require reboot.
Why do I must reboot when I change swap or registry size?
What about hardware changes? I'm *so* tired from "you've added/removed new hardware, reboot?" messages.
I'm doing a lot of disk swapping, and there is absolutely *no need* to reboot for that. If I answer no, the OS is perfectly capable of handling it. And if I answer no, it does.
As for BSODs, I'd oen with Win2k in nearly a year of using, (IRQL_NOT_LESS_OR_EQUAL), but that was when I was testing new hardware.
I'm currently using Whistler 2296 exclusively, and it's been very stable on me (although I would kill to get rid of the Comments?).
Qoute:
The 'control-alt-del' to login. Someone should pay for that.
On the contrary, there is a good reason for ctrl-alt-del. Unlike any other command keystroke and key combination, ctrl-alt-del cannot be caught programmatically. This means that on NT/2000, you cannot invoke a login prompt unless you are at the console or use something equivalent. This means that there cannot be some program placed on the system that will programmatically invoke a login prompt and brute-force attack the administrator's password. It's arguably more secure than linux's passing of runlevel at the lilo prompt (single mode).
End qoute:
IIRC, there are API calls that lets you verify user password, as well as list all users of the computer.
Gimme a second to find them...
I *love* Google!
http://www.vbsquare.com/api/winpass/
Okay, it lets you find only the current user password.
I'm not sure if you can brute force with it (I would have implemented 100ms delay in the API, but I don't know about MS)
I'm not sure it would work with NT.
I just tried the code sample, (XP beta) and it complained about not being able to find WNetVerifyPassword()
Slashdot Mirror
Then don't install the service pack, install the hot fixes!
Can't tell about NT, but 2000 gives you to uninstall SP & hotfixes via the Add/Remove programs.
He is correct, the idea is very nice, but the time configuration should be user contolable, I also don't think that it should scan the system every time, it can just as easily keep a small database and update it when a patch is being installed.
The programmer that wrote such code deserve not to be hired.
Asking to maintain that kind of code is a reason to quit your job.
Code must be more than understandable to the programmer who wrote it, it must be understandable to other programmers as well.
Bad movie: you lose couple of dolars.
Bad DB: you lose couple of thousands of dolars, minimum.
Since configuring a DB is an art, (How much time & money does it cost to get the knowledge, anyway? To be Oracle, or SQL Server, or DB2 official configurator.) making movies is also an art, but watching them isn't.
So, I can buy an AIX, get *less* perfomance, for *more* money, or buy a couple of 2K servers, get *more* performance, *less* money, and more reliability*?
Gee, what would I choose?
* If the AIX fails, my DB is gone, if a single 2K box fails, I only lose some performance, not the whole thing.
Can *any* open source DB do half what SQL Server, Oracle, or DB2 do? /or lots of moeny.
Would you put your life on an OS DB? You bank account? I'm sure that there are many people that can give examples of at least Oracle being used in places where failure cause loss of life and
DB2 and SQL too, for that matter, but Oracle is the most common for huge, faultless, database systems.
Absolutely.
You can come to MS and say, I don't like the terms of the EULA, I want other terms, and I'm willing to pay for it.
It's fully within their rights to say no, though.
If you offer enough money, I don't think that you would find refusal (unless you want 2k GPLed, in which case you would probably need more money than BG)
DB2 has this clause, and there is a *very* good reason why this is neccecary.
It's *easy* to misconfigure a database.
And it can be *very* hard to catch this.
How would you feel if there was another mindcraft test, that shows NT 100 times better than linux, and everything looks just fine in the configuration, software & hardware of the linux box?
What would that kind of benchmark do to people deciding between linux & NT?
That is exactly the reason why there is such a clause in every large DB maker EULA.
I don't know about IBM or Oracle, but MS has in their site free demo versions (fully functional, but stop working after 4 months) of most of their 2K line.
Win2K, SQL2K, Exchange2K.
Probably others, but those were the ones I d/l and tested.
Well, Apache (why does a web server is called "enemies", anyway) is not designed for speed, it is designed to be standard compatible.
A 4 year re-write of the OS is only a proof of cocept?
Wow, I sure would like to see what they do when they actually implements the concept.
XP is not *that* different from 2000, there is much greater difference between NT4 & 2K than 2K & XP.
Most of the differences are above the hood, new interface, skinable UI, new IE,OE. Expansion of CLI tools, so you no longer need GUI for anything, able to turn of GUI, better TS, IIS, etc.
I would say that they worked 4 years on the behind the scene stuff, released an OS superior to NT, and then went on to work on all the rest.
XP is more than just 2K + luna.
There are a lot of other changes there.
I agree, there is nothing wrong with dual stream files, the later two arguments are balantaly false.
I wish that other OS has multiply streams as well, so far, the only OS that I used that had them was NT.
Mac can only have two, NT can have unlimited.
They already did this, FWIW.
And for myself, I dread the day when MS will be as large as IBM.
They only forced them not to remove the IE icon on the desktop, not to avoid installing NS as well.
There was nothing that prevented OEMs from installing NS and giving the consumer the choice.
I agree about IE not being good enough until 4, but NS 3 was a good product, better than the competion, at least.
It's NS 4 that sucked.
It isn't that hard to turn NT into a unix look alike.
It's already POSIX compatible, and Cgywin is free.
Services for Unix is 99$, and should work quite well, as far as I hear.
> They could easily require that the CD key be recorded in a databse over the net and find out who is using illegal copies.
Um, I think that this is what XP copy protection is all about.
One could argue that MS has reached large enough market so that now it's MS consider it acceptable to make copying hard.
MS has *always* has been very harsh on mass copiers, but rarely has done anything to the casual copiers.
It may be the reason why they claim that casual copiers cost them so much money.
Of course, one could also argue that XP copy protection scheme will be easy enough to crack. Most products are, if you know where to look. (Google comes to mind, frex).
NT is POSIX compliant, and the NTFS (and possibly FAT) can be treated as case sensitive, it's just that most tools don't do it.
add:
#include
or
#include
to the code, btw.
I've tested the following code with 2K & Whislter (AKA XP), but not with NT4.
void main(){
while(malloc(32000));
}
The compiler was VC++ 6, using debug compilation.
It reached the limits of virtual memory (500MB 2K & 1.5GB XP {different settings for virtual memory}) and then it cleaned the mess.
I would hazzard a guess and say that this, on 9x, whould equal fork bomb on linux.
Although it might be the debug compilation that did the trick.
Personally, I would develop on the stable platform, and then test on the least stable platform I can find, make for good coding habits.
You mean that you *need* a DVD player in an OS?
Hmm, it would seem to me that OSX went to feature freeze and now being debugged.
This is a good thing.
OTOH, I think I like MS recent attidue better, not announce anything spesific about the dates, and just work on the product until it's as near perfect as times allows.
> I don't know of a single product install under w2k that asks for a reboot
IE (but I can sort of understand that, if not approve), WMP (Why, it's a media player, nothing more), Visual Studio (sometimes not, though), some world atlas (from MS, I think) wouldn't work without rebooting.
All the hotfixes that you can think of.
Why do I have to reboot for a patch in IIS?
Installing 128-bit encryption in Win2K require reboot.
Why do I must reboot when I change swap or registry size?
What about hardware changes? I'm *so* tired from "you've added/removed new hardware, reboot?" messages.
I'm doing a lot of disk swapping, and there is absolutely *no need* to reboot for that. If I answer no, the OS is perfectly capable of handling it. And if I answer no, it does.
As for BSODs, I'd oen with Win2k in nearly a year of using, (IRQL_NOT_LESS_OR_EQUAL), but that was when I was testing new hardware.
I'm currently using Whistler 2296 exclusively, and it's been very stable on me (although I would kill to get rid of the Comments?).
Sorry for the rant, hope you can answer.
Qoute:
The 'control-alt-del' to login. Someone should pay for that.
On the contrary, there is a good reason for ctrl-alt-del. Unlike any other command keystroke and key combination, ctrl-alt-del cannot be caught programmatically. This means that on NT/2000, you cannot invoke a login prompt unless you are at the console or use something equivalent. This means that there cannot be some program placed on the system that will programmatically invoke a login prompt and brute-force attack the administrator's password. It's arguably more secure than linux's passing of runlevel at the lilo prompt (single mode).
End qoute:
IIRC, there are API calls that lets you verify user password, as well as list all users of the computer.
Gimme a second to find them...
I *love* Google!
http://www.vbsquare.com/api/winpass/
Okay, it lets you find only the current user password.
I'm not sure if you can brute force with it (I would have implemented 100ms delay in the API, but I don't know about MS)
I'm not sure it would work with NT.
I just tried the code sample, (XP beta) and it complained about not being able to find WNetVerifyPassword()