It's different because you can't take GPL code and use it with another license.
MS at least gives you the choice of licenses, you can choose licenses which aren't GPL/like it.
Yuck! MS is more free then FSF? I think I need a lie down.
The part about not using their software in production machines is to prevent you from coming to them whining as say "Look, your evil beta software crashed and all my data is gone."
(Not that you could do that any case, but still.)
The part about not distributing the application talks about *their* application, not your code.
Since the University require a non-exclusive license, you can give them your code under whatever license you want, and also license it under the GPL.
The University can do whatever it wants with the code you handed them, and you can use your code under the GPL, or any other license that you want.
It's no different than the dual licensing of projects.
1) Ms has been using the BSD *API* for its sockets. But then, who didn't. The code itself is not portable to NT (maybe 9x, but I don't think so), because of the different models that they use.
Please note, all of the above are utitilies, not the stack, or winsock itself.
MS did port some utilities from BSD to NT, but that is different from taking the stack, as many seem to claim.
I can't think of any reason why an application would try to write directly to the floppy this way, unless it's something like Rawwrite.
And I don't think that this is how most applications works.
Do you understand the difference between porting a utility than porting that *stack*?
The API difference between BSD & NT are miniscule, the way they are implemented is totally different.
The BSD kernel is a monolitic one, NT's design is closer to mirco kernel, using layered I/O.
Totally different design, the BSD code wouldn't be of much use there.
Actually, I doubt that they could take BSD code and easily port it to NT. Well, not unless it was pretty high level, and even then, they would need some Unix portability layer. (They have several)
What they *can* do is to find out the weak points on NT/2K, etc. And work on them. This can be done in part by analyzing why BSD is better than NT in certain cases, and working on the design of those parts.
Maybe even studying the code to see how the BSD people solved those problems. But the design of Nt & Unixes is different enough to make code portability at the lower levels a PITA.
BTW, while Unix & NT have totally different design models (both in the way they are developed, and in the way they work), they both aim at the same targets, and have roughly the same capabilities.
The FSF disagree with you.
They say, and I qoute: "at least one application program is free software today specifically because that was necessary for using Readline."
No, IE also has a third-party cookies diable feature.
This mean that if I visit forbar.com and there is a DB ad there, the browser *wouldn't* transmit cookies to DB.
> Because TCP/IP is a standard, there should not be performance differences between stacks whereas a stack performs better speaking to another stack of the same design. TCP/IP should be completely interoperable.
TCP/IP is indeed interoperable, but some things will be faster on one system, and some on another.
Because of the different implementation of TCP/IP.
Slashdot Mirror
It's different because you can't take GPL code and use it with another license.
MS at least gives you the choice of licenses, you can choose licenses which aren't GPL/like it.
Yuck! MS is more free then FSF? I think I need a lie down.
--
Two witches watch two watches.
An SDK contain some MS code, header files, mainly.
But headers are enough to GPL the code as well.
So it's not like you don't have *any* code.
--
Two witches watch two watches.
The part about not using their software in production machines is to prevent you from coming to them whining as say "Look, your evil beta software crashed and all my data is gone."
(Not that you could do that any case, but still.)
The part about not distributing the application talks about *their* application, not your code.
--
Two witches watch two watches.
Hm, I know that NTFS has something like it, I wonder if it works over networks, anybody knows?
--
Two witches watch two watches.
It may be a medium sized company in the number of its employees, but it's a big sized company considerring its revenue stream and its profits.
Frex:
Did you know?
MS had more profits than IBM.
--
Two witches watch two watches.
Since the University require a non-exclusive license, you can give them your code under whatever license you want, and also license it under the GPL.
The University can do whatever it wants with the code you handed them, and you can use your code under the GPL, or any other license that you want.
It's no different than the dual licensing of projects.
--
Two witches watch two watches.
The instructor can do it, by requiring that you would use GPL library.
--
Two witches watch two watches.
Too much difference in the low level design of the OS to make it possible.
--
Two witches watch two watches.
1) Ms has been using the BSD *API* for its sockets. But then, who didn't. The code itself is not portable to NT (maybe 9x, but I don't think so), because of the different models that they use.
--
Two witches watch two watches.
Please note, all of the above are utitilies, not the stack, or winsock itself.
MS did port some utilities from BSD to NT, but that is different from taking the stack, as many seem to claim.
--
Two witches watch two watches.
It's how it goes beihd the scenes, yes.
/dev/kmem
\Device\PhysicalMemory -
Etc.
I can't think of any reason why an application would try to write directly to the floppy this way, unless it's something like Rawwrite.
And I don't think that this is how most applications works.
--
Two witches watch two watches.
Do you understand the difference between porting a utility than porting that *stack*?
The API difference between BSD & NT are miniscule, the way they are implemented is totally different.
The BSD kernel is a monolitic one, NT's design is closer to mirco kernel, using layered I/O.
Totally different design, the BSD code wouldn't be of much use there.
--
Two witches watch two watches.
Maybe beacuse Hotmail didn't *start* as a MS project?
a il /default.asp
They bought it when it run FreeBSD, they have been phasing it out for the last couple of years.
You don't just swap servers with something as large as Hotmail.
http://www.microsoft.com/technet/migration/hotm
--
Two witches watch two watches.
Actually, I doubt that they could take BSD code and easily port it to NT. Well, not unless it was pretty high level, and even then, they would need some Unix portability layer. (They have several)
What they *can* do is to find out the weak points on NT/2K, etc. And work on them. This can be done in part by analyzing why BSD is better than NT in certain cases, and working on the design of those parts.
Maybe even studying the code to see how the BSD people solved those problems. But the design of Nt & Unixes is different enough to make code portability at the lower levels a PITA.
BTW, while Unix & NT have totally different design models (both in the way they are developed, and in the way they work), they both aim at the same targets, and have roughly the same capabilities.
--
Two witches watch two watches.
When you've something as large as Hotmail, you don't just throw the old servers out and replace them with your own system at the blink of an eye.
a il /default.asp
http://www.microsoft.com/technet/migration/hotm
--
Two witches watch two watches.
It *is* legal, because the BSD license allows you to do whatever you want with the code.
The thing that they are being bashed about is that they claim that OSS (all of it, not just GPL one) is bad, and then they go and use it.
--
Two witches watch two watches.
Ftp, telnet, etc were ported straight from BSD.
The underlaying code has not, however.
--
Two witches watch two watches.
No, the NT design team was led by the same guy who designed VMS, and he brought many concepts to NT that where learned on VMS.
But NT isn't based on VMS.
--
Two witches watch two watches.
The *API* is the same.
The only places that don't use BSD's API (or a variant of) are *old* Unixes.
There is a reason why it's called Berkley's sockets.
I'm not familiar with the way the 9x TCP/IP stack is built, but the NT one most certainly isn't based on the BSD one.
Taking the BSD code and porting it to NT would be a tougher job than just writing your own, because the way BSD & NT works is totally different.
--
Two witches watch two watches.
The FSF disagree with you.
They say, and I qoute: "at least one application program is free software today specifically because that was necessary for using Readline."
http://www.gnu.org/philosophy/why-not-lgpl.html
--
Two witches watch two watches.
Yes, and it will probably fsck up your CGI.
But there are more illegitimate uses of 3rd-side cookies than there are legitimate uses.
--
Two witches watch two watches.
You can probably sue on false advertisiment, or something like that.
--
Two witches watch two watches.
No, there isn't such a thing as *global* cookie.
That is done for security reasons.
I can't think of a good way to do it that isn't exploitable.
--
Two witches watch two watches.
No, IE also has a third-party cookies diable feature.
This mean that if I visit forbar.com and there is a DB ad there, the browser *wouldn't* transmit cookies to DB.
--
Two witches watch two watches.
> Because TCP/IP is a standard, there should not be performance differences between stacks whereas a stack performs better speaking to another stack of the same design. TCP/IP should be completely interoperable.
TCP/IP is indeed interoperable, but some things will be faster on one system, and some on another.
Because of the different implementation of TCP/IP.
--
Two witches watch two watches.