Let's say you want to rename all *.jpeg files to *.jpg. How would you do that in Windows? In VMS that would be a piece of cake, in a Unix system it's more complicated, for i in *.jpeg; do mv $i `echo $i | sed s/jpeg$/jpg/ - ` ; done or something like that would do it, but the easiest way to do it in Windows that I can think of would be a VB program.
You picked an unfortunate example here... In Windows it would be "rename *.jpeg *.jpg" (sometimes the shell not expanding wildcards is a blessing).
Does the new licensing for Vista Enterprise only apply when using Virtual PC, or can you also install up to 4 extra copies when using VMWare's solution?
I know this is/., so you can't be bothered to RTFA, but it seems now even a 4-line summary is too much. From the summary (you know, the blurb right at the top of this page): Even better, nothing in the license requires that Microsoft Virtualization technologies be used - if you want to use a competing product as your Virtualization solution, you still get the four extra licenses for use with VMs."
You can always run a Tor node yourself. That way, it's impossible to distinguish between traffic which originates from you and traffic which you're just relaying.
When you use Tor, you don't build a direct connection from your endpoint to the site you want to visit. Instead, you build an encrypted connection from your endpoint to a Tor server. That Tor server in turn builds an encrypted connection to another Tor server. And, you guessed it, that Tor server builds an encrypted connection to a final server, the exit server. That exit server finally builds a connection to the target you wanted to connect to.
Snooping at your ISP won't reveal what your final connection target is. It will only show that you're connecting to a Tor server. Snooping at the exit server will only reveal that someone is connecting to the target, it won't reveal your IP address (the exit server doesn't know your IP address, it only knows the IP address of the Tor server before it).
I know you're being sarcastic (and you even got a chuckle out of me), but my question was quite serious. It's easy to make blanket statements like "The Windows architecture sucks", but what I'm trying to figure out is exactly what people think sucks.
Personally, I think the architecture is quite nice. It is the implementation and the business practices of MS that leave a lot to be desired.
If the bug/non-standard behavior is present in all NT-derived Windows versions then we simply do what Windows does. If it was fixed (or introduced) in later Windows versions, we need to look at the compatibility setting for the particular program (which NT version does the program expect?) in order to determine whether or not to emulate the bug.
The code we re-used from Wine certainly helps to speed up development. But we can't re-use everything and have to develop a lot of stuff ourselves. For example, Wine uses X11 primitives to do BitBlts (bitmap copying). Since ReactOS is not X11 based we had to write that stuff ourselves. Also, in Wine, everything is in userspace, while ReactOS follows the NT4 (and later) design where there is a split between userspace and kernelspace, with most of the work actually being done in kernelspace.
Still, we try to co-develop as much stuff as possible with Wine. Ofcourse, since Wine has been around much longer and the number of Wine contributors is larger than the number of ReactOS contributors most of the shared code originates in Wine.
Ge van Geldorp, ReactOS developer.
Slashdot Mirror
On my screen the summary is 6 lines. Please don't exaggerate. I'm not that lazy.
:-).
Sorry, in good Slashdot tradition I didn't bother to actually count them
Does the new licensing for Vista Enterprise only apply when using Virtual PC, or can you also install up to 4 extra copies when using VMWare's solution? I know this is /., so you can't be bothered to RTFA, but it seems now even a 4-line summary is too much. From the summary (you know, the blurb right at the top of this page): Even better, nothing in the license requires that Microsoft Virtualization technologies be used - if you want to use a competing product as your Virtualization solution, you still get the four extra licenses for use with VMs."
You can always run a Tor node yourself. That way, it's impossible to distinguish between traffic which originates from you and traffic which you're just relaying.
When you use Tor, you don't build a direct connection from your endpoint to the site you want to visit. Instead, you build an encrypted connection from your endpoint to a Tor server. That Tor server in turn builds an encrypted connection to another Tor server. And, you guessed it, that Tor server builds an encrypted connection to a final server, the exit server. That exit server finally builds a connection to the target you wanted to connect to.
Snooping at your ISP won't reveal what your final connection target is. It will only show that you're connecting to a Tor server. Snooping at the exit server will only reveal that someone is connecting to the target, it won't reveal your IP address (the exit server doesn't know your IP address, it only knows the IP address of the Tor server before it).
Of course, with svn you can just create a release branch where only bugfixes are allowed.
I know you're being sarcastic (and you even got a chuckle out of me), but my question was quite serious. It's easy to make blanket statements like "The Windows architecture sucks", but what I'm trying to figure out is exactly what people think sucks.
Personally, I think the architecture is quite nice. It is the implementation and the business practices of MS that leave a lot to be desired.
Windows over complicated and badly designed architecture(s)
What would you have done differently if you were Dave Cutler?
The ReactOS cmd.exe will run happily on Windows. Not all features are implemented though.
If the bug/non-standard behavior is present in all NT-derived Windows versions then we simply do what Windows does. If it was fixed (or introduced) in later Windows versions, we need to look at the compatibility setting for the particular program (which NT version does the program expect?) in order to determine whether or not to emulate the bug.
No, not at this moment anyway. We're concentrating more on getting the features in than on speed.
Yes :-)
The code we re-used from Wine certainly helps to speed up development. But we can't re-use everything and have to develop a lot of stuff ourselves. For example, Wine uses X11 primitives to do BitBlts (bitmap copying). Since ReactOS is not X11 based we had to write that stuff ourselves. Also, in Wine, everything is in userspace, while ReactOS follows the NT4 (and later) design where there is a split between userspace and kernelspace, with most of the work actually being done in kernelspace. Still, we try to co-develop as much stuff as possible with Wine. Ofcourse, since Wine has been around much longer and the number of Wine contributors is larger than the number of ReactOS contributors most of the shared code originates in Wine. Ge van Geldorp, ReactOS developer.