It just means a virtual address space, like your main memory already uses. The graphics chip will reference polys and textures in a massive virtual address space and the data will be paged in from the 256mb of main system memory when required.
You just need an MMU on the graphics chip somewhere.
Yep, in Longhorn they appear to be moving GDI and some other graphical stuff back into user mode.
Obviously some part of the graphics drivers will still have to be kernel mode to talk to the hardware, but that will be mostly through the Direct3D pipeline not all the bit twiddling that GDI drivers used to have to do in kernel mode.
Looks like GDI rendering will be done in user mode, mostly in software, and then passed as a texture through the 3D graphics pipe.
That only shuts down explorer.exe, not the whole gui. You still have stuff like winlogon.exe and csrss.exe running which depend on the GUI subsystem being present.
A lot of the supposed "undocumented apis" are little more than workarounds for bugs in older versions of Office. I mean why would they put something in Windows 2000 to give special advantage to Excel 5? How exactly does an OS api benifit an spreadsheet anyway? A faster statistical calculation? You might as well build it into the application code.
Other things, like the task panes in the explorer shell are undocumented for the very reason that Microsoft does not want to have to lock down the APIs and maintain them in perpetuity. These things are undocumented because Microsoft does not consider them to be a part of the platform for developers as they want to be able to implement them in an entirely different way in subesquent versions.
They have enough trouble trying to maintain binary compatibility against hundreds of potentially buggy programs written over 20 years against 3 different kernels, dos, 9x, nt.
Read some stuff from Raymond Chen http://blogs.msdn.com/oldnewthing/ who works on compatibility for shell extensions and see how much trouble they have to go to to keep everyone elses code working.
There are special fixes built into Windows for most of the major application developers, Borland, Adobe, Macromedia etc, they do this because they know people won't buy WindowsXP if Photoshop 4 stops working, even if the fault was in photoshop itself for relying on an undocumented internal function or even just a attribute of that particular binary version.
Linux has a much easier job as source is available for virtually all applications. Make some kernel changes, change LibC to match and recompile. Windows programs are virtually all destributed in binary form so the compatibility requirements are much more stringent.
Is that quicktime the program or quicktime the media framework that can be removed? AFAIK any sort of media playback in any application on a Mac uses the quicktime subsystem for rendering.
It would be just as hard to remove that as it would be to remove DirectShow/DirectPlay from windows.
Apple doesn't force retailers to bundle Quicktime? Huh? Only apple itself sells computers with OSX and all OSX computers have quicktime preinstalled. Who are these retailers that can make customised distributions of OSX stripping out Quicktime and swapping in a customised media subsystem?
If I don't want spyware on my system, and I know about the issues I can CHOOSE not to install Kazaa or similar. But the vast majority of people out there may not be aware that spyware exists or its potential for abuse, to them Kazaa is just a way to get something for nothing.
Isn't it right to protect people from corporations taking advantage of their ignorance?
In the perfect capitalist model where everyone has perfect knowledge and can make rational decisions weighing up the relative importance of privacy and conveniance then its OK to leave it to market forces to decide. But the world doesn't work like that. Having a click through license or privacy policy doesn't really work either.
At the very least these discussions serve to make more people aware of the implications of having that much personally sensitive information potentially available to marketers, governments and corporations.
Well the new "Media Transport Protocol" that will be used to connect to mp3 and media players in Longhorn is actually a lot less dependant on the file system of the device than the previous Windows Media Player integration.
http://www.microsoft.com/whdc/hwdev/tech/DMD/Med ia Port.mspx
So you could create a Media Player that uses ext3 or ReiserFS and still integrate perfectly with the new media player and shell.
"create a test environment that is for all intensive purposes "live" "
definately seems more like you were trying to say "practically the same thing in all circumstances" than "the same only in extreme circumstances" to me, why would it make sense to create a test environment that was not generally representative of the system you were trying to simulate?
well "intensive purposes" does not actually mean the same thing as "intents and purposes"
By saying it out loud we can recognise that what you meant was not actually what you wrote, but that does not make what your wrote mean the same as what you think it does.
"intensive purposes" could be seen as meaning something like "the same thing in extreme circumstances" whereas "intents and purposes" is more like "practically the same thing in all circumstances"
You can get the vendor and device IDS to show up in device manager.
At the command prompt set DEVMGR_SHOW_DETAILS=1 then devmgmt.msc to restart device manager
Now the properties dialog of each device will have a "Details" tab that lets you see device IDs plus a lots more information like ACPI capabilities. You can see this for all devices including those marked "Unknown Device" to discover the hardware that you do not yet have drivers for.
Microsoft is working with hardware manufacturers to get them to include a much more descriptive string in their firmware. The "Unknown device" entries will be replaced by a unicode string chosen from a lookup table of several languages.
I agree The code names are more interesting than the final product names. Freestyle = Windows XP Media center edition Mira = Windows powered smart display Media2Go = Windows mobile for portable media centres
It think there are signs that they may begoing to do something like that. When a non-priveleged app tries to write to a restricted area of the registry the call can be redirected to store the setting in a file within the users profile. The next time that exe is run the value stored in the config file will masquerade as the registry entry for that key. This will only be a per-app setting for compatibiliy, to allow legacy apps to be run without breaking under non-admin accounts. New program written in.Net tend to store configuration in an xml file in the users profile anyway. The GAC replaces it for componant registration and it is being deprecated as a means of storing program configuration.
parseing xml files is a lot slower than calling a function to read an in-memory hierachical store, but it will be safer and more robust.
Thats just not true. The difference between 10.0 and 10.3 could be reckoned to win95-ME, but comparing osxs incremental releases to the fundamental shifts between 3.1 and 95, and 98 and XP is just over the top. When if osx ever jumps to an entirely different kernel, API or memory model in a.1 release you may have a point.
How do you read encrypted mail through a web interface?
Don't you need to cut and paste the contents of the email into a client program and supply the key to decrypt it? Doesn't that defeat the point of using a web based interface to begin with?
Can you keep your key on the local file system and have a jscript decrypt the mail locally? Can you be sure that the script is not uploading your key to the server.
Or you can do it with a GUI. Control Panel -> Administrative Tools -> Local Security Policy -> Local Policies -> Security Options -> Shutdown: Clear Virtual Memory Page File -> Enabled
All Microsoft products are xxx Next. Yukon is SQL Server Next, Whidbey is Visual Studio Next, Longhorn is Windows Next. Its just a way to refer to the version currently being worked on.
It just means a virtual address space, like your main memory already uses. The graphics chip will reference polys and textures in a massive virtual address space and the data will be paged in from the 256mb of main system memory when required.
You just need an MMU on the graphics chip somewhere.
Completely restarting the GUI would require winlogon to be restarted. Thus it can't be done without a reboot.
Posix and OS/2 subsystems were removed from Windows XP. Though you can now get the posix one back for free buy installing SFU 3.5
l /XP_ker nel.mspx#XSLTsection136121120120
See
http://www.microsoft.com/whdc/driver/kerne
Yep, in Longhorn they appear to be moving GDI and some other graphical stuff back into user mode.
Obviously some part of the graphics drivers will still have to be kernel mode to talk to the hardware, but that will be mostly through the Direct3D pipeline not all the bit twiddling that GDI drivers used to have to do in kernel mode.
Looks like GDI rendering will be done in user mode, mostly in software, and then passed as a texture through the 3D graphics pipe.
That only shuts down explorer.exe, not the whole gui. You still have stuff like winlogon.exe and csrss.exe running which depend on the GUI subsystem being present.
A lot of the supposed "undocumented apis" are little more than workarounds for bugs in older versions of Office. I mean why would they put something in Windows 2000 to give special advantage to Excel 5? How exactly does an OS api benifit an spreadsheet anyway? A faster statistical calculation? You might as well build it into the application code.
Other things, like the task panes in the explorer shell are undocumented for the very reason that Microsoft does not want to have to lock down the APIs and maintain them in perpetuity. These things are undocumented because Microsoft does not consider them to be a part of the platform for developers as they want to be able to implement them in an entirely different way in subesquent versions.
They have enough trouble trying to maintain binary compatibility against hundreds of potentially buggy programs written over 20 years against 3 different kernels, dos, 9x, nt.
Read some stuff from Raymond Chen
http://blogs.msdn.com/oldnewthing/
who works on compatibility for shell extensions and see how much trouble they have to go to to keep everyone elses code working.
There are special fixes built into Windows for most of the major application developers, Borland, Adobe, Macromedia etc, they do this because they know people won't buy WindowsXP if Photoshop 4 stops working, even if the fault was in photoshop itself for relying on an undocumented internal function or even just a attribute of that particular binary version.
Linux has a much easier job as source is available for virtually all applications. Make some kernel changes, change LibC to match and recompile. Windows programs are virtually all destributed in binary form so the compatibility requirements are much more stringent.
Is that quicktime the program or quicktime the media framework that can be removed? AFAIK any sort of media playback in any application on a Mac uses the quicktime subsystem for rendering.
M acOSX/Co nceptual/SystemOverview/SystemArchitecture/chapter _43_section_2.html#//apple_ref/doc/uid/20000980/CH DBJCFH
See:
http://developer.apple.com/documentation/
It would be just as hard to remove that as it would be to remove DirectShow/DirectPlay from windows.
Apple doesn't force retailers to bundle Quicktime?
Huh? Only apple itself sells computers with OSX and all OSX computers have quicktime preinstalled. Who are these retailers that can make customised distributions of OSX stripping out Quicktime and swapping in a customised media subsystem?
This is great, someone mod this up please.
If I don't want spyware on my system, and I know about the issues I can CHOOSE not to install Kazaa or similar. But the vast majority of people out there may not be aware that spyware exists or its potential for abuse, to them Kazaa is just a way to get something for nothing.
Isn't it right to protect people from corporations taking advantage of their ignorance?
In the perfect capitalist model where everyone has perfect knowledge and can make rational decisions weighing up the relative importance of privacy and conveniance then its OK to leave it to market forces to decide. But the world doesn't work like that.
Having a click through license or privacy policy doesn't really work either.
At the very least these discussions serve to make more people aware of the implications of having that much personally sensitive information potentially available to marketers, governments and corporations.
No on Windows its just "c:\windows\fonts\", how different is that?
Well the new "Media Transport Protocol" that will be used to connect to mp3 and media players in Longhorn is actually a lot less dependant on the file system of the device than the previous Windows Media Player integration.
d ia Port.mspx
http://www.microsoft.com/whdc/hwdev/tech/DMD/Me
So you could create a Media Player that uses ext3 or ReiserFS and still integrate perfectly with the new media player and shell.
They have been having difficulties getting things off the ground
hey thats funny
keep up the good work
Someone mod this funny please.
Thanks
"create a test environment that is for all intensive purposes "live" "
definately seems more like you were trying to say "practically the same thing in all circumstances" than "the same only in extreme circumstances" to me, why would it make sense to create a test environment that was not generally representative of the system you were trying to simulate?
well "intensive purposes" does not actually mean the same thing as "intents and purposes"
By saying it out loud we can recognise that what you meant was not actually what you wrote, but that does not make what your wrote mean the same as what you think it does.
"intensive purposes" could be seen as meaning something like "the same thing in extreme circumstances"
whereas "intents and purposes" is more like "practically the same thing in all circumstances"
exactly
right it as thay here it
You can get the vendor and device IDS to show up in device manager.
At the command prompt
set DEVMGR_SHOW_DETAILS=1
then
devmgmt.msc
to restart device manager
Now the properties dialog of each device will have a "Details" tab that lets you see device IDs plus a lots more information like ACPI capabilities. You can see this for all devices including those marked "Unknown Device" to discover the hardware that you do not yet have drivers for.
Microsoft is working with hardware manufacturers to get them to include a much more descriptive string in their firmware. The "Unknown device" entries will be replaced by a unicode string chosen from a lookup table of several languages.
It should be "for all intents and purposes", thats what he's getting at. Nevermind, you're not the first person to write it as they here it.
I agree
The code names are more interesting than the final product names.
Freestyle = Windows XP Media center edition
Mira = Windows powered smart display
Media2Go = Windows mobile for portable media centres
It think there are signs that they may begoing to do something like that. When a non-priveleged app tries to write to a restricted area of the registry the call can be redirected to store the setting in a file within the users profile. .Net tend to store configuration in an xml file in the users profile anyway.
The next time that exe is run the value stored in the config file will masquerade as the registry entry for that key.
This will only be a per-app setting for compatibiliy, to allow legacy apps to be run without breaking under non-admin accounts.
New program written in
The GAC replaces it for componant registration and it is being deprecated as a means of storing program configuration.
parseing xml files is a lot slower than calling a function to read an in-memory hierachical store, but it will be safer and more robust.
Thats just not true. The difference between 10.0 and 10.3 could be reckoned to win95-ME, but comparing osxs incremental releases to the fundamental shifts between 3.1 and 95, and 98 and XP is just over the top. .1 release you may have a point.
When if osx ever jumps to an entirely different kernel, API or memory model in a
How do you read encrypted mail through a web interface?
Don't you need to cut and paste the contents of the email into a client program and supply the key to decrypt it?
Doesn't that defeat the point of using a web based interface to begin with?
Can you keep your key on the local file system and have a jscript decrypt the mail locally? Can you be sure that the script is not uploading your key to the server.
Windows can do that itself anyway.
s /q 182/0/86.asp
http://support.microsoft.com/support/kb/article
Or you can do it with a GUI. Control Panel -> Administrative Tools -> Local Security Policy -> Local Policies -> Security Options -> Shutdown: Clear Virtual Memory Page File -> Enabled
All Microsoft products are xxx Next. Yukon is SQL Server Next, Whidbey is Visual Studio Next, Longhorn is Windows Next. Its just a way to refer to the version currently being worked on.
The 40 million lines of code is for the whole system. That would be like the kernel, xfree86, KDE, Mozilla, and all the system tools and utilities.