Paul Thurrott has posted yet another
look at Windows Vista. Ever since the first alpha and beta releases
of Longhorn/Vista hit the web, Paul has been giving us regular
updates on the progress of the operating system. Paul's articles are
usually for the most part positive with a little hint of negativity
thrown in where appropriate.
article though lays everything out on the line when it comes to
Vista. Now that Vista is supposedly feature complete and many things
will stay as is when the final product ships, promises that Microsoft made in regards to features in the
operating system, usability issues and application blunders are now
fair game. Here, Paul rants about missing features that Microsoft promised:
There are so many more
examples. But these two, WinFS and virtual folders, are the most
dramatic and obvious. Someday, it might be interesting--or depressing,
at least--to create a list of features Microsoft promised for Windows
Vista, but reneged on. Here are a few tantalizing examples: A real
Sidebar that would house system-wide notifications, negating the need
for the horribly-abused tray notification area. 10-foot UIs for
Sidebar, Windows Calendar, Windows Mail, and other components, that
would let users access these features with a remote control like Media
Center. True support for RAW image files include image editing. The
list just goes on and on.
I must say, I've tried and tried to
give Vista more than a second glance. I've tried every beta release
that Microsoft has issued, but every time I find myself being less
productive and utterly frustrated using the operating system compared
to Windows XP. Fortunately, it looks like Microsoft has a few more months to get some of these issues under control.
quote: So you are basically totally unhappy with current Vista situation, yet feel confident that totall rewrite of the code ground-up by the same very company will magically change this state...
quote: AFAIR they took some parts of the BSD kernel and hired folks from BSD team when they were "designing" NT ...it's that simple.
quote: Writing new kernel is anything but simple, no matter how much money you put into it. It takes time, patience, vision etc.
quote: If it were _that_ simple as your post suggests why Microsoft haven't done it already for Vista ?
quote: And where are the figures that show how Vista scales with number of cpu's increasing ?
quote: From what I know Linux kernel _easily_ scales up to 32 cpu's (with regard to tasks that require interoperability between cpu's), and probably well beyond it, so there is kernel that may do just what you ask for
quote: What about internet ? I thought is't about TCP/IP stack, which is what ? 25 years old (hint - not everything that is old is worth s#%@)
quote: "pervasive wireless" - isn't it a job for drivers regardless of the kernell ?
quote: "no-execution bit" - and I thought that Service Pack 2 already added support for this...
quote: "e-commerce" - web browsers role to make it happen, along with conforming to standards
quote: "SSE2/SSE3" - what does it have to do with kernel ? Do you ask for specific optimisations for this instruction sets to be made in kernel ? They are multimedia specific (hint - in kernel there will be no improvement), it's up to the apps to utilize it.
quote: "Intel VT and AMD Pacifica" - and your point is ... ? They will make it into Vista
quote: "all the new hardware" - drivers _again_
quote: "gigabit" - what about it ?
quote: "3D accelerators in every mobo" - you will have your eye-candy in Vista so what's your point ? Isn't modern accelerators fully supported on WinXP already (by respective manufacturers) ?
quote: "NEW WORLD requires a NEW KERNEL" - no it does not ! - it's Microsoft that need's one :) (joke)
quote: WinXP was made on a basis of Win2000, Win 2000 on WinNT and WinNT used BSD code... when exactly someone wrote an entire kernel from scratch that made it to mainstream ?
quote: Linux kernel for example is constantly being developed for the last 15 years and it is work in progress...the point is - this can be made, but takes time and a lot of it for that matter.
quote: So what ... do you want to have every task to be at least 2 threds ?
quote: XEN is not ready yet, nor is it a Microsoft's baby.
quote: In what way not ditching support for 32 bit makes platform suddenly dated ?
quote: Don't like floppies ? - don't use them
quote: Drivers for COM, LPT and PATA don't make kernel that much bigger do they ? (definitely not slower !)
quote: support 3D accelerators from the start (HAL layer for 3D hardware anyone?)
If Microsoft starts to manufacture graphic cards then yes. Otherwise - no... It's the manufacturers who provide drivers for their stuff.
quote: Dave Cutler has nothing to do with BSD, sorry to disappoint you.
Care to share a link to where I claimed it ?
quote: Microsoft just took a part and incorporated it into their kernel
quote: After you patch the code there is no distincition from the situation when you would make look the code look like this from the beginning.
quote: And example of linux kernel shows that this problem can be easily overcome.
quote: Yes, WPA supplicant and hostapd USE hardware drivers for Wi-Fi chipsets, but would you include 802.11i support in the DRIVER or in the KERNEL???
What about driver in the kernel ?
quote: And you would design whole system around "no-execution bit" ?
quote: and when the whole industry moves forward to another more secure "standard" this whole pile of legacy shit within kerenel would do what ? - sit and lough in your face... How would you like it then huh ?
quote: 4) 3D audio processing, other hi-def audio algorithms
Yes - definitely this would make sever admins endlessly cheer with this " advanced" "feature".
quote: I can imagine some appliances (like speeding some disk operations assuming that ethernet transfer>> disk throughput), but certainly nothing realistically beneficial/killer feature. And certainly nothing that would warrant writing new kernel for.
quote: You know why GPU's are not commonly used in general computations ? Becouse vendors won't open up their spec. Your "stupid" example in no way support your former post.
quote: Where are sutudies about feasibility of such things and ... why on earth Microsoft havn't hired you already ? You present such deep insight to the subject, that it just blows my mind.
quote: In what way keeping support for legacy stuff makes platform suddenly new and exciting? I didn't say anything about it being exciting. It's just real life necessity.
quote: You know that virtualising "costs" - the currency being computer resources? Stupid nonsense - as you said.
quote: It takes development resources and time away from really important task of implementing things noone implemented before, and this is BAD, so screw floppies, PATA and other artefacts like these.
No it doesn't. Nobody is working on support for floppies - there is no development on it - just support.
quote: And please note that this standard appeared WITHOUT MS manufacturing any 3D hardware. Sorry to disappoint you here again, buddy.
Sorry for not being that much into gaming.
quote: I read once that they (MS) hired some BSD folks, but I cannot find any prove it as of now. Not talking about Dave Cutler :)
quote: in reality it's more difficult to drive whole project with them to different architectural design
quote: But they don't start _everything_ from the beginning - maybe some stuff has to be removed once in a while, but nothing THAT radical as you claim. The overall kernel design stays untouched.
quote: If for example the second computer being yours sister's she wouldn't be very pelased to see her task crawling becouse you started ripping your porn :)
quote: DirectX API is not meant to perform general computations. It is good for transforming triangles, but reverse discrete fourier transform computations are not made that easily on it, right ?
quote: DOS run doesn't has to run fast, since everything that was written for had been done with slow computers in mind, now we have much faster rigs. Don't know details about MacOS9 emulation so I won't comment on this one.
quote: Maybe because there is no need for it ? What exactly task would you perform on GPU's on network ? Gaming won't do, becouse of latency. Rendering is OK - but it is ALREADY being done by apps, so ... ? And why this god damns DX ?!? If you would like a real "standard" than pick OpenGL - it's platform independent and more suitable for "seriuous" tasks
quote: is a powerful command line tool (POSIX compliant shell ?) , new file system without constant need for defragmentation (ext3 ? - neee - that would be to good to be true), easy support for other files systems (JFS, XFS, ReiserFS), more flexibility during installation (e.g. more MBR options, more other systems awarness), changing registry in sth. of more ellegant design, enforcing and better handling of multiple accounts (now amost everyone on desktop is using root accounts), more modular design (separating text mode from graphical interface?, those modules you've mentioned...)