Post a Comment
1) apart from aero I didnt see anything else new ?
2) The search functionality is better then spotlight ? are you sure about that?
3) NX in avaliable in winxp ?
the funny thing is if you follow wikipedia link to new features it only has 3 ??? and even those arent great ? so.... why upgrade unless you just want to ? there isnt anything really significant ?
hard to tell realy. only thing that i have picked up so far is that its using managed code for the whole system (iirc).
still, that could make backwards compatiblity simpler. in that when a program askes for a old interface, it will be handed over to a special sandbox where it can go nuts. but if it dies, nothing will happen to the rest of the system...
There are lots of new OSs being developed in research. There is no reason to believe that any more willc ome out of Singularity than has come out of all the other research OSs that have preceded it. The only reason Microsoft exists today is because of backwards compatibility. Without the enormous market advantages of application compatibility, Microsoft never would've made it through the 1990s. Without application compatibility, Microsoft is nothing, and they know that. They'll never field an OS that isn't compatibile with the vast body of Windows software that exists.
Yes, we all know about singularity; and like all R&D things, this was mearly a proof of concept; a rectum pluck along the lines of 'what would we do, if we could design something from scratch'.
Why not base a new operating system off Plan9? I mean, it corrects all the problems and limitations of UNIX; tries some new things; it would be simply a matter of Microsoft fixing up the GUI, and voila, you have the problem solved with minimum fuss.
Deploy it first as a server OS, then gradually start to sell it as a client operating system as new programmes come on board; backwards compatibility can be done via VirtualPC.
This statement makes no sense. The .NET framework is a runtime for writing WINDOWS programs, not a separate OS. Versions of the .NET framework can be written to run on other OS's, but on WIndows, the runtime converts C#/VB.NET/managed C++ code to the apropriate API calls
Now that Microsoft has said it would include hardware virtualization techniques into Vista, why didn't they do software virtualization to keep legacy APIs separate from the core system?
Something a la Win32s or WINE, but sandboxed into a lightweight virtual machine, where the sandboxed kernel would forward translated API calls to the host kernel.
It is often said that an added layer of abstraction has at least a 10% impact on performance. But isn't a one-shot 10% slowdown acceptable since we're talking about a next-generation operating system bound to be better-performing on next-generation computers, which by the powers of Moore Law will just trump the 10% hit with a 50% increase in performance anyway? It's still a net increase of 35% in the first hardware refresh...
Don't take me wrong, this is an Athlon XP machine and the other computer by my side is a Pentium 133. I'd be glad if Vista could degrade gracefully to support both computers to some decent level. However I wouldn't really mind if the system core could run in "autobahn" mode most of the time and "city traffic" when I run a legacy app.
What ticks me off is that all the compatibility layers added to Windows make it run in "bumpy countryside dirt road" all the time. Getting rid of it might as well more than make up for the theoretical 10% speed hit.
Edit: clarified hardware/software virtualization distinction
Edited 2006-04-23 18:22
Now that Microsoft has said it would include hardware virtualization techniques into Vista, why didn't they do software virtualization to keep legacy APIs separate from the core system?
The Win32 subsystem is separate from the core system. I don't see how moving to a virtual machine solution would offer any meaningful benefits, given that it will still need to be able to access the user's files and documents.
How come Apple has compatibility cleaner and less bloaty then Windows has?
Apple's backwards compatiblity is a laughing stock. Applications written for Panther had to be updated to work on Tiger, for instance. And in order to load a pre-OSX app you need to load up OS9!
I wouldn't call that 'cleaner and less bloated'.
99% of the apps I used on 10.1 run just fine on Tiger, including large app suites like Office X. It is usually the reverse that is an issue: apps that take advantage of features exclusive to Tiger won't work with older OS versions. Most apps written today are designed to run in Panther & Tiger, some offering more functionality when runing on Tiger (Spotlight integration, Core Graphics use, etc.)
So your solution of splitting up the beast that is windows seems just a little naive.
If Windows were to divide into a backward compatible version and a non-backward compatible version, that would possibly mean that the two "current" versions in your view of the world, might not even be compatible.
Imagine the consequences of that:
1) Microsoft developers would need to be divided between the two versions.
2) 3rd parties would have to divide their time between the two versions.
3) Users would be even more confused as to which version they should buy.
4) Less people working on each version equates to slower development.
Clearly all of these would show microsoft in a bad light, and risk the company losing a great deal from the fall out. So how exactly do you think splitting up the process would help?
Maybe its just me but your posts seem to be rushed and follow quickly on from other posts on other sites, or maybe I'm just biased against you because you kept announcing SuSE releases and stating there was no release notes when a bit of research would have shown you where they were.
More haste, less speed?
1) Microsoft developers would need to be divided between the two versions.
4) Less people working on each version equates to slower development.
The legacy version would need little development-- since all that needs to be done on that version is bug fixes and security fixes. It's a legacy version-- no need for new features. And since the modern version does not require work on legacy stuff, greatly simplyfying the development, the end result would be that no more developers are needed than there are now.
2) 3rd parties would have to divide their time between the two versions.
That did not stop the transition from 9x to NT, now did it? It didn't stop Apple's move from PowerPC to Intel, now did it?
3) Users would be even more confused as to which version they should buy.
Welcome virtualization, as someone else has already mentioned. The modern version could allow users to install their old 2000/XP copy, to run the legacy apps. Similar to what Apple did with OS9. And since all new PCs will ship with the modern version, developers will think twice about not updating their apps to run on this new version.
The legacy version would need little development-- since all that needs to be done on that version is bug fixes and security fixes.
In that case they should not create a version of Windows with legacy support. They should just maintain a current version of Windows (say XP Home) indefinitely.
Unfortunately this wouldn't make a great deal of fiscal sense, they'd lose money from the lack of Windows and Office sales (the big one-two that pays for almost everything in Seattle). MS already complains that their greatest competition comes from their past products.
For what it's worth, they've already done what you suggest. Windows 2003 Server lacks most of the compatibility stuff that Windows XP has, as far as I know. However selling it as a desktop operating system, even for businesses, would be very difficult. They tried that in the mid-nineties with NT 4 for business users and Windows 9x for traditional users, and the cost of keeping two code-bases in sync was deemed to be counter-productive.
The only reason I (have to) "care" about Vista is because I work on a Windows network. At home I don't care about it at all because I run FreeBSD and Gentoo, and use one Windows 2000 laptop for the occasional game.
Our German colleagues are already looking for people who want to pilot the latest Office with Vista, I wonder what the outcome of that will be.
"You can take an application designed for Windows 95, and run it without any problems on Windows Vista."
Backwards capability is a myth.
We supported a bank's front counter system written on Win 95 using standard Win API, C++ and Foundation classes and also a lot of C and VB programs - not all of those would run when we tried to run them on XP. The ones that didn't run had to be rewritten and that then, in turn, required the rest to upgraded and enhanced to work in a proper mutli-tasking environment.
So backwards capability in large systems is a complete non-starter. I bet you'd never be able to take a complete system written on XP and put it on Vista and expect it to run with no problems.
Does it make you feel good saying that?
Call me back when a Linux distro can hibernate and resume my laptop out of the box. The last time I played with Linux's suspend it hard-locked on the resume, and ended up corrupting my ReiserFS partitions. I was real pleased.
In fact, call me back when wireless, 3D acceleration, MP3, and DVD work out of the box on either of my two laptops.
LOL.
Linux works with what it works with, and doesn't with what is doesn't. I suspect strongly that (checking title of post) that Linux significantly more compatible than Vista upon launch.
I would argue strongly that there are very few laptops people are using that currently that have hardware compatible or upgradeable to that to Vista.
DVD's on any Microsoft from the get go is a non-standard by a third party, and MP3's are a little tricky from the get go as for some reason MP3 wants to promote its own format
You cannot blame Linux for that. Blame the hardware vendors that don't release drivers and specifications for the hardware they make. I have an HP NX 8220 laptop and it it is able to hibernate when I use the proprietry ATI FGLRX driver but it doesn't resume from sleep mode. If I use the open source driver I am able to sleep/hibernate. That's because unlike some other hardware vendors, HP actually give a shit about their Linux users.
Edited 2006-04-24 04:51
Sure I can blame Linux for that. Pretend I'm a regular user for a second. I'm not going to care that it's so-and-so's fault and this-and-that company's hardware not working with Bob's driver.
All I'm going to see is Linux not hibernating. If Windows and OS X can do it, but Linux can't, it looks pretty bad.
Sure I can blame Linux for that.
You could, but you'd be mistaken.
Pretend I'm a regular user for a second.
Uh, no. Let the regular users speak for themselves.
I'm not going to care that it's so-and-so's fault and this-and-that company's hardware not working with Bob's driver.
You don't know that for a fact. Most "regular" people would complain about the hardware, because they don't know what an OS is.
All I'm going to see is Linux not hibernating. If Windows and OS X can do it, but Linux can't, it looks pretty bad.
If you're buying a laptop and you're a clueless user, you won't install an OS on it. Regular users don't install OSes. Again, you're comparing apples and oranges to suit your anti-Linux agenda. It's getting really old.
BTW, hibernating does work on many Linux laptops.
Most USB devices work out-of-the-box. On my Kubuntu laptop I've tried USB drives, iPods, digital cameras, scanners and printers, and they've all worked.
YMMV, of course, but I think you've just been unlucky (or have problems with your hardware setup). USB support in Linux is excellent, whatever's missing is the equipment manufacturer's responsibility.
Then again, it's easier said than done. And they would probably have made some users get pissed off with the outcome.
But it's hard -- with tens of thousands of employees, Microsoft can't coordinate anything anymore.
Where I don't agree, is that Windows Vista isn't needed in my opinion. People should just use Linux and MacOSX. Forget about the ease of use of Windows, because it's easy for you and for the hackers and bad guys in general. Ask yourself: how good are you? And the people you know? Would you trust your data/computer to all the people you know? Windows is too "permissive" to bad guys.
Can Linux run all the programs that are available to Windows users? Nope. But on the other hand, your data/computer will be safer.
I understand that users want/need the Windows applications, but it's a double edged sword -- innocent people will get cut.
The same innocent people want to run the Windows programs on Linux, but it's not going to happen, not without the consent from Microsoft.
Do you know about the Catch-22? If you keep using Windows, developers will continue to create Windows-only programs and you will never be able to leave Windows. On the other hand, if you start using the alternatives (Linux), the developers will be pushed to support Linux and then you will be able to choose whatever OS is best for you. Until then, you are stuck in a Catch-22.
What happens if you make Windows much more secure and much more stable.
Linux has the same old problems that Unix had for the most part so it has a lot of legacy there as well.
Why do we need Linux? If Windows runs all of my applications and linux doesn't and Windows is what everyone uses then what is the point?
Users don't have any reasons to move to Linux after Vista comes out.
so _what_ is it you wanted to write about? you are indifferent about vista, because you're not _not_ missing something in vista? i've no idea what you're talking about, really, but i guess i'm indifferent about your article. i'm not not missing anything not about not it. *nod*. thx.
"Memory usage was comparable to ordinary XP's as well."
No way. I have tested several of the builds of Vista available to developers & partners. In each case a freshly booted clean install logged in and running only task manager has showed over 500Mb for the commit charge.
I recently did the same check on an XP SP2 system, and it was 130Mb.
I wouldn't class 4x the memory usage as "comparable".
I wouldn't class 4x the memory usage as "comparable".
During standard usage (one Firefox window, MirandaIM, Google Desktop, AV), my XP leaves about 230MBs of my mem untouched. Vista left the same untouched, running the same amount of apps (excl. Google Desktop obviously, as Vista has its own equiv.).
What I found strange was the large difference in memory usage when I resized a simple Explorer window. Windows would consistently use about 30-40MB more RAM when I resized, say, a Control Panel window from a quarter screen to nearly full screen. The resize was not very smooth, either. This was on a dual-core Athlon with 1GB RAM. Radeon X300 video, though, which is a little weak.
Also, to be fair, this was the old beta, and the 64-bit version. I haven't tried the new beta yet. But my point is that not everyone has had good experiences with Vista resource use.
The main thing I noticed about Vista is that isn't shipping with the real stuff that was to make it useful for power users such as the SQL based file system. I have read that it was so slow MS had to remove it for acceptable performance. I am sure all these great new features will get in there eventually but the 1.0 release seems kind of a let down. Sure there is the great new aero glass look with compositing features but then again Vista users are complaining they can't tell which windows are in the front and back now. I wouldn't doubt Vista users will have to wait for service pack 3 before all the real Vista components get added in. A lot of the bloggers out there are saying that MS really dropped the ball on Vista. I heard the main problem with Vista was the design plan. Originally it was coded as an extension to the XP code but it got so complex and buggy MS had to start over again from scratch. They lost a lot of time when they did this and the incomplete version of Vista is what resulted. Supposedly the media center code, which was supposed to be a highlight of Vista, isn't stable yet.
http://www.theinquirer.net/?article=30516
Edited 2006-04-23 19:06
Yep. Vista is a disappointed and Microsoft dropped the ball on it.
What people don't seem to understand though, is that it can be a disappointment, but STILL have things to look forward to it and STILL be a large improvement over XP. The problem is that it was so hyped up, it simply couldn't live up to expectations, and anything short of a masterpiece will have people bitching about how bad it is.
Unfortunately not that way. It was the C++ programs that held the infrastructure of the system together that caused the first problems. The VB apps were bad enough on Win95, you just can't tell some people not to develop complex apps in VB. Fortunately I got out of that system upgrade. They thought that could use a XP workstation as a file/print spooler in each site - they had fun and games in sites where they had more than 10 workstations.
Will Vista workstation only allow a max of 10 connections like Win XP? That is such a rip off - forcing you to by a server edition.
Nice article, sir. I'll admit, though, that I get blurred vision and a headache whenever Vista is mentioned. To be fair, I'll feel the same about a new Mac or Linux OS upgrade too.
Until the new boy hits the streets, it is all talk talk talk and pie in the sky. We need a real OS with real and final features/decisions, then a wait of six months for the worst bugs to be squashed, before it's worth forming any view. Just my 2 cents.
In my case the other reason is that I think "omg, no" when hearing about Vista's (or Mac OS's) 196,758 new and exciting features. I know the vast majority will be to hand-hold Joe and Jane User as they explore their new world, and flash them dire warnings if they step outside it. I equally know that I don't want or need most of these features, which slow everything down, preferring stuff that is simple, fast and slick. The legacy issue lies not only in the code-base but in the assumptions about how much the user knows about computers and computing.
The idea of a legacy-free Windows is extremely appealing. Personally I don't think it will happen, but it would be great fun if it did (and probably fabulously good for internal morale at MS, too, as their best people get to show what they can really do).
Couldn't you really find another subject to write your column about?
This story starts as a good intro to a Vista compared to *insert other OS here* or something similar but then you bring up some lame topic such as backwards compatibility! As if you really don't have a clue on what you are going to write about today.
I think it's a good idea to let someone else write a Sunday Eve Column once in a while, for those sundays where you don't have anything to say. I love coming to OSNews but this kind of articles are a waste of my time because they really don't have any value what so ever. I think there are a lot more people out there with some useful information and opinions to share with the OSNews readers. Maybe you could invite someone to write one.
And by the way, why should Linux engineers learn from backwards compatibility designs implemented in Windows? Isn't Linux backwards compatible enough?
Edited 2006-04-23 19:07
Couldn't you really find another subject to write your column about?
Yes, I could.
bring up some lame topic such as backwards compatibility! As if you really don't have a clue on what you are going to write about today.
You post started out pretty well, but then you bring up some lame topic such as as if I don't have a clue what I'm going to write about!
Explain why you find the topic lame. Saying "I say it is lame, therefore it is" is pointless.
Isn't Linux backwards compatible enough?
I'll just pretend I didn't read that, ok?
binary incompatibility for modules and applications across various glibc implemenations, ABI changes, gtk changes, qt changes, deprecated kernel features, removed kernel features...
Yes, that's lack of backwards compatibility for Linux distributions as well as the kernel.
I agree with everything you said, except I believe it can be done much better than by creating a new division. Apple tried that with the Apple III/Lisa/Macintosh.
Apple had it right, however, with the introduction of OS X: System 9 Compatibility layer.
You know, with all this hullabaloo about virtualization on the chip and everywhere else (which isn't really hullabaloo ;P), creating a legacy environment within New Windows (for lack of a better name) would appear to be cake.
One of the big issues with Windows' backwards compatibility is that they need to reimplement all the bugs. How ridiculous! The beauty of the proposed method above is that Windows^2 (better name ;P) wouldn't need to include those legacy bugs and holes. All of that would be self-contained in a legacy guest OS. Any new applications developed for Windows^2 could then take advantage of the new, (hopefully) clean environment.
The idea is so simple, I can't believe they aren't doing it. They really do need a shift in management over at Microsoft.
Share what you're smoking with the rest of the class, Mr. Spicoli.
Desk Accessories were NOT programs, but DRIVERS. The old MacOS lacked multitasking, and as such, this hack was invented to give the MacOS the illusion of multitasking, since multiple drivers could be loaded at once (while programs couldn't).
Thom is correct on this one. Go read http://en.wikipedia.org/wiki/Desk_Accessory
Whilst desk accessories used the driver queue to get some processor time on a regular basis, it's a stretch to call them "device drivers".
Yes they used a hack but they were still tiny footprint programs offering widget like functionality, they were just activated by the driver queue.
The method to get them to activate might be considered a "hack" but to suggest they were not programs doesn't seem right.
This would be like suggesting some TSRs in the old DOS days were not "programs" because they were activated by interrupts to gain multitasking functionality not built into DOS.
Yes and if you go to a store with Apples what do they have on screen? iTunes. So what do you tie the iPod with, iTunes, what do you link iTunes with? Apple Macs and OS X. Its a complete package that fits neatly together. I go into peoples houses and what do I see? More and more Macs.
The DOS emulation is getting worse in XP (Had some apps stop working after patching) and I expect it will be worse again in vista.
I don't blame microsoft for this, who the hell wants to support an OS that they considered obsolute 11 years ago?
However they are the ones with the old MSDOS code, so It sort of falls upon then to provide some compatability.
The DOS / Win16 subsystems should move to an emulatated or virtualised layer. DOSbox already allows this to happen, but microsoft should do it in a way that can be supported (For some reason my management doesn't want us supporting a application that hasn't yet had it's 1.0 release).
A cutdown Virtual PC (doing emulation) + MSDOS 6.22 would allow microsoft to do similer to what apple has done with classic. Where the legacy environment is contained. Emulating it would have limited effect as the software was written for computers that are orders of magnitude slower then the computers today. With the added benifit of never having to really worry about security in the system affecting the OS, since it is all an application running in userspace.
But we all know this is not going to happen, microsoft would not want to make such a financial commitment because they don't have to. If a dos application stops working, the application is blamed, even if the OS has changed it's dos subsystem.
The DOS / Win16 subsystems should move to an emulatated or virtualised layer.
Unless I'm mistaken there is no DOS or Win16 subsystem in Vista. It has been removed completely. Windows 3.1 and DOS programs will not run period. They did this with Windows XP x64 edition also I believe.
Unless I'm mistaken there is no DOS or Win16 subsystem in Vista. It has been removed completely. Windows 3.1 and DOS programs will not run period. They did this with Windows XP x64 edition also I believe.
DOS has always been emulated. It's still there in Vista. Win1






What comes of it remains to be seen.
