Teflon Computing

[jeff_duntemann]

Pete Albrecht sent me a link to an article that means well but doesn't quite hit the bullseye. No one seems to be asking, Is Vista really necessary? What does all that complexity actually buy us? Why is Vista better than XP? Or how about that now mostly abandoned but still pertinent

Comments

the case for something new

[anonymous]

I contend that the innumerable complaints about Windows security issues are fundamental to the notions embodied in .NET. Moreover, I suspect that the research by MS has shown that high on the hit list is the problem of memory leaks, now to be solved by GC. Adding .NET onto the existing Win32 structure will be, at best, a partial fix. The better path will make .NET and GC intrinsic parts of the OS.

An argument against the need for something new might have been offered at the time of NT4, or Win2K, both of which were hugely better than the Win9X variants. But nothing remains constant; entropy increases. MS must either continue to make what they see as improvements, or lose ground to alternate OSes.

For my part, I have tried Linux, and found it wanting. Mac has never really been an option; I like -- indeed need -- open hardware. The most promising OS of the last decade was BeOS, which actually represented a largely new approach. But Be Inc. couldn't tough it out, and the pairing I would have considered ideal -- Delphi on BeOS -- was too dicey a notion for the folks in Scotts Valley. Lacking a "killer app", Be went down the tubes.

My main complaint with the evolution of Windows is that the more horsepower we get, the more MS wastes on merely painting the screen. On the other hand, the decoupling of the GUI from the OS in Linux (and for that matter, OS-X) makes for a rather sluggish experience.

Is Vista the best answer? Not in my view. But something new is needed, and evolution is preferable to stasis.

Re: the case for something new

[jeff_duntemann]

Maybe I'm just too old, but I consider ever-increasingly complex screen pyrotechnics a big drawback for operating systems. I like seeing one window at a time whenever possible--helps me focus.

The other major problem Windows has had is trying to suck every last little useful utility into the OS itself, and to tie every component to every other component in nonobvious ways. This makes the OS harder to extend back at the source, where MS has had huge difficulty making Vista happen.

I still haven't seen how either XP or Vista will make my everyday work as a writer/editor/publisher easier or faster. To the contrary, the snappy performance of Win2K on a 3 GHz machine with loads of RAM puts the performance of XP to shame.

MS OS performance is basically running in place, as complexity rises to the point where even MS has trouble making it work. OS/X is a good example of how to make a user-centric OS out of Unix, but I don't see anybody in the FOSS world trying to learn from Apple's example.

By the way, I liked BEOS too.

Re: the case for something new

[anonymous]

I don't know what the future will bring in terms of an OS, but the way things are going, I anticipate we're not far from having 2GB display cards with three times the processing power of that needed by the OS. One of the problems is that, 25 years ago, it was hard (for me) to imagine a processor being able to do anything useful to a video signal in real time. So much for my crystal ball!

I recall back around 1983, in a discussion with a young friend who was then a CS major at UCSC, complaining about the way functions were being stuffed into the hybrid CM/P - CDOS OS our client was using. He replied that it was easy to make the case for the OS to provide practically any service you could imagine, from date and time calculations to the more esoteric math needed for television timecode. It appears that such was the thinking in CS circles of that day, and Windows has clearly inherited from that philosophy. And yes, Windows does happen to provide functions in support of television timecode -- but not the math needed .

Then, of course, comes the legacy support issue. If every Tom, Dick and Harry is going to scream bloody murder when an old app no longer works, then how can we expect MS not to follow the lead of IBM in supporting -- even by emulation, if need be -- all the old function calls of the not so lamented DOS?

As an aside, I note that the HP49g+ calculator, now on an ARM processor, emulates the prior Saturn, to preserve the investment in the old software, at the expense of speed. No, Sir, it is NOT my old HP-25.

Odd that you find XP sluggish. I have Win2K on a 2.5GHz P4 with 1GB, and XP on a 1.8GHz Athlon with 1GB of RAM, and I find that XP is more responsive. Not dramatically so, but just a bit less sluggish.

I wish I could run BeOS on either of these machines, but of course, it won't support the hardware.

As to the FOSS world, which is still so much enamored of *nix... how could they be expected to be innovative, when they venerate chiefly an OS and architecture now 30+ years old. And yes, I appreciate that it has been successful on servers, but on the desktop, I'm afraid Linux still basically sucks. Running the GUI as an app is something we got past with the advent of Windows NT. And the "beauty" of running apps on the server where they reside was chiefly attractive when a potent computer cost more than most people earn in a lifetime. Sorry, but it's a model whose time has passed.

Now, on the other hand, BeOS did something new. And it did so in a lean, mean way that the Borland barbarians should have supported. IMHO.

Bill Meyer