W
O
R
D

Just thought I'd get a word in edge-wise.... gee you kids are silly.

Jim, that is funny!

Longhorn was very much a Bill Gates work, and Longhorn eventually ended up as Windows Vista.

Thank you for making my point.... Longhorn was the next Windows with real new features that many professional users were eagerly looking forward to (the new file system, etc.). Then in the course of events every single exciting new feature in Longhorn was dropped, one by one, until Vista remained.

The CEO doesn't take part in the design of an operating system. That is the job of the Chief Software Architect, the role Bill Gates took when he handed things over to Ballmer.

Of course not, but the CEO still has the highest responsibility for the Execution of what the company does. If the development of the company's next #1 cashcow is in trouble, it is his responsibility to make sure it is fixed. If he doesn't, he'll be replaced by the board, and if the board won't do it soon enough, the shareholders come running to replace the board.

...a large company has to rely on process, not on people. People come and go, process has it's own forward motion.

History is full of large companies that relied only on process.

Process is very important, and if you're running a family restaurant where say the short order cooks are totally interchangeable, I would agree it's the only thing that matters.

If you're running a creative enterprise, you soon find that some "people" (many bosses don't like this term, they prefer to refer to the meat as "human resources") are 5 to 50 times more productive than others.

After that you find from experience, over and over again, that there are some rare people who can singlehandedly develop 80% of anything over a weekend of focused work.

And you discover that it takes these same people a year to develop the remaining 20%...

So you find a way to combine these people doing what they do best, with other specific people that you pick because they are able to work with the sometimes somewhat socially inept characters who develop 80% of anything over a weekend.

And then you find that with a small team of carefully selected people (and processes), you are able to beat teams of 250 developers at the biggest companies in the world. And you do it many times over.

The above is from my own experience, but I have seen the same thing in many other companies.

So I trust we can agree that both people and process are necessary, and one without the other cannot work.

For an example of people without process, you just need to look at "All Star Bands" which is a very old concept that sounds great until you try it.

For an example of process without people, take, umm, never mind (so I don't have to mention the current, hopefully temporary, state of Microsoft again).

Unix, probably one of the best and worst things ever to happen to the computer industry.

I've been using Unix since the early years, and still appreciate the almost poetic beauty of its design. I always liked the way everything connected with pipes and streaming instead of blocks, and the way you could create a chain of I/O-ed (it is a word, right? :O) processes in the most amazing way on a command line.

QNX seems very cool, I wasn't familiar with it. Now you made me want to build something with it...

Microkernels are beautiful too, and will spread even more when Windows 7 replaces what may be the last behemoth monolith.

Ooooh, now I understand that scene in Planet of the Apes!!! The monolith was Vista!!! That explains everything!

:O)

Terje

The main feature that was dropped from Longhorn was WinFS, and it was dropped somewhere in early 2004. Long before Bill stepped down from anything at all. He was very much the chief architect then.

If the development of the company's next #1 cashcow is in trouble, it is his responsibility to make sure it is fixed.

Not everything is fixable from the top actually, particularly if you have taken over a company with a certain "culture". The problem is that the culture at Microsoft doesn't lend it self to quality and predictability, which is a problem that is the responsibility of Mr. Gates. Ballmer can try to fix it, but it is going to be very hard, and there is no way he could even start such a thing while Bill was still there.

If you're running a creative enterprise, you soon find that some "people" (many bosses don't like this term, they prefer to refer to the meat as "human resources") are 5 to 50 times more productive than others.

Absolutely, which is why large companies generally are unable to perform. No matter who is in charge. As I said, as a large company you have to rely on process, and if you are in an industry where process is not going to deliver you are screwed.

So I trust we can agree that both people and process are necessary, and one without the other cannot work.

The problem is that since a large company is forced, by its size, to mandate process, the superstars will leave. They always do. This is why large companies and organizations can not deliver. The only way they can is by spinning off stuff and buying it back, a process CISCO is probably the brightest example of. An example that works some of the time.

I've been using Unix since the early years, and still appreciate the almost poetic beauty of its design

I have been using Unix since Sun was BSD and not SysV. The simplicity of Unix was brilliant in its day, and as an example of KISS it was beautiful. The problem is, it is no longer in any reasonable way possible to say that there is anything KISS about Unix, and since almost everything else about Unix is not particularly well designed, there is no longer anything particularly beautiful about Unix, IMHO. For operating system beauty I tend to look to QNX and Amoeba (not related). The big, fat, bloated monolithic mess that is Unix doesn't even come close to cute, let alone beauty.

The beauty and terror of Unix was "everything is a stream of bytes". Your screen is. The disk is. The network is. Everything. Everybody copied it and we ended up where we are today, with horrid file systems that are still trying to catch up with what DEC had 20 or so years ago. It was a sad-sad day wen DEC, one of the most innovative computer companies ever, was acquired by Compaq, the ultimate copy-cat.

For reasons I can not understand, but that probably has to do with the extreme pragmatism of Bill Gates, David Cutler didn't influence Windows as much as I would have thought he would. Or perhaps he just grew old and tired.

Terje,

You have a competent and useful perspective on Vista, the Myriad Organization, and Unix.

So let's see if I can talk some sense into ya!

:O)

There were plenty more architectural features of real substance in the first Longhorn versions, but they all had to be dropped because the programmers were drowning in spaghetti.

You could certainly argue about whether that was a process problem, a people problem, or both. You could even argue that Bill Gates was in charge of the Longhorn design, even though Jim Allchin became "the Longhorn Vice President" after he was successful with Windows NT and more (and DEC before that).

ComputerWorld: Longtime Windows development chief James Allchin wrote in a January 2004 e-mail to Microsoft CEO Ballmer and co-founder Gates that the software vendor had "lost sight" of customers' needs.

I don't have time to pull the list of everything that was dropped from Longhorn to get Vista out while Bill G. was still alive, but I remember it being substantial.



I'll have to agree with this, with the qualification that it is possible but may not be worthwhile (because you have to replace a majority of the staff and move the offices).

I remember two large organizations that were facing this problem: The National School Board of Sweden, and ditto in New Zealand. The organization was rotten to the core, but the government had to make it work, so they did the only thing that could fix the problem:

They closed down the entire organization and fired all employees. Then two months later, they announced the formation of "The School National Board" and hired fresh people. Everyone lived happily ever after (except the grumps of the old Dilbert-shaming organization).

Of course Unix has gotten more complex, and the elegance of the CLI gets lost in the GUI era (although I'm happy to have a full BSD shell at my disposal anytime in OS X).

...large companies generally are unable to perform. No matter who is in charge. [...] As a large company you have to rely on process,

While this could easily be seen to be generally true, look at Adobe, where they are idolizing their top developers. Not just putting their names up for everyone to see in the splash screens of their products, but with actual appreciation shown in many ways inside the company.

As we know, this hasn't led to perfection, but if Adobe had been relying only on process, they would have been sold off bundled with WordPerfect.

Of course streams are not appropriate when you are seeking out maximum performance. But don't forget that this made I/O so simple, that a lot of new software wouldn't exist today without it.

Huh? Before streams, the block I/O of the day was not trivial on large machines and each app had to write a lot of I/O code. The simplest of I/O such as printing out "Hello Cruel World" on a line printer hooked up to an IBM 360/370 main frame was taught in the SECOND YEAR of programming school.

I wrote a lot of block I/O routines for max. performance, for example in a Z80 microprocessor interface to a 3280 reel-to-reel tape drive, allowing a desktop computer to read and write mainframe computer tapes. There was so little time between blocks flying in, that it was mathematically impossible to process character by character what was coming in, even CPU block transfers couldn't do it fast enough, but I finally found a sneaky DMA mode that made it possible to get in under the wire, and this became a hot selling product.

I'd say that this success was a result of "people with process," but I also found in managing a lot of superstar talent (in several fields) that it was always worthwhile to adapt the creative development environment to their way of working, even when their way wasn't "correct" in the sense of the latest fad in programming (and there have been a lot of those).

And David Cutler? I think he just didn't have a champion at Microsoft, somebody with enough clout to remove company internal obstacles before him so he could be effective. People without process, you could say.

DEC was really a fantastic company in its heyday. I was both a customer and a supplier to them, so I saw quite a bit of them.

The last DEC machine we bought was an 8600 to replace the VAX 11/785 we had used around the clock for 3 years (the 785 cost us something like $700,000 new, when we wanted to sell it 3 years later, the best bid was "$500.00," which turned out to be what we would have to pay to have it removed and processed as toxic waste.). Sigh.

Since this is extremely OT now, I'll end it here and cede to you that Bill was innocent in the mess that started as Longhorn and ended up as Vista.

Since I have two relevant new items to add to this OT thread....

1) I have two IDENTICAL laptop computers, except that one runs Vista and the other XP - both 32-bit. EVERY application runs slower on the Vista laptop than on the XP laptop. EVERY ONE. As Dvorack says in this month's issue of PC Magazine: "Vista is a prettier and slower version of XP"

2) My wife works for a multinational company with tens of thousands of PC's. The IT department says on their corp internal website that they have "absolutely no plans to switch to Windows Vista before 2010." Serious incompatibility with thousands of internally written programs and the huge expense to upgrade thousands of PC's to the minimum hardware that Vista requires is mentioned as reasons why not. They also mention that it would take months to complete the transition to Vista and during the intervening time, incompatibilities of versions of Microsoft's own applications between XP and Vista would create new problems of their own as users try to exchange documents. They further state that this would be a multi-million dollar expense with at best, a negative ROI. They conclude that "... there is no business advantage to the upgrade [to Vista] for [the] Corp. None." At the end of the page is the most succinct, most sensible business decision ever made: "If it ain't broke, don't fix it".

I was not supposed to be able to see the internal webpage, but her laptop was open to the site and I saw the link. That is why the company is not being revealed.

"Vista is a prettier and slower version of XP""

yes this is true. Vista is slower for more than just reasons of prettiness. The new search feature for example takes a fair bit of horsepower to run. The thing is though... all the neat and pretty things (including search) can be turned off if you want speed, and with Vista trimmed down it's every bit as fast.... and in some cases even faster.

XP doesn't have all these extra things, so you really have no choice but to leave them off :)

John

Have you tried HDLink to capture/simultaneously convert HDV with one of your Vista 64 machines? Any issues?

Thanks
Mark