Sivar

Windows Longhorn in 2006

Recommended Posts

Links here and here.

Microsoft is scaling back Longhorn (again) in order to make a 2006 release date, and will ship an update with WinFS after Longhorn's release. Possibly two years later.

According to Infoworld, "...Microsoft has set a new timetable that includes a major OS release every four years, with an interim update following two years behind.", thus the WinFS update may come about two years later.

Share this post


Link to post
Share on other sites

That's sad; I guess it all comes down to money instead of quality. There was supposed to be quite a few enhancements in technologies; then the setback’s what's left is a watered downed version. What if anything is going to be new? :unsure:

Share this post


Link to post
Share on other sites
I guess it all comes down to money instead of quality

That really sounds like a premium scientific extrapolation.

Especially in the software industry, quality is money. Marketing helps, surely, but is not a replacement.

You can be the leader, spend a lot of money in marketing, if your product sucks, you will lose money and stop being the leader (even if the later may require some time, depending on how far you were leading).

Share this post


Link to post
Share on other sites
You can be the leader, spend a lot of money in marketing, if your product sucks, you will lose money and stop being the leader (even if the later may require some time, depending on how far you were leading).

I agree with you quality is money. Does anyone think Windows ME was a quality Operating System?

Share this post


Link to post
Share on other sites
I agree with you quality is money. Does anyone think Windows ME was a quality Operating System?

There are two distinct points here.

The first is the definition of quality. The fact that Win16 OSs had stability issues (in today standards) have almost nothing to do with quality. They fulfiled their specifications. Those specifications were designed conforming to what their target was expecting.

The second point is : how much money did microsoft with Windows Me ?

Actually i have no clue, but for sure they spent more time working on it than selling it.

The fact that so much code were reused is an evidence of quality. You cannot reuse crap. You just cannot maintain it. There were a lot of both win16 and win32 code reused.

Any senior software designer will confirm that quality code costs less than crap code.

Any senior software designer will confirm that crap code is not maintainable and hardly reuseable.

But i know that this kind of subject will lead nowhere but to a flame war. So i won't try to convince you (nor even go on with this subject). Learn a little bit about software design or keep your hype position, it's up to you and i don't mind...

Share this post


Link to post
Share on other sites
I agree with you quality is money. Does anyone think Windows ME was a quality Operating System?

There are two distinct points here.

The first is the definition of quality. The fact that Win16 OSs had stability issues (in today standards) have almost nothing to do with quality. They fulfiled their specifications. Those specifications were designed conforming to what their target was expecting.

The second point is : how much money did microsoft with Windows Me ?

Actually i have no clue, but for sure they spent more time working on it than selling it.

The fact that so much code were reused is an evidence of quality. You cannot reuse crap. You just cannot maintain it. There were a lot of both win16 and win32 code reused.

Any senior software designer will confirm that quality code costs less than crap code.

Any senior software designer will confirm that crap code is not maintainable and hardly reuseable.

But i know that this kind of subject will lead nowhere but to a flame war. So i won't try to convince you (nor even go on with this subject). Learn a little bit about software design or keep your hype position, it's up to you and i don't mind...

I might know very little as you stated, but I believe Windows NT was a completely different platform from Windows 98 or ME and the NT Platform has been the one built upon. If I'm wrong I would like to learn. :D

Share this post


Link to post
Share on other sites
You can be the leader, spend a lot of money in marketing, if your product sucks, you will lose money and stop being the leader (even if the later may require some time, depending on how far you were leading).

That isn't necessarily the case. Look at the success of Symantec's Norton Utilities, many Microsoft products, Easy CD-Creator, Intervideo's WinDVD, Bose, and Ford Motor Company.

That's sad; I guess it all comes down to money instead of quality. There was supposed to be quite a few enhancements in technologies; then the setback’s what's left is a watered downed version. What if anything is going to be new?

In this case I do not understand this argument. I think it is exactly about quality--if Microsoft were to release the product with their original ambitions intact, I doubt they would have the time to effectively test it. Would you rather have an incredibly flaky OS with loads of incredibly flaky features, or a mildly flaky OS with a number of enhancements that actually work?

Share this post


Link to post
Share on other sites
That's sad; I guess it all comes down to money instead of quality. There was supposed to be quite a few enhancements in technologies; then the setback’s what's left is a watered downed version. What if anything is going to be new?  :unsure:

Engineering is always about the trade off between quality vs time and money. It is very likely this type of decision is based on how those portions of the project are going and what the feedback from testing is.

If nothing is new it isn't a new version.

Share this post


Link to post
Share on other sites
In this case I do not understand this argument. I think it is exactly about quality--if Microsoft were to release the product with their original ambitions intact, I doubt they would have the time to effectively test it. Would you rather have an incredibly flaky OS with loads of incredibly flaky features, or a mildly flaky OS with a number of enhancements that actually work?

I'm just curious about what the enhancements are; all I've really heard is what's not going to be there and that's the three major improvements.

Share this post


Link to post
Share on other sites
The fact that so much code were reused is an evidence of quality. You cannot reuse crap. You just cannot maintain it. There were a lot of both win16 and win32 code reused.

The vast majority of Win16 code in Win9x was re-used, as-is, precisely because it was hand-written ASM in most cases, and totally unmaintainable. The workaround that they instituted, in attempting to wrap a psuedo-pre-emtive 32-bit multitasking OS kernel around what was still really a 16-bit protected-mode ASM code core, was to institute an outer global mutex to prevent re-entrancy problems in that ball of single-threaded code, called the "Win16Mutex". I'm sure that you've heard of it. Code re-use is no evidence of quality, nor maintainability. Only the ability of said code to be hacked at enough to be re-used. It was also the biggest single point of weakness in the Win9x OSes.

Any senior software designer will confirm that quality code costs less than crap code.

I just can't really believe this. Developers get paid by their time, not by the number of lines of code that they write (at quality shops, that is), and writing better, more maintainable, re-usable code takes time. More time that just spewing barely-working non-reusable crap code.

Oh, actually.. if you are speaking of the overall lifetime costs of the code, including future maintainability, rather than just the initial costs of production, then yes, I would say that you are correct, sorry.

Not that many code shops recognize that fact though, which is probably one of the reasons for the interest in outsourcing - firms only look at the initial creation cost of the code, instead of the maintainability and longer-term lifetime costs of the codebase. Maybe, like everything else, firms are looking more towards "cheap" and "disposable".

Any senior software designer will confirm that crap code is not maintainable and hardly reuseable.

Yes, but if the company wants the code now, and wants it at a low cost, or the lowest cost possible (outsourcing), then that code produced will be "good enough", even if it is crap code on a more global scale. Who cares about re-usability? They'll just commission another outsourced development firm to write the next version from scratch, and still save money (as opposed to paying for real software-engineering/programming work, rather than an offshore sweat-shop coding shop).

But i know that this kind of subject will lead nowhere but to a flame war. So i won't try to convince you (nor even go on with this subject). Learn a little bit about software design or keep your hype position, it's up to you and i don't mind...

Flamewar? Talking about Win16 and code quality in the same paragraph? Nah... :) Now outsourcing, when discussing between programmers, that's flame-bait.

Share this post


Link to post
Share on other sites

One thing about M$ - whatever compromises they have in Longhorn probably don't stem from ANY of the following:

1) A hesitancy to spend money to get the best programmers they can hire

2) A hesitancy to spend money to build the best infrastructure to develop within

3) A hesitancy to spend money to hire the best architects they can find

When all is said and done, it is hard to beat M$ as a well-funded development shop. Historically, their real weaknesses have been around their need to maintain compatibility with a huge installed base of commercial and custom applications, many of which have been coded with a blatant (and sometimes intentional!) disregard for Microsoft's standards and public APIs. In essence, they are hamstrung by thier own success in the past.

I think that if M$ thought they could get more features and functionality into Longhorn by 2006 if they simply doubled the development budget they would do so. NOT doing so makes them smart project managers. It's always important to remember "9 women don't make a baby in a month." Team size and number of interconnects is an important project success and project management parameter - and I think M$ tends to like very flat team structures. Which probably limits them to development teams of 20-40 developers, plus a separate corps of testers and QA folks.

This contrasts to my old days at Accenture (then Andersen Consulting). We had hierarchical project structures, very rigid. We had teams of several HUNDRED developers on many projects - successful projects in many (but not all!) cases. Of course, these were by no means creative software development projects - so input from the developers upwards could be minimal...YECH!

I'm glad M$ is holding back on the features - it will make Longhorn more reliable, and some of us have to use it and support it. It will ALSO give Linux, Gnome and KDE a chance to better catch-up and make in-roads into the desktop space... :D

Future Shock

Share this post


Link to post
Share on other sites

Another determinant as to whether code is reused is the size of the codebase. Much of the code in PeopleSoft and SAP is legacy and is crap by today's standards (some of it was crap when it was developed). But they are mammoth packages, and the cost to redesign and rewrite all of the components is so huge that they just add layers around the old code, and rewrite essential parts on version upgrades.

Share this post


Link to post
Share on other sites
In this case I do not understand this argument. I think it is exactly about quality--if Microsoft were to release the product with their original ambitions intact, I doubt they would have the time to effectively test it. Would you rather have an incredibly flaky OS with loads of incredibly flaky features, or a mildly flaky OS with a number of enhancements that actually work?

I'm just curious about what the enhancements are; all I've really heard is what's not going to be there and that's the three major improvements.

What you've heard isn't accurate. WinFS will likely be available in beta form for Longhorn's release, making it the only 1 of the 3 major updates to not be completely ready. Indigo should be available well before Longhorn is available, and there appears to be a pretty decent chance that Avalon in some form will make it out before Longhorn as well. Both will certainly be fully implemented in Longhorn when released.

Share this post


Link to post
Share on other sites
The vast majority of Win16 code in Win9x was re-used, as-is, precisely because it was hand-written ASM in most cases, and totally unmaintainable

That's a real (and interresting) point...

if you are speaking of the overall lifetime costs of the code, including future maintainability, rather than just the initial costs of production, then yes, I would say that you are correct, sorry

Yes and now. The point where quality invested begins to be a cost saving is getting closer and closer to the beginning as the project gets more and more complex.

Yes, but if the company wants the code now, and wants it at a low cost, or the lowest cost possible (outsourcing), then that code produced will be "good enough", even if it is crap code on a more global scale. Who cares about re-usability?

The first major impact on costs from quality is in debugging (and the more complex the project gets, the more it impacts it).

That isn't necessarily the case. Look at the success of Symantec's Norton Utilities, many Microsoft products, Easy CD-Creator, Intervideo's WinDVD

Well, i was just happy with all thoses pieces when i was using them. You have to define how they suck, because they all met their specifications to me...

I know about Norton, it seems it is hype to say norton sucks the same way that it is hype to say microsoft sucks...

might know very little as you stated, but I believe Windows NT was a completely different platform from Windows 98 or ME and the NT Platform has been the one built upon. If I'm wrong I would like to learn

You are misreading me. A lots of code were reused from earlier releases of windows in win95, then in win98 and Me. There was also a lot of code reuse from the earlier releases of Win NT in Win 2000 and XP. I was not talking about Win16 code reused in Win32.

It's always important to remember "9 women don't make a baby in a month."

I like this :)

Another determinant as to whether code is reused is the size of the codebase

This is a good point...

Much of the code in PeopleSoft and SAP is legacy and is crap by today's standards

Actually, i don't know anything about their code, but what you are saying is very likely to be true, as software engineering have changed, and grown a lot those last years (the OO paradigm is a major step in this regard, even if it is far from being the only one).

All in all, what is getting me sick from times to times is the unrational critics against some companies, just because it is hype to do so, like the "Pentium IV is crap" and "Microsoft's software sucks" clichés.

Share this post


Link to post
Share on other sites
It's always important to remember "9 women don't make a baby in a month."

I can tell you right now though that nine women in one month is one helluva fun month!!! :P

Share this post


Link to post
Share on other sites
I can tell you right now though that nine women in one month is one helluva fun month!!! :P

Not really... Scheduling nightmare.

Share this post


Link to post
Share on other sites
QUOTE

That isn't necessarily the case. Look at the success of Symantec's Norton Utilities, many Microsoft products, Easy CD-Creator, Intervideo's WinDVD

Well, i was just happy with all thoses pieces when i was using them. You have to define how they suck, because they all met their specifications to me...

I haven't used anything Norton, but I've used Easy CD Creator and WinDVD and I quickly switched to Nero Burning Rom and PowerDVD, the I was happy...

Share this post


Link to post
Share on other sites
I can tell you right now though that nine women in one month is one helluva fun month!!! :P

Not really... Scheduling nightmare.

Trying to love nine women is like a balling chain.

-- Rick

P.S. Or is that "ball and chain?" I can never remember.

Share this post


Link to post
Share on other sites
The vast majority of Win16 code in Win9x was re-used, as-is, precisely because it was hand-written ASM in most cases, and totally unmaintainable. The workaround that they instituted, in attempting to wrap a psuedo-pre-emtive 32-bit multitasking OS kernel around what was still really a 16-bit protected-mode ASM code core, was to institute an outer global mutex to prevent re-entrancy problems in that ball of single-threaded code, called the "Win16Mutex". I'm sure that you've heard of it. Code re-use is no evidence of quality, nor maintainability. Only the ability of said code to be hacked at enough to be re-used. It was also the biggest single point of weakness in the Win9x OSes.

Given fixed development resources, though, wrapping the old code may have been far more reliable than what would have resulted from starting from scratch.

-- Rick

Share this post


Link to post
Share on other sites

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now