Originally posted by: Lord Evermore
Dump support for 32 bit in 4 years? They started migrating away from 16 bit 11 years ago. We're still supporting 16 bit apps 6 years after the last version of Windows that was even partially based on 16 bit.
Originally posted by: Nothinman
But why do they care? They get their licensing fees either way.
It's not about just about the license fee and yes, of course backward compatibility is a major issue (btw how much 'official' support there is for 16-bit apps starting from win2k when microsoft started saying bye bye to 9x versions?), but guys, I'm talking big picture here
🙂.
Think about the reasons that made microsoft dumped the 9x line, they did it so they can focus more on a single line (or a smaller line) of general use desktop OS and to have better bang for every man-hour in creating / supporting & patching / adding value as in support new technology in existing OSs and by new technology I mean both hardware and software (i.e. like XP SP1 added support for USB.2 and better wireless encryption or is that from SP2?).
Imho microsoft may be big but vista seems to be bigger.
My personal theory is that with every new generation of OS microsoft will have more difficulties for three reasons:
A. In every new generation of OS, microsoft tries to add as much features(*) as they possibly can (hell XP is almost a classical example of
creeping featurism, next to win2k, and so is ME compared to 98SE, and I have no doubt that vista will also be like that).
B. Windows as a whole is not modulated
enough.
C. They
have to offer extensive backward compatibility.
When you consider that a lot of these added features(*) are new APIs or at the API level (API like, if you wish) like WMI (
wiki) and DCOM (
wiki). This, combined with the other two issues will (imho) make any subsequent OS development harder in an almost exponential way.
If you look at the pattern of timing that microsoft announced about both dropped features and pushing the launching date back, it easy to see that they are reaching the point where they
have to make some cost-effective decisions ("
adding/debugging this feature is starting to take too much time, is this feature important enough for us to push vista back again?"), and they are hitting that point a LOT.
There are ways out of this (my education forces me to either offer a solution whenever I have some criticism or to STFU, so bare with me). Respectively (from the previous listed):
A. They (microsoft) can stop adding so much goddamn features into every OS they make, but that isn?t going to happen. This is a well known established strategy for microsoft, to offer as much software as possible, to have a foothold in any software market under the sun, to eventfully reach a point where windows and Office will together cover 99% of the needs of 99% of users. Now, because microsoft is so big they can actually go for that kind of strategy and still offer some decent software (sometime when i see in microsoft documentation the quote "
to offer our partners" it is replaced in my mind automatically with "
to offer to our soon to be new division").
However, vista (and its successor I'm sure, to an even greater extend) is reaching to a point where even microsoft is starting to have lots of difficulties, and is starting to make a lot of compromises.
B. I'm pretty sure that vista is more modulated then 2k/XP but considering how much time and other resources it take to fully modulate a behemoth piece of software like xp/vista, I wouldn't be surprised if it turned out that vista isn?t far off from XP in the modulation department, and I'm not talking about just better separation between the kernel and the user land, but a better modulated API structure, where the cross dependencies are not complicated enough to give Albert Einstein a migraine (is there a way to write a networking program to XP and lunching it without triggering more then a third of the background services and reading from less then a several dozens of windows DLLs), I wander how bad vista will be in that department, and I'm not optimistic.
C. Eventually something has to give. Like when microsoft said "enough is enough" and started to pull away from the 9x family, or like in XP SP2 when it broke some apps (although some will argue that some of the changes in XP2 should have been in XP from the start). A relatively good example of this is DirectX where microsoft has reached the point where you can't add an additional floor to the existing building because they have reach the limitation of the foundation, and the only way to build a higher building is to tear it down so they can build a better foundation, its only a
relatively good example because the a LOT has changes in computer games and multimedia since the first DirectX 1.
Now, if microsoft can push a critical mass of users to 64-bit, it can in 4-5 years time announce a shorter Life-Cycle Policy to the 32-bit vista (faster retirement), and for their next OS to offer less backward compatibility (like NTs kernels support to 16-bit, it's there, but there isn?t a lot of it), giving them the ability to stop support for a TON of old code in their next generation OS (the one after vista), scheduled for somewhere like 2013.
However if micosoft can't pull that off, they will need to offer support for legacy software that was wrote to an 32-bit APIs microsoft created for more then 15 years ago, and there is a TON more 32-bit software then 16-bit, and supporting
all the older 32-bit APIs will be programming
hell (there are a lot more 32-bit APIs, and even if there aren?t a lot of them, they are much more difficult to support then DOS), if you think vista took a lot of time to make, wait for its successor! Something like that will take more then 3-4 years to make, think more like 6-7 years if not more, this is why i said it would be
exponential more difficult.
It's in microsoft best interest to leave 32-bit behind as soon as possible, and they will try to '
encourage' people to move to 64-bit, using well known tactics, even if it isn?t true in this particular case.
edit
big post = lots of errors
add links