Gossip: Windows 7 To Break With Backward Compatibility

[Mike]

Previous versions of Windows provided backwards compatibility with older versions of windows, making it easy for users to upgrades as their favorite programs would still work.

However, with Windows 7 Microsoft are approaching backwards compatibility differently, and Windows 7 will not be binary compatible with older versions. This won’t mean that older programs won’t run, but that virtualisation will be used to make them run. The key drivers for this are performance improvements, but also in an attempt to sidestep regulatory programs.

Vista struggles with performance because Microsoft choose to incorporate regulatory challenged features like Internet Explorer and Windows Media Player using a modular design, rather than the monolithic processes used in previous versions of windows. The greater use of componentisation led to performance issues due to the increased number of libraries within the operating system. Simply put, the more separate files your hard drive has to load, the slower your system is going to run.

Another reason why Vista runs so slowly is that the operating system stores multiple copies of core system libraries, as each revision of a library typically adds/removes functions, and applications compiled with dynamic links to a specific version of a DLL file may call on functions not present in the currently installed library. Vista aims to solve this issue through the WinSxS collection; essentially a massive store of every differing version of libraries present on the system. That way, when an application makes a call for a dynamically linked library, Vista queries the WinSxS cache for the correct version, which is then loaded into memory. On the average system, this directory can be several gigabytes in size, with much of the code duplicated between the separate versions many times.

By not being binary compatible with previous versions of Windows and being built from the ground up, will give Windows 7 significant performance improvements. This will make it easier for developers to produce faster apps as they won’t have to worry about backwards compatibility, but it will increase the upgrade pain for users who may find that some of their existing applications will no longer work and will need upgrading.

Where do you lie on the trade-off between compatibility and speed? Do you think Microsoft are making a mistake here?

[Wraak]

I honestly think MS has to start from scratch with W7. Do not care about backward or forward compability. It has to be fast and reliable.

[DDH]

IMHO its basically down to the 18 round slug out between marketting and technical. If you want the best technical solution start afresh and don't live with the sins of the past. If you want to sell lots of it now be sure you have an easy upgrade path and plenty of things to run on it.

Given the great advances in technology I think for the long term good of the product you need to start over and not worry about compatability. Vista is pretty mature, it and XP can run commercials and MCE now, let it be and get a new good start I say.

[arkay]

and yet OSX was able to completely change hardware platforms. i.e. Not even remotely compatible chipsets, and still maintain backward compatibility. Might not have been fast but it worked.

At least this seems to be the first time in a long time that MS actually are thinking about what needs to be done to continue to remain a part of IT's future. It may be too little too late.

Unfortunately for them if they go this way I think they'll only continue to lose customers.

Vista was the "re-write" that everyone was prepared to accept. The end result was a lot of cost on hardware if you wanted it for a less than stellar result.

If Windows 7 is the "other re-write", who's cost will be software based, due to lack of backward compatibility.. Well. You can understand if people just aren't prepared to wear the brunt and costs of MS's mistakes in not doing with Vista what they now appear to be willing to do with seven.

Only time will see where this leads.. I also have my reservations about the technical ability of the devs/project managers/anyone within MS to develop significantly differently to what they are accustomed. It's difficult to not just repeat the same mistakes when those methods and ideologies and ingrained.

Of course they burned a lot of the hardware driver coders with Vista, look how long it's taken to see a stable video driver.

What will the software market do when they all have to re-compile and re-work their code for this new release?

Cheers,

Arkay.

[Ararat]

Anyone think we'll end up with the same thing as we did last time we had this sort of rewrite? The situation where we had the NT OSes and the 9x OSes alongside each other for half a decade. We could well have Windows 7 with this new rewrite, followed shortly by a redressed Vista (or a Windows 6.1) for those who want the new features (assuming we get some with Windows 7), but don't want the incompatibility. It could run like that again for several Windows versions before we end up in todays situation again - one main OS (like XP and Vista), and people wanting a rewrite of that 'old' code.

[ozbear]

I think the key things are:
1) Whether this virtualisation gizmo ensures that all my old programs work and
2) what kind of performace hit running them in that mode incurs

Most of Joe Sixpack's applications are just not CPU hungry...word processing, email, web surfing. Games are a different story but those kinds of apps I would think would be of the "disposable" kind (i.e., they aren't huge investments, after you've completed it, it becomes shelfware) and the games industry seems pretty adaptable to generating games for multiple platforms/operating systems, so just add Windows 7 to the list.

I don't particularly care if my MS Outlook 2003 email program runs more slowly since it isn't a performance application. Ditto with the majority of my applications. However, if virtualisation is going to require significantly more time to load the applications and if I am going to get pauses between keyboard presses then that is going to be an issue, for me.

Oz

[ExtremePC]

Using past precedence when MS say "break backwards compatability" are we to expect a host of new propriatary file formats, multimedia etc? If they are they can keep it .... no thanks.

To me it sounds like Windows 7 is either going to be the last straw that pushes a large percentage of the market to an alternative or simply retains market share. I strongly believe MS has had its day in the sun and will now be in a declining position. The unfortunate thing about that for the short term is that there is no viable alternative able to fill the same spot. linux .... no its too far to the "left" , mac osx ... maybe but ...no its too far to the "right", Windows ..... is having an identity crisis.

[arkay]

Originally Posted by ExtremePC

linux .... no its too far to the "left" , mac osx ... maybe but ...no its too far to the "right"

In their current forms maybe but something will emerge from it all as the next user OS. MS won't disappear overnight (though I think it will happen fairly swiftly if it does).

There are plenty of options out there in terms of OS. The only thing any of them lack in terms of being the next OS is vendor support. If the world suddenly decided to redevelop all their software and drivers for syllable then it'd be a top OS. Likewise if it happened for Linux. OSX though wreaks of the same commercial lock-in of Windows and I don't think any of us ever want to be locked in to proprietary formats again. Ever.

It's funny in a way but their insistence (both Mac and Windows), on DRM and protecting their lock-in has very quickly brought about a backlash that will force them into openness.. It's a great time to be watching the goings on.

Just look at the bungle that was ooxml, MS running about bribing people making a complete sham of the iso certification process. Now MS, after all their tactics have failed, are implementing ODF in Office 2007 SP1. For the first time in a very long time the commercial entity is having to go with an open standard document format that is known and can be implemented by anyone. That's a huge step given Office has been MS's mainstay over the years. Lets see how Office competes on a level playing field after that.

It's really down to us, the end users, to say no to proprietary formats of any kind. If we (the collective we, including business and home users), won't use it, they can't sell it. It's that simple.

Personally I think new laws need to be drawn up specifying that the format of any datatype that defines the way in which audio/video or textual media is stored must be published to the public domain and cannot be patented. How something is stored shouldn't be the marketable component. What you are able to do with it should.

This would put a stop to the "thousand format wars" and we'd end up with very few, well designed formats to work with.

Cheers,

Arkay.

[ExtremePC]

Originally Posted by arkay

In their current forms maybe but something will emerge from it all as the next user OS. MS won't disappear overnight (though I think it will happen fairly swiftly if it does).

There are plenty of options out there in terms of OS. The only thing any of them lack in terms of being the next OS is vendor support. If the world suddenly decided to redevelop all their software and drivers for syllable then it'd be a top OS. Likewise if it happened for Linux. OSX though wreaks of the same commercial lock-in of Windows and I don't think any of us ever want to be locked in to proprietary formats again. Ever.

It's funny in a way but their insistence (both Mac and Windows), on DRM and protecting their lock-in has very quickly brought about a backlash that will force them into openness.. It's a great time to be watching the goings on.

Just look at the bungle that was ooxml, MS running about bribing people making a complete sham of the iso certification process. Now MS, after all their tactics have failed, are implementing ODF in Office 2007 SP1. For the first time in a very long time the commercial entity is having to go with an open standard document format that is known and can be implemented by anyone. That's a huge step given Office has been MS's mainstay over the years. Lets see how Office competes on a level playing field after that.

It's really down to us, the end users, to say no to proprietary formats of any kind. If we (the collective we, including business and home users), won't use it, they can't sell it. It's that simple.

Personally I think new laws need to be drawn up specifying that the format of any datatype that defines the way in which audio/video or textual media is stored must be published to the public domain and cannot be patented. How something is stored shouldn't be the marketable component. What you are able to do with it should.

This would put a stop to the "thousand format wars" and we'd end up with very few, well designed formats to work with.

Cheers,

Arkay.

Oh yeah, agreed MS's structure is such that it NEEDS to sell big in order to exists, The dollar value of MS is predominantly in shareprice or in other words percieved value. If that were to ever collapse MS cannot sustain itself the implosion would be phenominal.

As for how soon we would NEED an alternative FULLY able to fill Windows shoes after such an implossion, I'd say 1-2 months MAX.

Please lets not get the government involved in already murky waters. If anything how about removing the ability to patent or copyright a file type, file system etc. works better I think.

for any other OS to take marketshare away from MS in a big way its not enough to offer more of the same but with an open standard because as far as the majority of people are concerned we already have that ... the MS standard, no, what is needed is for an OS to be soooooo HUGELY and obviously better including the "spit and polish" with an almost 0 impact transition path to workflow PLUS all its toolsets and developer tools also need to be refined, and fully GUI, PLUS a complete back catalog of applications that are as equally refined and complete including the "spit and polish" AND be data compatible with the apps they're replacing, AND it will need to have some "killer" games that are BETTER than the current crop of DX10 based games at the SAME time as the OS release. ........... of course what we have is linux which is ... alway almost maybe, could be ready for the mass market if it wasn't for ALL of the other things mentioned above.

To replace MS/Windows we are talking about 2 - 4 years of development work across the board of all facets of linux.

As for hardware support from manufacturers for linux, here's a novel idea .... PAY them!

[ExtremePC]

How about starting a thread on what linux has to evolve to so it can be an actual contender for a mainstram OS?

[Ararat]

Originally Posted by ExtremePC

Oh yeah, agreed MS's structure is such that it NEEDS to sell big in order to exists, The dollar value of MS is predominantly in shareprice or in other words percieved value. If that were to ever collapse MS cannot sustain itself the implosion would be phenominal.

As for how soon we would NEED an alternative FULLY able to fill Windows shoes after such an implossion, I'd say 1-2 months MAX.

Please lets not get the government involved in already murky waters. If anything how about removing the ability to patent or copyright a file type, file system etc. works better I think.

for any other OS to take marketshare away from MS in a big way its not enough to offer more of the same but with an open standard because as far as the majority of people are concerned we already have that ... the MS standard, no, what is needed is for an OS to be soooooo HUGELY and obviously better including the "spit and polish" with an almost 0 impact transition path to workflow PLUS all its toolsets and developer tools also need to be refined, and fully GUI, PLUS a complete back catalog of applications that are as equally refined and complete including the "spit and polish" AND be data compatible with the apps they're replacing, AND it will need to have some "killer" games that are BETTER than the current crop of DX10 based games at the SAME time as the OS release. ........... of course what we have is linux which is ... alway almost maybe, could be ready for the mass market if it wasn't for ALL of the other things mentioned above.

To replace MS/Windows we are talking about 2 - 4 years of development work across the board of all facets of linux.

As for hardware support from manufacturers for linux, here's a novel idea .... PAY them!

Pay them? We DID! We bought their hardware! Anything other than that, and it becomes very difficult, as afaik, noone actually OWNS linux, so who pays them? God? Santa? Not me I hope.