Slashdot Mirror
How Microsoft Plans To Get Its Groove Back With Win7
shawnz tips a blog post up at thebetaguy that details Windows 7's huge departure from the past, and the bold strategy Microsoft will be employing to maintain backward compatibility. Hint: Apple did it seven years back. There are interesting anti-trust implications too. "Windows 7 takes a different approach to the componentization and backwards compatibility issues; in short, it doesn't think about them at all. Windows 7 will be a from-the-ground-up packaging of the Windows codebase; partially source, but not binary compatible with previous versions of Windows."
Over ambitious as always. I say work on improving XP . Make it more efficient and add features. Perhaps get all those other features that were promised 10 years ago working. Like WinFS. Like a dozen other things. MS is just digging itself deeper.
Skiffy is Spiffy, but Ort is tort.
...releases lost the game long ago. It is useless to think in an OS as a package, much less something you put in a box. Given that the OS is the first software building block of a system and due to the sheer complexity of the thing, it has evolved into a continually updated and polished piece of engineering, where you take snapshots of the development and call them releases.
An operating system evolves and you don't sell it. You either provide it as a service, or provide it for free, so that you can hook people on some service you offer.
I'll tell you why Win 7 will be a huge flop: since it breaks almost all compatibility between itself and previous windows releases, it has to compete on the same grounds as Linux, *BSD and OSX. Which means, that without the massive inertia of the previous windows releases, those three will kick the living crap out of Win 7 in terms of maturity, usability and price.
It takes a man to suffer ignorance and smile
Be yourself no matter what they say
Transparent emulators (should they even be called that?) are very fast - ever run a VM? They just pass through code into the native processor and make sure functions get routed to the appropriate library. Not quite as fast as running natively, but if you are able to significantly increase your "native" speed, the tradeoff is usually worth it (at most it's about a 20% hit - real world is usually much less).
Where you DO run into problems is with I/O, meaning we get the driver headache again. I believe that is one reason Vista pushed a new driver model - an attempt at future-proofing for this new OS model.
The plus side of a VM is you get a layer of stability for free if you do it right (I don't count on MS to do anything right, especially the first time...) - crashing the VM doesn't necessarily crash the native OS (depends on what caused the crash - bad memory crashes everything).
From the article: On traditional hard drives, the more separate files which the operating system has to load, the more seeking across the hard drive is required, and therefore overall performance takes a hit. ... In Windows 7, Microsoft will break from the Windows' norm by breaking previous API compatibility, offering new API frameworks as a native solution, and providing support for legacy frameworks (COM, ATL, .NET Framework, etc) through monolithic libraries designed to provide the functionality of all previous revisions of the modules in question.
And so, the answer is to put everything in one bloated DLL?
It apparently hasn't yet penetrated to the Windows 7 group that computers aren't going to get much more powerful for years to come. That stopped once laptops started outselling desktops. In laptops, what matters is size, weight, and battery life. The future is the OLPC and the Asus Eee. In a few years, laptops in bubble-packs for $89.95 will be hanging on racks at the drugstore. Microsoft isn't ready for that.
Progress now will come from reducing software bloat. Microsoft has, in desperation, extended the life of Windows XP for little machines. That's only a stopgap measure. Now they need to de-bloat their whole product line and get their costs down.
If Microsoft was at all smart, they would use a light weight "Windows on Windows" strategy similar to how they implemented 16 bit Windows on the NT base on a new VERY stripped-down 64 bit Windows kernel and use virtualization of every Windows application.
In this day and age, it makes no sense to me to write another massive OS.
Looks like things are playing out as Joel predicted. It should be interesting to watch.
Lacking <sarcasm> tags,
Maury
I think that the big problem is that Microsoft is calling this new operating system "Windows". If they were to break with the past, and continue offering and supporting XP for the installed base, they would find a lot of benefits.
First, a non-Windows operating system would probably free them from the anti-trust agreements. After all, the old Windows line, that was the monopoly -- this new OS is competing with Windows.
Second, freeing themselves from the name allows them to experiment with new changes to the OS experience, which in turn would allow them to make much better use of their in-house R&D and their UI experience from their gaming division.
Third, it puts them in the position of underdog again, a position in which Microsoft historically thrives. They're a competitive bunch, and they just write better code in a competitive environment. With Vista, there was no real pressure to get it right, because they assumed that everyone would just upgrade from XP. If they're competing against XP, however, that frames the development process quite differently.
In a way, it's kind of a cheap trick, but I think that it would be very good for Microsoft to break out of this rut and break away from Windows. If they make a product, and compete fairly to get people to use it, they have the cash, talent, and reputation to pull off something good.
My Slashdot-fu fails me but I seem to recall, circa 2002, an article almost exactly like this, but the speculation was on "Longhorn" (i.e., Vista). The predictions, the most notable of which was that Longhorn would completely break all compatibility with everything that came prior, was pretty much identical. Then, as now, it seemed like the single stupidest idea ever. And then, as now, it was in an article using no sources on what was essentially a blog. And then, as now, the Slashdot submitter posted it as if it was the Gospel and the first several submitters carried it as if it was going to be the death blow to Microsoft they needed. Then someone (like yourself) clued in that this is just something that some blogger pulled out of his ass.
And the best part about the circa-2002 article was that either in that post or on another post on the site the author railed on about how you can be a 40-something programmer and lose out on a job to a 28-year-old programmer because the 28-year-old has "social skills" and you don't and don't want to because if you wanted to have "social skills" you would not have become a programmer in the first place. His "about" page revealed that he was a 40-something programmer, complete with a laughably awful photo of him, morbidly obese, sitting in front of his PC.
So essentially this was a bitter old man making a bunch of shit up. I'd almost guess that this "betaguy" is the same person with some better web design skills.
Schnapple
And...I'm all for them trying something new. Start over! Look at apple. They've started over a few times, and I think it's been worth it...there's just not as much community pain felt because the install base is relatively small.
If you want a stable, mostly command line, system that'll be backwards compatible for decades to come, use your flavor of *nix...but if you want a fancy graphical interface with pretties (targeted at an audience who enjoys them)...you're gunna have to deal with sdk's and API's...that's just smart/efficient programming...where have you seen anything else?
In my opinion, it's marketing that screws the tech of MS. They come out with stupid as claims before knowing what the final product will be, over hype everything, and seem to get their hands in determining code paths. Their sdk's and api's (directshow for instance) and are mostly pretty neat. Marketing makes it so abstract and burried in coined tech terms that somehow make their way into the msdn (I consider this in the marketing goup...cause an intelligent software engineer would never make something like msdn) that it takes all the fun, desire, and some ability to learn it (at least for me)!
I agree, they are admitting defeat...but that comes with a realization that the customers (us) obviously want something better (sales of vista), but are limited with the current platform/code organization/model that they use now. Sounds like innovation/renovation to me...and that should be something constant in any field.
I am seriously curious twitter, you spend a lot of time on Slashdot, you talk incessantly about honesty - when did you decide you were going to turn it into a mockery and a circus by organizing these "bad zealot-good zealot" clusterfucks where you use the troll accounts everyone knows about (twitter and Erris) to give your other sockpuppets an opening to blabber their way on to karma heaven?
The problem here is not what you're saying on this particular post for example, which I suppose might be considered halfway insighftul without the "fuck shit rape fuck M$ Winblozes LOLOL" tone of your earlier accounts. The problem is your blatant gaming of the comment and moderations systems. You call Slashdot a community and you spend a lot of time talking about "us" and "we", but you sure seem to spend a lot of time lying (and therefore ridiculing) to everyone as well.
How long do you figure this can last?
The twitter monologues. Click on my homepage and be amazed.
Well, maybe there's your answer - MS wants to drop backwards compatibility specifically BECAUSE of things like ReactOS.
If Apps manufacturers are forced to follow suit, all new apps will have no (or poor) XP compatibility and thus will not run on the likes of ReactOS - in other words, end-users MUST use Win7 in order to run the latest apps.
AT&ROFLMAO
No, just no. It's a mystery to me why MS hasn't done this sooner. There's a lot to be gained for end users by throwing out the old code and starting from scratch with a set up which is designed for modern processors.
It's hardly a credit to MS that they've stuck with what is a bog of broken code and APIs for this long. ReactOS and wine just aren't large enough competitors to warrant this sort of radical "fix."
One can throw around a lot of paranoid speculation, but the reality is that a lot of the flakiness of Windows has been a byproduct of having all that stale code and 3rd party software interaction. Doing a redesign now with VM processor extensions and an awareness that right now things are moving to a multi-core 64bit environment makes this a good thing. Many of the design decisions would have been handled differently had the engineers known where things were going even 3 or 4 years down the road.
In terms of threat, the biggest threat here is that win 7 will not only not suck, but will do a genuinely amazing job at providing the end users and support staff with what they really want.
Apple didn't introduce a new OS that was only source-code compatible with existing applications. Apple introduced a new API that was very similar to the old API, restricted in some areas, expanded in others, designed to run efficiently on both the old and new operating systems. They did this before the new OS was released. Then, in the new OS, old applications that were not written to the new API ran in an emulator, and old applications that were written entirely to the new API ran native on the new OS.
.NET this way: that .NET code would run on some future Windows platform, but Win32 code would only run in an emulator.
.NET, or Microsoft is doing something completely different from Apple... and what Apple did was risky enough to start with.
At the same time they introduced two more APIs, one that was an enhanced version of the old compatible one that took advantage of the new OS, and one that was new to the new OS. They also introduced a new development environment that generated code for the new APIs.
When they introduced the Intel-based Mac, they abandoned the oldest API, provided an emulator for existing code, and code written in the enhanced API using Apple's development tools could be recompiled in a mode that supported both Power PC and Intel processors.
At no point was there a stage that broke code written within the previous two generations of APIs.
I was under the impression that Microsoft was planning on using
Either the article is wrong about Microsoft abandoning