Engineering 7: A view from the bottom

re: Engineering 7: A view from the bottom

Asesh — Thu, 16 Oct 2008 05:31:15 GMT

Yes guys, make Windows 7 the best OS out there. Let it be a Mac OS X killer!! :)

re: Engineering 7: A view from the bottom

Asesh — Thu, 16 Oct 2008 05:34:02 GMT

I have heard that Windows 7 will only be available to some beta testers. Why? It should be open to all the users who would want to test it. That would help Microsoft to fix more bugs and security holes and result in an initial release of a more stable operating system. As usual the initial release will be full of annoying bugs and many won't just upgrade to Windows 7 until it's SP1 is released. SO THE BETA VERSION OF WINDOWS 7 SHOULD BE AVAILABLE TO ALL THE USERS!!

re: Engineering 7: A view from the bottom

Vistaline — Thu, 16 Oct 2008 06:43:37 GMT

>>"cutting is shipping"

THEY'VE TAKEN THE SCISSORS TO W7 ALREADY!!!!!111!1

All kidding aside, it looks like management of the operating system has gotten much more efficient this time around. Windows, prior to W7, sounds like a mess from what I gather in that post! It's good to hear things have gotten better... and that you're having so much fun, Mr. Osterman. :)

Thanks

TimOR — Thu, 16 Oct 2008 06:48:09 GMT

Thanks Larry - a great post and pick-me-up for a Windows user. As born-again XP user, I was really depressed, thinking that my experience with Vista signalled the beginning of the end of my use of Windows, that XP was the best I'd ever see from MS. You've helped me understand some of the reasons why Vista was so poor out of the gate. From what you write it seems MS will ship Win 7 thoroughly tested and more robust than Vista.

My only plea AS AN XP USER now to MS management and programmers and all is that you make the move from XP to Windows 7 as pain-free as possible, or even (dare I say) a joy!

Thanks again

Tim

re: Engineering 7: A view from the bottom

magicalclick — Thu, 16 Oct 2008 09:45:33 GMT

Good to know the process if more transparent internally. Even though the difficulty to finish WPF early is a big problem, the development process is also needed to be changed.

I wonder how Win7 will be like. Could it be filled with tons of WPF based goody apps?

re: Engineering 7: A view from the bottom

lyesmith — Thu, 16 Oct 2008 09:48:16 GMT

With earlier windows I got the impression that at MS the right hand does not know what the left does. And the head does not know anything. There were so many inconsistency and duplicates. Lets hope that this time "transparency" made a difference.

Do you know what will make Windows Seven 7.0 and not 6.2?

re: Engineering 7: A view from the bottom

Domenico — Thu, 16 Oct 2008 11:11:00 GMT

Nice post Mr. Larry

Many many Thank's!!

re: Engineering 7: A view from the bottom

d_e — Thu, 16 Oct 2008 11:57:38 GMT

Thanks for the very interesting read.

You wrote that input for feature no longer comes from senior management but from the feature crews themselves (=developers, testers). Isn't there a danger that devs will focus on new features and bugfixes and don't invest time on "boring" things like consistency and fit & finish?

http://shellrevealed.com/blogs/shellblog/archive/2006/09/18/The-Fit-_2600_-Finish-Balancing-Act.aspx

But considering all things I have to say: Wow! You changed the whole process completly since the first years of Longhorn. Pretty impressive stuff! I wished I could work on Windows 7 too.

re: Engineering 7: A view from the bottom

Tihiy — Thu, 16 Oct 2008 14:00:00 GMT

Very interesting post indeed...

But reading and realizing it makes my hair rise.

I dunno, you refer to developers like they are robots or monkeys. Maybe i feel that because you don't describe any particular change specifically, or this "6 weeks of development and 6 weeks of integration" really frightens me out.

Who's the creator there? Where is the generator of the ideas? How deep feature can be cut if there's time out? Where's the GUIDELINES?!

What, you don't have time and leave only a dummy UI window in some temporary library just because you can't fulfill specifications and test everything? That smells like Vista, full of unfinished taste.

You have to fall back to previous milestone and 'polish' it for 6 weeks? Or team is broken apart?

That bad taste is especially because you don't talk specific... sorry.

re: Engineering 7: A view from the bottom

adondai — Thu, 16 Oct 2008 14:23:11 GMT

That was very interesting... great to hear in more detail the development process...

You said it's been a blast building Windows 7 and I imagine it would be, especially compared to the complications around Longhorn/Vista whatever...

However surely it's also been a little stressful as well? Considering the fuss over Vista (justified initially, totally unjustified now) there must be a fair amount of pressure to improve Windows image and remove the stigma Apple has given Windows etc?

re: Engineering 7: A view from the bottom

dreimanis — Thu, 16 Oct 2008 16:22:48 GMT

As I said earlier. If Windows 7 is gonna be at least as half as good as this Engineering blog, it will be the best OS ever!

Keep it up, guys!

re: Engineering 7: A view from the bottom

lyesmith — Thu, 16 Oct 2008 17:11:26 GMT

"ASUS readying touchscreen Eee PC and laptops for 2009 Windows 7 launch?"

http://www.engadget.com/2008/10/16/asus-readying-touchscreen-eee-pc-and-laptop-for-2009-windows-7-l/

Does this mean that Win7 will have much lower system requirements than Vista?

What can you tell us about system requirements?

re: Engineering 7: A view from the bottom

Pendragon — Thu, 16 Oct 2008 17:23:04 GMT

I'm a user not a programmer, as such I have no right to criticize the process.

Having read you post I'm pleased that a shift I'm the management process puts the emphasis on good code. We can all dream up the killer feature but if you can get it to work then what is the point. If management were to focus on the unattainable feature at the expense of good code. All it achieves is delays and poor customer experience.

From a users point of view Windows, historically always been buggie. With most geek guru's saying wait for SP1 whenever talking about new versions of Windows. If this new way of working reduces that then great. If it keeps the people working on the project happy then that's also good. We hear some odd things from Redmond about what it is like to work for MS. Having happy workers also means better code...

re: Engineering 7: A view from the bottom

Kosher — Thu, 16 Oct 2008 19:02:28 GMT

The feature cutting process is always iffy. Some guy usually "has it in" for some other technology, feature, or idea because he thought of it first or someone else gets to work on it besides him. If that jealous person is in a position of power, they will often axe at will.

Perhaps this has become easier in Windows 7? Hopefully the backing from management is on the developer's and users sides. They both want the best thing in the end. It's the management that sometimes only thinks about the green and ship.

The end result should be what was originally envisioned IMO and that means getting some floating teams involved to provide additional push on the fronts that need a boost.

re: Engineering 7: A view from the bottom

Prixsel — Thu, 16 Oct 2008 20:17:56 GMT

Can't wait for the new Win7 release for public :P

Wiser behaviors of the new core , lowered down size of the OS , new program rules for program architectures , better support for old and new drivers, and speed like WinXP when it's at least running on lowered down settings is one of many dreams :(

Here are some sugegstions and problems -->>

Make that programs work and give out popups and settings what actually can't be used by quest or non administrative accounts!

During installation system files should not fragment pagefile and MFT palcements on disk and should be optimized on disk from the very first start for optimal experience! ( Perfectdisk 2008 website says it uses the info what Microsoft gaved for optimal speed with MFT and Pagefile with correct disk placement ).

System restore points should be excluded from defragmenter programs and be placed on the most inner disk part so it would not make confusions or unnecessary movements during defragment.

These are just problems and I don't know what things are possible do make or change :(

Cuting is Shipping

Yert — Thu, 16 Oct 2008 20:22:23 GMT

Just as Messenger, and soon Movie Maker got cut, we will see more and more cuts, but the features will still be out there. Way to go. The biggest feature of Windows 7 will have to be the code shrinkage. Even if the requirements are the same, and tons of new features added, I expect the code won't grow much if at all, if not shrink a little. While that isn't the holy grail, it will still be very interesting.

Also, I have to say that Seven is also a great name, and if I don't see a Bond (i.e. 007) joke before this thing hits SP1, I will have to give up on Microsoft's sense of humor or marketing.

re: Engineering 7: A view from the bottom

Domenico — Thu, 16 Oct 2008 21:12:24 GMT

@lyesmith

Yes I love It UMPC MT+ Windows 7 = Surface consumer :D

but see this

http://www.tendancehightech.com/blog/en/index.php?2007/01/31/54-canova-dual-screen-laptop

Go Micosoft GO!!!

Discoverability

woodycodeblue — Thu, 16 Oct 2008 22:10:35 GMT

The phrase "delivered in a coherent and discoverable way" caught my attention. As said, there are thousands of features implemented in Windows releases. The problem is that a lot of them are not easily discovered.

The Welcome Center in Vista makes an attempt at helping to discover features but falls short, in my mind at least, of being truly useful. The "What's new in Windows Vista" entry talks about Searching And Organizing, for example, but never gets to any meat. It says you can search for stuff and points out the search box, but never talks about the cool points of how to use it. It links to a help article "Demo: Working with files and folders", which doesn't even touch on search.

I wonder how many people know that you can search from the start menu for emails in outlook. How many "power users" wanted to turn off UAC and had trouble finding the switch when they could have simply used the search box in the Control Panel window to search for "uac" and be led right to it? How many know that you can use natural language for searches to narrow down searches, like "email from SomeGuy@hotmail.com" or "Songs by SomeArtist"? For those that do, how many are really familiar with the extent of that usage? I'm certainly not. I just know that it can be done, and I haven't taken the time to find a listing of the possibilities.

This leaves a lot a room for improvement. Of course it's probobly a bit late in the development lifecycle for Win7 feature suggestions, but here's one anyway, maybe for the next release or a service pack or something. A mechanism to help in discovery of prominent features by tracking which have been hit. Privacy concerns abound with any tracking so it'd have to be opt-in or whatever, but I'd certainly make use of it if it did a good job.

Say there's a set of features that you want to make sure that users discover. Have a checklist somewhere of what points of what features are hit by the user. Let the user opt in, and track which features have been touched on, and how well they're being utilized. Monitor certain clues indicative of a place they could work faster or easier.

Take search, for example. If a person is always going into All Programs in the start menu, prompt them in a non-intrusive way and ask if they know they can search for programs in the start menu. If they're exploring a lot of Control Panels but never doing anything, suggest they try searching for what they want. If they search for *.mp3 suggest they use "songs" instead, and offer a link for an in-depth guide to really using the search functionality.

Different levels of pushiness might be desired, ranging from on-demand bubbles to a daily or weekly suggestion dialog to turning the feature off completely. The point is to actually help people discover the features that make the OS great.

There are so many things where it seems you have to know to look for them, like search semantics or keyboard shortcuts for frequently performed tasks (win-char combinations, anyone?), that a feature to help discover other features would really be useful for a lot of people. Power users can turn them off or leave it on for discovery. Grandparents can almost always use the extra tips on the things that make the system easier to use.

Overall, I think a lot more could be done to showcase the things that make each new version of windows better. It hurts me inside, but Apple really does a good job at that. The user manual for the iPhone, from what I hear, is basically a guide for how to demo the machine for your friends. Everybody I knew that got it when it first came out seemed to say "Well, all I know how to do is look at this picture, turn the phone sideways and pinch it to zoom, but doesn't it look neat?"

Apple has a great following for a phone that never impressed me because they did a great job at making a couple things work well and look pretty (and let's face it, displaying images isn't very special by itself), and showing their customers how to show their friends.

I mean, there were too many people saying "man, my older-than-dust printer doesn't have updated drivers" or "Vista doesn't work with my video card". How great would it have been if they were saying something like "Isn't this cool? I can search for any program right here", "Watch this, I can find all my favorite music by typing 'songs with more than 4 stars' and THEN I can save the search and have it right there ready for me", or "Check this out, these popups make sure I know something's messing with the system, so I can stop malware and viruses before they do anything serious." If we taught people how to use all the amazing features that get put into the releases, rather than letting people focus on and get frustrated about things like the frequent UAC prompts, the overall attitude would be so different.

re: Engineering 7: A view from the bottom

soukyuu — Thu, 16 Oct 2008 22:13:36 GMT

The old way of building windows kinda sounds like a huge mess to me... the new one sounds much better.

Maybe i misunderstood a part of the post but did the old way really force some features to be merged into the winmain branch even though they were not ready/tested completely yet? O_O

I hope that win7 will be much better than Vista & XP and will use less space on the HDD compared to Vista. As I understand it, Vista is taking so much space because all the drivers/language support packages etc get copied to the HDD so the user doesn't have to insert the DVD every time a component is installed. Please, please make it an option to NOT do this. I prefer to insert the windows DVD over having all the components i'll possibly never need occupy the space i could use for other things.

re: Engineering 7: A view from the bottom

someone — Thu, 16 Oct 2008 22:22:52 GMT

Hey I thought you'd talk in general about audio instead of this, although this was an interesting read. (Milestones 2 and 3 did ship with under-the-hood features!!!, only thing ppl could do is post UI screenshots). Coming back to audio, I just hope your teams build something new for MIDI and USB interfaces and audio pros in Windows 7. (The pro audio area has been somewhat neglected by Microsoft and although Vista introduced WaveRT, it only supports PCI devices). And something new with HTPC and gaming audio as well. (Make XAudio 2 part of the OS).

re: Engineering 7: A view from the bottom

LifeOnTitan — Thu, 16 Oct 2008 22:39:40 GMT

Great to see engineers participating in this blog(go Larry, go!!) It's just awesome that senior Windows executives are so engaged here (and Sinofsky's writings are excellently written, informative and unusually candid.)

More devs on Engineering Windows please!

Cheers,

Charles

http://channel9.msdn.com

re: Engineering 7: A view from the bottom

Unknown2 — Thu, 16 Oct 2008 23:32:42 GMT

Thankyou for some insight into the process of developing Win7.

The current development process is very similar to the one that I am familiar with in my limited experiences as a developer. I must say that I have found that environment much less stressful and more productive then some of the other models.

What happens though with features that are not ready by the launch? Do they just disappear? Are they released later as an update or as part of a service pack?

re: Engineering 7: A view from the bottom

VistaLover — Thu, 16 Oct 2008 23:33:44 GMT

This is excellent stuff, and I'll bet Windows 7 is a great OS, just like Vista before it.

Thanks for keeping us informed, these posts have really got me excited for what you'll be posting in the future on specific technologies.

MS is truly responsible for the dream of 'a PC in every house' coming true, I hope when Windows 7 is done and the press is falling over themselves to praise it, you guys take a nice, well deserved vacation. :)

re: Engineering 7: A view from the bottom

jrronimo — Fri, 17 Oct 2008 00:02:13 GMT

Another great post, thanks!

It sounds like the building of Windows 7 is much more dynamic than previously, and that sounds like a good thing. I'm really pleased to hear that it could essentially be shipped today and all features would be complete and tested, even if the OS was not as complete as everyone might want.

I love Vista, but there's an excitement around 7 that I just can't escape...

Features

smartpatrol — Fri, 17 Oct 2008 00:32:43 GMT

See a lot of talk about features which is nice but what about core OS? are we getting vista with different and more features? Hopefully this time they are actually useful!

re: Engineering 7: A view from the bottom

Eusebio Rufian-Zilbermann — Fri, 17 Oct 2008 03:27:16 GMT

I have been myself on the short milestone model for a while and it is a model that I don't feel it can work for the long term. There are important changes that seem impossible to fit in, and they never get done. Then each discipline starts looking for workarounds when the other disciplines are busy and not looking (devs implementing "stealth" rearchitecture, PMs steamrolling tasks in the schedule, etc). In the end you have a team that works like a person with multiple personality disorder where the disparity in perception of reality (and tensions because of it) just keeps growing. I'm not sure how it will blow up but I'm almost sure it will. Of course, I may be wrong and a reorg will save the day when someone believing in a completely different model gets put at the helm and a new process model (that probably doesn't work for the long term either) gets implemented

re: Engineering 7: A view from the bottom

BigBadBombasticBob — Fri, 17 Oct 2008 04:34:28 GMT

back in the old Win 3.0 days Windows was innovative, with a capital 'i'. It made computers easier to use and more powerful simply by being there. Then 3.1, WfWg, Win '95, Win '98, and Win 2000 all showed vast measurable improvements over their predecessor. People _WANTED_ to upgrade, because it meant "more". XP was only slightly disappointing, but still had enough eye candy, hardware support, and '9x compatibility to get a general thumbs up.

But then... along came ".Net". What was already working well just _HAD_ to be re-invented, and in a bass-ackwards way that made everything that ran with it INEFFICIENT out of the gate (by comparison to what it would have been like). Never mind that senior developers such as myself were _NOW_ faced with the prospect of having to 'learn it all again' (so that we could become JUNIOR developers again, thanks a lot). It was, in my view, the beginning of a disaster waiting to happen.

Products generally go through a life cycle. In the beginning there is REVOLUTIONARY change, where early adopters get on board but that's about it. Then there's INNOVATION, which puts the product into the hands of the power users and early adopters. Then there's the "mass market" which puts the technology into everyone's hands. But without another development cycle of revolutionary change and TRUE innovation, you get _DECLINE_. This is what Vista is doing.

Lessons to be learned:

1. TRUE INNOVATION. Why can't I speak to my computer like they do on Star Trek? Why must I mousie-clickie _everything_ at a rate that is sometimes 10 times slower than it would be by just typing it (like MS-DOS)?

2. FASTER and BETTER (not SLOWER and MORE CUMBERSOME)

3. Eye candy is for SALES/MARKETING execs (regular folk just think it's gaudy and unnecessary). Eye candy doesn't sell OSs and Software.

4. What about the "Killer App" concept? Us developers were _SUPPOSED_ to be the reason Windows was successful! Back to the '.Net' gripe on THIS one...

5. Computer users aren't idiots. They don't need condescending menus and terms. 'Nuff said.

re: Engineering 7: A view from the bottom

Kosher — Fri, 17 Oct 2008 04:53:53 GMT

Bob, well said. I think we need to stop innovating and just focus on what we know best.

re: Engineering 7: A view from the bottom

Asesh — Fri, 17 Oct 2008 05:53:51 GMT

"BigBadBombasticBob: 1. TRUE INNOVATION. Why can't I speak to my computer like they do on Star Trek? Why must I mousie-clickie _everything_ at a rate that is sometimes 10 times slower than it would be by just typing it (like MS-DOS)?"

You can control Vista using by using your voice

Please make shape run properly on Windows7

florindonttouchme — Fri, 17 Oct 2008 07:18:20 GMT

You have broken DX7 badly in Vista. Compare the start transition with WinXP. I can work with you to fix it. Do not say port to DX9, there's good reasons for 7.

LINK: shapeofidentity.com

Mail me for details. Thank you.

re: Engineering 7: A view from the bottom

thecolonel — Fri, 17 Oct 2008 11:19:40 GMT

it sounds to me that you are finally cutting out the right code from win7. with vista all that ever seemed to get cut was the stuff we actually wanted, whereas you left in all the bloat. this new way of doing things appears much more logical

re: Engineering 7: A view from the bottom

GRiNSER — Fri, 17 Oct 2008 12:16:53 GMT

I'm not sure but how do the User Interface- and User Interaction Designers fit into this triad grouping?

re: Engineering 7: A view from the bottom

vikagg — Fri, 17 Oct 2008 14:54:46 GMT

Windows 7 is a make or break OS for Microsoft. If Microsoft does not deliver big time, then it is the end of Microsoft. This is the last opportunity for Microsoft to prove to the world that it is not a has-been company. If Microsoft does not deliver big with Windows 7, then I will shift to Mac OS.

re: Engineering 7: A view from the bottom

michaelljones — Fri, 17 Oct 2008 16:30:15 GMT

Steven,

All great stuff!

I'd really like to see you guys post a little more info on the tools you use internally to manage all of this as well. I know historically Microsoft has been a big 'eat our own dog food' kind of company, so are you managing all of this with Sharepoint? Project Server? Analytics? What?

There are more tools in a PM's toolbox than just one, and I'm curious as to how you've used technology to help those groups interact with one another and keep track of all those interface dependencies.

Thanks and keep going!

re: Engineering 7: A view from the bottom

MattJC — Fri, 17 Oct 2008 21:20:54 GMT

Thanks for this very informative "how the guts of MS Development has worked and is working now" post! I'm glad the re-organization is working out well, and that this guarantees a LEAN MEAN OS this time around, right!? I have three questions I hope you can answer:

1. I've heard that parts(all?) of the GDI interface wasn't hardware excellerated in Vista, will GDI(gdiplus?) be hardware excellerated in Windows 7?, Or is the WPF going to be the new standard.

2. Will there ever be a DirectX10 for Windows XP?

3. Except for the Direct3D interface, are all the other DirectX interfaces completely dead now? I keep reading that they are depricated, but are they truly dead, never to be updated again? I personally hope not!

re: Engineering 7: A view from the bottom

marcinw — Fri, 17 Oct 2008 23:52:48 GMT

@Steven,

Friday, 17 October 2008

Steve Ballmer promises Windows 7 will be better than Vista: "Windows Vista is good, Windows 7 is Windows Vista with clean-up in user interface [and] improvements in performance,"

Windows Vista according to some estimations has got 3 x more entries in Registry. What like what, but it can make even the most powerful PC slow. Add DRM, Indexer, UAC and you have "wow" effect.

I really hope, that Microsoft will think about it and I really hope, that recession will make, that people will think about wasting hardware resources by wrong architecture.

re: Engineering 7: A view from the bottom

marcinw — Fri, 17 Oct 2008 23:57:24 GMT

3 x more entries in Registry than XP...

re: Engineering 7: A view from the bottom

sokolum — Sat, 18 Oct 2008 03:50:20 GMT

Very nice reading, actually every blog was pleasent to read, good work there.

I know that there people are working who u will never hear here, but for u, keep up that good work!

Although, would be nice if there's is a credit page in W7 :)

Making a OS that's more costly than a average Hollywood movie and no credits shown? U should be more proud of our delivered work!

re: Engineering 7: A view from the bottom

lyesmith — Sat, 18 Oct 2008 08:46:01 GMT

"Windows 7 is Windows Vista ..."

This is very disappointing news. I was hoping for fundamental changes in kernel, registry, system requirements, file system etc. Windows 7 will be the same old stuff with more lipstick. :(.

As for home use this will put Windows from 80% below 50% in 5 years.

re: Engineering 7: A view from the bottom

sokolum — Sat, 18 Oct 2008 13:55:41 GMT

"Windows 7 is Windows Vista ..."

Agree, it did sound like an change of lipstick. Sry guys, but that's how is Mr. Balmer made it sounds like.

When reading here the effort u guys are making and the comment of Balmer, sound like a counterproductive commentary of Balmer to me.

There reason choosing the name W7 is of course very bolt to me. No more fancy name's like XP, Vista, Longhorn. But again a plain number 7. To me it sounds like a distance of the existing product name want to be made, because this is such more then another lipstick..... i hope.

re: Engineering 7: A view from the bottom

Domenico — Sat, 18 Oct 2008 16:38:00 GMT

Clarify one thing , Vista is (for now) the Best OS. 360 degrees

People should learn to speak when the final product in our hands ,we can understand from where the Windows 7

but none of us can imagine where arrives.

Sinofksy was clear regarding the compatibility, Imagine having to face new problems Driver, Software, Hardware etc.

Let lose any comment PREFUD, Remember only that there are 2500 engineers to work for us.

9 day for PDC

@Windows TEAM ... Unleash Hell

GO!

re: Engineering 7: A view from the bottom

burgesjl — Sat, 18 Oct 2008 16:51:14 GMT

Well, I guess Ballmer has let the cat out of the bag officially now. "Windows 7 is Windows Vista" and the rest of his comments, shows that he doesn't understand how his company works and what its core processes are. He's a sales and marketing guy - he just wants to sell code and get $$$. He doesn't understand the development process, which is THE key business process at MS; he also doesn't understand why his customers do or don't use his product, which therefore means he's simply a sales guy. I think it would be very instructive for MS to review how Apple make their products: you can tell, they've thought everything through from an integration perspective, and what you have in your hands when its done shows that - its simple, it all fits together. I think its incredible (as an engineer) that the old MS way was to develop code and then toss it over the wall to the testers. However, the current model also has its issues, and primarily its about having everyone on a single team and getting into 'group-think'. I'm going to hazard a guess that a developer likes the current structure because he is, in fact, the center of attention. The testers are now part of 'his team', and they're probably going to knuckle under when he tells them what to worry about or not worry about. Testers need to be the voice of the customer. There are still numerous issues about how they decide what features should be in the product or not, and then how the overall design comes together as a coherent whole. You knew Vista was in real trouble when they starting dropping stuff that should have been foundational like WinFS. At that point, it just becomes a scramble to get something together so you've got something to ship and get revenue against all the hours you spent on it. As they talk about features here, they are very small elements within a wider framework: if the framework is badly conceived much time can be wasted polishing the little elements that then aren't going to get used. There's clearly a hierarchy here of how less important stuff is layered on top of more important stuff, without which the less important stuff (which may be in front of an actual user) won't be delivered. At least there's mention of a 'service agreement' between teams and a clear definition of the interface, beneath which the actual implementation should be allowed to change. This is a core engineering element of 'modularity'. I'm suspecting there's still a lot of legacy stuff in Windows which will carry over and which doesn't follow modular rules, and so that runs the risk of creating massive inter-dependencies at a lower level that can unravel the entire edifice. Until that's addressed, the developers can't make the assumption that what they've done won't be broken by some other change. Its clear MS have issues deciding whether to implement a fix because it requires a need for regression testing, which shows the level of inter-dependence in the code. A simple bug fix shouldn't need this; there's actually clearly a change in behavior/specification so the problem is more profound. It's not the fact that the code doesn't work as designed, its that the design wasn't complete or there's side-effects of the implementation that weren't appreciated. Vista had huge issues with 3rd party graphics drivers at launch because they kept changing the underlying driver model on their partners - there's nothing in this blog entry that suggests that still isn't the case. Given the fact that MS has stated "Windows 7 is Windows Vista" there are fundamental building blocks that are still exposed, the whole thing is built on a shaky foundation. Add to this that entire features are being 'retired' (no more Windows Mail, you've got to go to the 'cloud' of Windows Live to get this functionality - a dreadful decision because people want to work disconnected, like on a plane) or not touched at all from version to version (Notepad, Paint, Movie Maker) shows a rampant disregard for a good design/feature selection process for a coherent product. We get a bunch of stuff that's half-baked and then left to wither, but meanwhile, its a drag on everything else. MS has virtually unlimited resources and unlimited cash, so its a management issue of how to harness all that to deliver significantly new products on a regular basis. With Windows, you're kind of getting the impression MS thinks its largely good as-is and just needs a bit of spit and polish. Ballmer then says the goal post-7 is to get closer to future processor capabilities from AMD and Intel; meanwhile, the existing product still isn't using the capabilities in current generation hardware. Its only just getting around to 64-bit versions regularly shipping on hardware thats been 64-bit capable for about 4 years, and there's still no sign of a 64-bit Office suite. Too slow, too ponderous. And the reason its too slow is because the development process fundamentally doesn't support being fast and innovative, and neither do their sales and marketing practices. Ballmer again: buy and implement Windows Vista today even though 7 is only 12 months away... and thats still 2.5 yrs between releases and they wasted much effort in fixing a botched release. And everyone is conditioned to wait for SP1 (another year) before things are really working as expected.

re: Engineering 7: A view from the bottom

marcinw — Sat, 18 Oct 2008 18:27:10 GMT

I didn't want to make FUDs here, but when I hear, that "Windows 7" (which should be revolution), will be Windows 6.1 and more and more people are speaking about it as "Vista" with some improvements only, I have a little bad feelings...

All current Windows systems are based on NT and have some disadvantages (like shared Registry). After PDC we will see some new facts. But I will read about thousands of new API only, it will be important sign for me, that I should start thinking more about other platforms than WIndows for ALL my tasks.

The truth is too, that people will think much more now, before they will buy new hardware. Wrong architecture will mean, that resources are wasted. The more people will notify it, the more will think about Linux, MacOS or even about leaving x86 for some other solutions...

re: Engineering 7: A view from the bottom

marcinw — Sat, 18 Oct 2008 18:29:27 GMT

in previous post I wanted to say of course:

But, when I will read about thousands of new API only, it will be important sign for me, that I should start thinking more about other platforms than WIndows for ALL my tasks.

re: Engineering 7: A view from the bottom

Domenico — Sat, 18 Oct 2008 20:30:17 GMT

@marcinw

I see more and more and more person declare Windows Server 2008 an operating system phenomenal!

Windows Server 2008 is Vista Sp1 +different Service STOP.

Now think about this system already perfectly stable , 2500 engineers continue to work for 3 years direction By Steven Sinofsky!

You just have faith and do not add anything in speeches futile

re: Engineering 7: A view from the bottom

marcinw — Sat, 18 Oct 2008 20:58:44 GMT

@Domenico,

NT systems are stable starting from beginning. When you don't play with strange drivers or apps, you can make even NT 4.0, 2000, XP, 2003 or Vista rock stable. I don't say, that it's different.

I don't say too, that Microsoft needs to rewrite full system or I don't say: "Linux great, Windows wrong". No.

But:

1. I'm tired with Vista system (resolving many issues among friends and coworkers), Vista word and Vista interface.

2. NT based systems have some disadvantages. If Microsoft want to make revolution, need to work on them. Shared Registry or giving access to WIndows directory for 3rd party apps is very wrong idea. When I see many solutions around it (System Restore, protecting system dlls, etc. etc.) and not main problem resolved, I can not say, that it's resolved.

3. what about DRM ? what will be implemented in main system (how many CPU cycles will be used for it ?) and what in the added apps ?

4. I don't say, that Steven is making wrong work. I say only, that in this moment even his manager is speaking, that we will see not revolution.

Generally speaking: I hope, that Steven will answer on all these difficult questions (yes, he can read them now and because of it this is not futile) after PDC. This is "faith". I don't speak, that WIndows is phenomenal and I don't speak, that it's totally wrong. That's all.

re: Engineering 7: A view from the bottom

Klimax — Sat, 18 Oct 2008 22:57:42 GMT

Hello,

I would like to ask for some things about Win 7:

1.Would it be possible to have "XP-style" dialogs/controls?I mean to be able to bypass "connection center" for example.When in XP I double click network connection(concret device) I'll get status of it and from there I get properties.In Vista there is another layer/window to get past.Can we get back old way?

2.Ability to completely disable any DRM services?(Maybe you got rid of them,but still...) And Patchguard and such if Admin wants(It might not be good thing,but there are still things where this can be neccessary and using 3rd party tools to disable is not good.)

3.Get back ability to chose what parts to install as in 9x or at least get back size on HDD as in XP.(Vista is too space hungry from what I have seen)

4.better built in tools.As Windows were more and more advanced some tools were loosing configure-ability.See defrag.In 9x it had some options,in XP almost none.

5.Optimization.Since Win 7 will for sure require advance CPU,why not to use some extended instruction sets properly.(Like in FFmpeg)

6.Will it be possible to use unsigned drivers with no limitation in x64 version?

7.Since HD is coming surely will there be no limits in usage of HD movies be it on Blu-ray or aired on DTV?

That so far is all.Thank you for answers.

Daniel Klima

P.S.:I have only Win Vista on one notebook as part of offer and based on that experience when setting up it for company network that Vista is not good replacement for XP.

re: Engineering 7: A view from the bottom

bluefisch200 — Sat, 18 Oct 2008 23:16:14 GMT

Nice post about working as a developer on Win7 and how MS work inside...

But i have some questions about Seven:

1. I read about change the Registry to a SQL database...is there something in the air?

2. Direct3D 11/GPGPU and .NET: will be there an easy way to work with GPGPU from the .NET Freamwork

3. Also about the .NET Freamwork...when i work with Aero in my Apps i need to use "dll"s from Windows self...will you put an easyer way to work with Aero in the next .NET Freamwork?

4. Steve Ballmer talks about Seven and by the question about the support of multicore CPUs he say that MS think about...its a little bit hard to hear that becose multicore CPUs are great this time and Vista can not support to mutch Cores(on my 4 Core System its great but i dont now about 8 Core Nehalem and Phenom(45nm) Systems)...

re: Engineering 7: A view from the bottom

wtroost — Sat, 18 Oct 2008 23:36:36 GMT

Interesting article. Vista's favoring security over backwards compatibility makes it impossible to deploy for many organizations. All you need is a single accounting package that doesn't work with Vista and you're screwed.

I hope you're making W7 more compatible than Vista! If W7 can run more XP programs than Vista, that could be a big thing.

re: Engineering 7: A view from the bottom

steven_sinofsky — Sun, 19 Oct 2008 01:36:35 GMT

Howdy,

A number of folks have specifically asked about the DirectX APIs. We have 3 sessions at the PDC on DirectX and of course more at the WinHEC conference.

All the sessions will be available via the PDC site as per the information on the site.

--Steven

re: Engineering 7: A view from the bottom

ndiamond — Sun, 19 Oct 2008 02:47:39 GMT

"As the product moved further along, it got harder and harder to get bug fixes checked into winmain (every bug fix carries with it a chance that the fix will introduce a regression, so the risk associated with each bug fix needs to be measured and the tolerance for risk decreases incrementally)."

I sense contradiction. As the product moved further along, known regressions were perpetuated rather than take the risk of an unknown degree of fixing, and the tolerance for bugs increased to 100%.

"the feature has to be capable of being shipped before it hits winmain"

Where capable of being shipped includes having known bugs. How does this differ from any other policy?

"stuff was checked into the tree that really didn’t work completely"

No ship-it. I assure you, customers are already reminded of this fact, every day.

Asesh: "I have heard that Windows 7 will only be available to some beta testers. Why? It should be open to all the users who would want to test it. That would help Microsoft to fix more bugs and security holes and result in an initial release of a more stable operating system."

No it would not. Here's why: "As the product moved further along, it got harder and harder to get bug fixes checked into winmain..." Allowing more beta testers would increase the number of status changes from "unknown bug" to "known bug", from "not discovered" to "won't fix", from "not yet reproduced" to "reproducible only outside of Microsoft". It would not increase the number of fixes, or stability, or anything like that.

re: Engineering 7: A view from the bottom

steven_sinofsky — Sun, 19 Oct 2008 03:00:46 GMT

@ndiamond -- @Asesh had a good question worthy of a good answer.

Early in the product stages if you have too many testers you end up finding the same issues over and over again, and you can't "break through" and see the functionality of the product. The classic for Windows would be setup and device installation. Even with user interface changes, if the features aren't in a stable and known state, then everyone just hits the same issues.

Thus you don't broaden the test coverage and improve the product quality. Rather you frustrate everyone and create a large volume of bug activity without really any progress.

We announced a pre-beta for attendees to the PDC. The beta will be available broadly as we have promised.

I think your analysis mixes regressions with new bugs. As a product moves further along we do reduce the number of code changes--that's just good software engineering. If something is a regression relative to existing and known functionality / behavior then we address it. However, if new functionality has "bugs" (a bug is defined as "any time anyone experiences anything they did not expect") then you always weigh the risk of new bugs (or new regressions) with the benefit of the code change. There's nothing out of the ordinary about how we approach this.

--Steven

re: Engineering 7: A view from the bottom

murtaza110 — Sun, 19 Oct 2008 11:41:38 GMT

As pointed out by GRiNSER, I would like to know why User Experience and Design people are not part of the triads. I think UX should be integrated in every feature and not just UX intensive features.

Engineering 7: Language & Blueprint

graham.lv — Sun, 19 Oct 2008 12:33:20 GMT

I am not a programmer, (just Delphi code copier) but just wondering if young people would like to know what coding language you use for Windows? I'm supposing mostly C++, but you have so many available.

Your C# is a 'knock-off' of Delphi, right?

You'd be more modular with self contained Delphi wouldn't you? Bigger exe but you could use a 'packer' (to reduce code 50%) and this would provide more security?

I have never used an Apple, but I understand from what I read that you just drag a package to install - then drag it to the bin to uninstall. This is very Delphi like. (Delphi appears to be dying?)

Do you use the same throughout, or mix and match languages?

Interesting that the team is now responsible for the final GUI as well, this leads to the build looking like the final. Normally I would think you would try and hide the final 'look & feel' behind the previous O/S GUI. It's good to know what it looks like a bit, stops critics saying it's just the same as the old one, and doesn't lead to over expectations. But it still leaves room for a few little 'surprises' to be included in the RTM.

As each team is now implementing the UI for their project, you must get a 'blueprint' of the 'look & feel' down to forms, dialogs and glyphs? (No, I've got a big round purple button - Yeah, well I've got a small square yellow one!)

It's good to know where you are all going and not living in isolation - next you'll be having a one-a-month Friday afternoon BBQ together (rain permitting :-) - no sanity please!

re: Engineering 7: A view from the bottom

d_e — Mon, 20 Oct 2008 20:32:07 GMT

@steven_sinofsky

Thanks for reading trough all those comments. Regarding small beta:

Betas for enthusiasts would be a good idea however. First and foremost because you'd make them happy - and some of them are influential. I'm a strong supporter of Vista primarily because I went trough all the betas. And reported non-duplicate bugs. I wasn't using legal ways to obtain ISO images, however. And that's frustrating as well...

If you don't want too many bugreports on the same subject just introduce some form of punishment for duplicates. Like: Every tester can report 2 bugs. For every ACKd non-duplicate bug, a user may report another bug.

re: Engineering 7: A view from the bottom

Mantvydas — Mon, 20 Oct 2008 20:38:05 GMT

While it sounds really good, and might do it for Windows, but really, I do keep my fingers crossed for Windows 7 to be dramatically better in its performance than Windows Vista.

If you, in your triads, are constantly thinking on how to improve the performance of your deliverables, then thanks God, you are on a right track. If you get triple speed in a milestone 3, comparing to what you had in a milestone 1, then I believe it should be good.

If again, if it's only about stability and security, then it's not good enough. Performance is also a must now.

I hope you can pull this off with your new working style.

re: Engineering 7: A view from the bottom

locolorenzo — Wed, 22 Oct 2008 21:29:16 GMT

I am very excited to see how this all goes, MS has always done a more efficient and powerful OS, and bar non Vista has been the best so far.

I look forward to getting my hands on the pre-beta.

I have been reviewing a lot of MS research projects, if you keep on track with what I can clearly see as the future of computing! Imaging Technology/IR is so important to this future. For the doubting lot...look at SideSight, why boink your screen when you can wiggle your finger.

Windows 7 is way better than Vista...

@steven your team is doing a great job, thanks for a what will be a great product.

re: Engineering 7: A view from the bottom

stryqx — Thu, 23 Oct 2008 01:56:59 GMT

Interesting about the SLA between feature teams that share dependencies.

It would be good to see this process extend to all feature teams across product lines - this means that the changes to the Remote Desktop Client for example would have been less likely to break console access to an SBS 2003 box using RWW.

re: Engineering 7: A view from the bottom

Aetyr — Tue, 04 Nov 2008 14:51:31 GMT

Only just catching up on this blog after a couple of weeks being unable to read it, but I just wanted to say that this post is single-handedly the most reassuring post of the entire lot. I like the sound of your management's new perspective. The fact that you ('you' in the plural sense) are not letting rushed, unfinished features into the OS just to boost the list of things Win7 can do is great. I'd much rather know that any feature that wants to get into the OS needs to be of sufficient quality first than know that there were a few more features overall.

I can also see the idea of triads being very helpful. If we had that kind of co-operation between the different teams at my workplace, I can already think of a handful of projects that would have run into far fewer problems.

Thanks Larry, for making me aware of why I should have more faith in Windows 7.

re: Engineering 7: A view from the bottom

wonen — Wed, 05 Nov 2008 17:11:00 GMT

You shouldn't be so negative :)

re: Engineering 7: A view from the bottom

cirurgia plastica — Sat, 15 Aug 2009 17:34:57 GMT