Robert McLaws: Windows Edition

Blogging about Windows since before Vista became a bad word

August 2004 - Posts

  • Longhorn Shakeup - Part One: The Social Perspective

    This is the first in a series of commentaries I'll be writing about the changes that are coming in Longhorn. There's going to be a lot of speculation over how this will affect the Microsoft roadmap for the next decade, so I thought I should throw some inside insight into the fray. Now, much of what I have to say will probably appear to follow the "party lines" that Microsoft laid out on Friday. What I'm hoping to discuss over the next several entries is an expansion on what has already been revealed, in hopes of connecting the dots for the rest of the community.

    While most people will be focusing on the removal of WinFS from the main release, I want to take the time in this first article to talk about some of the backstory that lead up to Friday's decision. The past year has been very interesting for Microsoft. MSBlaster shook the company hard, and really forced everyone in the organization to take a hard look at how they did business. Up until that point, there was a real disconnect between the company and it's customers. Well, Microsoft caught an earful from those disconnected customers in the hours, days, and weeks following the attack. And, in the company's defining moment, they started listening.

    But Microsoft had been developing Longhorn in secret for at least 18 months, maybe longer, prior to MSBlaster. This development track had been predetermined, and predestined to follow the same old Microsoft dev cycle: built in the Echo Chamber, with few outsiders defining how the OS was going to function. Following that same ol' path was going to introduce the same ol' problems, but a burgeoning new movement inside the company would soon change that. 

    Two months after MSBlaster came the PDC, where Microsoft did something it had never done before: it laid it's plans bare in front of 3000 developers, reporters, and IT planners years before the finished product would be released. The plans were ambitous, and the world cheered. When Microsoft said it was betting the company on Longhorn, it was only partially accurate. MS really took it in the shorts with their "let's take over every tech company" strategy, so they really needed the whole "corporate transparency" thing to work for them. While it would appear to the world that Microsoft was betting the company on a revolutionary new platform, the bigger bet was that the world at large would welcome Microsoft's newfound benevolence.

    So in October 2003, Bill Gates & Co opened the kimono (and, surprisingly, Don Box didn't) to show the world Windows Codename 'Longhorn'. They touted new architectural strategies, new capabilities, and a new slew of confiusing acronyms (WinFS and WinFX... come on guys) that would make the computing world a better place. It was the "revolutionary" new operating system that would usher in the "Digital Decade", with large swaths of the company's product lines being delivered in synchronized waves.

    But as Bill was touting the future, already the need for change was apparent. Features that were being developed for Longhorn suddenly needed to be done yesterday, and Windows XP Service Pack 2, previously only a hotfix rollup, now had to break the mold and introduce a significant new feature set. Never before had Microsoft fundamentally changed the operation of a product in a service pack, and the ramifications would be huge. So Microsoft made one of the most expensive business decisions it ever made. They ground development on their next OS to a halt, and moved most of the Windows team to solving the security problem in Windows XP.

    Many people underestimate the significance of Service Pack 2. I personally believe that Microsoft should have called it Windows XP2, because it really is a new version of Windows, but I don't determine that stuff. A lot had changed since it was released in 2000. There were better compilers that did a better job checking for "buffer overruns" and other trivial bust nasty coding flaws. So Microsoft completely recompiled the codebase to take advantage of the increased protection. There were hundreds of bug fixes, as well as vast changes to the core of IE, which forms the basis of most of the Windows browsing and navigation experience.

    But I digress. So Microsoft didn't just stop working on Windows while SP2 was in development. They had evangelists across the country talking to ISVs, IT implementers, and end users, getting feedback on the public and private versions of Longhorn that had been released. They had internal teams prototyping Longhorn applications to get a feel for what ISVs would be going through. And they were surprised with the results.

    Compared with the difficulty in installing, using, and maintaining Windows; the prevalence of spyware, "browser helpers" and other malware; and the lack of consumer education on the perils of a connected society; the thought of a new desktop rendering system wasn't all that exciting. On top of that, it was becoming increasingly apparent that WinFS was not as far along in the development cycle as it's pillar counterparts. So Allchin did exactly what he is paid to do: he caught wind of the problem and realized that action needed to be taken.

    What happened next is the same thing that happens in every software development cycle. The problem is, while Microsoft went great lengths to educate people on what they wanted to do with Longhorn, they virtually ignored educating people on how software development works. This miscalculation will prove costly in the PR war that Microsoft battles daily. The armchair quarterbacks of the tech world will rip MS a new one, but Microsoft will learn from its mistakes, and the process will go on.

    In the weeks leading up to Friday's announcement, Microsoft had to look at what it wanted to do, versus how long it would take to do what it wanted, and how much that would cost to accomplish. They also had to factor in the new feedback, and adjust their priorties in regards to usability. So you would think "Well, Microsoft has $70B in the bank, they can afford to do whatever they want." In an ideal world, that would be true, but Microsoft is a publicly traded company, answering to shareholders that will not tolerate a lull in profits. Something had to give, and Microsoft came up with an unorthodox way to do it.

    In the end, Microsoft got what it wanted, but not in the way it wanted. They said that Longhorn was going to be a revolutionary operating system, and I believe that it will be. The architectural side will definitely improve, but the revolution will take place where it really belongs, in ensuring that computers are as easy to use, easy to manage, and resiliant to attack as possible. And that can only be good for everybody.

    In my next installment, I'm going to talk about the death of the 'software wave', and what it means for the Microsoft roadmap through the end of the decade.

  • Gates On Longhorn gets way too many scoops, especially since they like to spin stuff negative. Ah well. Read what BillG had to say about today's announcement.

  • Windows 2006 Cutbacks: The Long[horn] And The Short Of It

    Ok, so the press is going to have a heyday for the next month over the news announced today. This has been an open secret for quite some time, and it's good to see some actual information on what has been going on. I just got done with a round of interviews with some of the major news outlets, and I'll be assembling a post with links to all the comments all over the web about what's going on. But I wanted to take a minute and clear the air about what is happening.

    1. Microsoft is decoupling the development of the "Longhorn Pillars" to speed up the development process. I'll go into more detail later, but basically there are too many interdependancies for the Longhorn development process to be manageable, even by a company as large as Microsoft. If anything, the scale makes management of the process infinitely more difficult.
    2. Avalon and WinFX will still be delivered as a part of Longhorn. They will not, however, be dependent on Longhorn. This means that they will also be available to run on Windows XP and Windows Server 2003.
    3. The .NET Framework 2.0, code named "Whidbey", will be installed by default in Longhorn. At this point, it does not look like there will be any major parts of the OS that will be dependent on Whidbey by default. This may change, and those kind of details are still being worked out, so don't take that as static information.
    4. WinFS is being shipped as an out-of-band add-on pack (a la ADAM) for Longhorn, to be shipped after Longhorn is released. I'm going to go into detail on that tomorrow, but here's the drill. Microsoft won't publicly confirm this, so don't ask, but it only makes logical sense that Microsoft will now deliver it as part of Longhorn Server, with an "integration pack" for Longhorn Client, sometime in 2007. As I said, I'll go into the hows and the whys in a detailed post later this weekend.

    That's it for now, as I said, I'll be culling resources together to put together a more accurate picture shortly. In the meantime, I would recommend that everyone refrain from going nuts until more information is available.

  • Revisiting Longhorn's 'Priorities'

    "SP2 was a major milestone for the Windows development team," [Microsoft] said in a statement Wednesday to CNET "Now that it has been released, it is a natural time to revisit Longhorn priorities."

    The needle on the Longhorn Richter Scale keeps rumbling, and change is in the wind. Microsoft opened the kimono on Longhorn way early... but was it too early? Refocusing happens in every software project, but expectations on this version of Windows are so high, how will everyone react if the things that alludes to are true? If the inevitable happens, and there is a backlash, will Longhorn scalebacks be followed by a clampdown on the flow of information from Redmondites? One can only hope that rational thought will prevail.

    Stay tuned.

  • XPSP2 RTMs, What's Next For the Windows Team?

    It's been a long and bumpy road for the Windows team this year. I was in Redmond around this time last year when MSBlaster hit the scene, and let me tell you, it was quite amazing to watch. It had to have been a lot like the attack on Pearl Harbor. Microsoft was calling all hands in on nights and weekends, doing emergency customer service training courses... it was crazy. I might be in the cheap seats of the Diaspara, but from here I think it was the best thing that ever happened to Microsoft. I don't remember if they've caught the original writer or not, but as much damage as it did, you gotta give they guy snaps.

    Now I know I know, I'm not going to sit here and condone hacking, that's just irresponsible. But they did more than just expose a crack in Windows, s/he exposed the gaping wound that lay just below the White Knight's glistening armor. Let it not be said that one person can't change the world, because it just isn't true. That person awakened the sleeping giant that has sole responsibility over 90% of the worlds computers. And just as the world is a better place because of Pearl Harbor and September 11th, the over the next several months, from this point forward security will always be at the top of Microsoft's priority list.

    So the work on Service Pack 2 is complete. You all have been watching with bated breath for Longhorn news, leeching leaked builds from the Usenet, and what not. As tech-savvy users, it is your responsibility to make sure everyone you know has installed this update as soon as you get access to it. It's more important than anything in Longhorn, because it's here today, and it improves the security and stability of your machine TODAY.

    Now, at the same time all this has been going on, there has been a lot of complaining about the constantly slipping Longhorn release date. I haven't weighed in on that too much yet, but I think it's time to break my silence. Microsoft shifted between 80-90% of the Windows Client Team off Longhorn development and onto Windows XP SP2. Now, a lot of the features, like IE popup blocking and the No-Execute technology were originally supposed to be in Longhorn. Following the Battle of Blaster, Microsoft regrouped and decided, and rightfully so, that these features could not wait until 2006/2007.  Now you may not have known this, but XP2 was even compiled differently to prevent buffer overflows.

    So, in the Longhorn space, this is why we've only seen two public builds, 4051 and 4074. There were only 23 internal builds in the 8 months between the PDC build and the WinHEC build. 23 in eight months. I tested practically that many builds of XP2. It's the end of the first Friday in August, and you better bet the Windows Client Team is taking a long weekend off. I bet the probably even have the day off on Monday. But next week the WCT begins an even more daunting, and extremely expensive, task: Moving the XPSP2 code BACK into the Longhorn codebase.

    You see, when Microsoft started working on Longhorn, they forked the Windows codebase. So, for the past year, Windows has had 2 parallel development tracks, Longhorn and SP2. Longhorn development started with the Windows XP SP1 codebase a few years ago, and I'm sure at that time, Microsoft pretty much figured that was it. But one person changed a $70B company, and now you have the next revolution on the Desktop that is actually behind what is currently available in many respects.

    If you've been fortunate enough to play with Longhorn 4074, you'll notice that it doesn't have No-Execute protection yet, it doesn't have the Windows Security Center, and it doesn't have the brand new WiFi/Bluetooth stacks. All these great new features need to be brought back on the Longhorn train so that it can pull out of the station before 2007.

    So why did I just go off about all this? Well, I guess it's so that everyone can get a little perspective on the situation. The media talks up and down about how Microsoft's ship dates affect their bottom line, and that's what they are paid to do... talk. But events of the past year have made it readily apparent that, while the inevitable delays cost money, using software that was hurried out the door over a blip on the calendar can be exponentially more risky.

    Besides, Microsoft needed to give Apple enough time to make itself look retarded by saying that its next OS is Longhorn, without even coming close to having a comparable feature set.

    But that's just my opinion, I could be wrong.