Nick Gall's Weblog

Friday, October 24, 2003

My schedule for the Microsoft PDC 2003 I am keeping an updated PDC schedule available on my blog for those who want to know what I am interested in and for those who want to meet up during the conference; and so I can access it myself from any browser. (Hint: the title of this post is the link.) 9:29:04 AM

Monday, October 20, 2003

Communication: less control = more effectiveness. Microsoftian Randy Holloway writes regarding Microsoft's new blogging friendly policy: The more you try to control the message, the less effective the message will be. The reason is that people today can tell the difference between valuable information and canned marketing material. Picked this up via a post in Richard Gayle's blog. I wish all companies were so open to giving up some control (hint, hint). That reinforces a good article ("A Better Longhorn Through Blogging") on how Microsoft bloggers are opening up Microsoft's development process and making it more transparent by writing about it as they do it. 6:01:21 AM

Benchmarks and Bathing Suits. Reposting a great analogy from Werner Vogel's blog: Benchmarks are like bathing suits: they reveal a lot, but always manage to cover up the more interesting parts ... I changed bikinis to bathing suits to make it gender neutral. 4:46:35 AM

If I had twenty minutes with Bill and Steve... (I've been meaning to post this for a while. My recent visit to Microsoft prompted me to finally do it.) If I had twenty minutes with Bill and Steve, I would discuss the following recent farewell letter from ex-MSer David Stutz. I think this is a tremendously insightful critique of MS's core strategic vulnerabilities. This letter reinforces my belief that the biggest strategic vulnerability staring MS in the face is a software paradigm shift that is outperforming MS at its own game -- commoditizing software. This shift is what IBM calls the "network" era and what some of us a META call the xWeb era. David Stutz refers to it variously as "networked ways," "the network integration model," and "networked software." Whatever you call it, open network protocols and the open source implementations enabling them are redefining the economics of software. The (indirect) "network effect" value of widespread Windows use pales in comparison to the (direct) "network effect" value of the Web services protocol stack being built under the leadership of MS. But just as Sun has not been able to monetize its leadership of the Java API/software stack as it is being built, there is no guarantee that MS will be able to monetize the WS protocol stack as it is being built. As document formats drive towards XML-based standards and collaboration protocols do the same (e.g., SIP, Jabber IM, BPEL, UIML) and open source implementations of these protocols arise, how does MS low pricing strategy still save the day? This cuts across IW, .Net, and Windows product families. So my question would be, "Bill and Steve, Do you think you really turned MS around to catch the Internet wave back in 1995, or did you just survive the initial "tidal wave"? I am linking Bill's infamous May 26, 1995 "Internet Tidal Wave" memo because it reinforces for me how clearly Bill saw the magnitude of the change being wrought by the Internet, but how poorly he or anyone else saw at the time the details of how the Internet would upend everything. (FYI, the ONLY place I could find it was at the DOJ site listing government evidence in the case!) You can see in the memo an understanding of the threat that Internet "protocols" are to MS APIs, but a lack of understanding that simply offering up a MS protocol to run over the Internet was not going to cut it. Microsoft seems to only now be aware that only open protocols, not owned by MS, will achieve significant "network effect." And so how will MS thrive as a software company as the economics of software fundamentally change underneath the commoditizing effect of open Web services protocols? This is ultimately a question about how Microsoft gradually reinvents itself from a software products company to a services company. No not professional services (although it will need to have a strategy here); rather a services company in the direction of MSN, Hailstorm services, Passport services, MSDN developer subscription services, Project Summit, etc., but moving beyond the challenges and failures of these initial attempts. In doing so, MS's biggest challenge will be to change its image. I would pick up on David Stutz's comment that "Recovering from current external perceptions of Microsoft as a paranoid, untrustworthy, greedy, petty, and politically inept organization will take years." It certainly took IBM several decades and a near death experience and outside leadership to overcome its monopolistic image. Thus the three big strategic opportunities/challenges for MS over the next decade or so are: (1) the changing economics of software--MS being out-commoditized; (2) transforming MS into a services company, especially network-based services; and (3) changing public perception of MS from a monopolistic bully to a trustworthy strategic service provider. 4:41:30 AM

Thursday, October 09, 2003

Different types of dynamic "parametric variation"? In discussing Microsoft's Dynamic Systems Initiative (DSI) and its System Definition Model (SDM) with someone from Microsoft, he made the point about the importance of "dynamic configuration" (DCon). I thought about my definition of service-oriented architecture (SOA) to see if I had dealt with DC as a first class concern. It struck me that I had put "dynamic binding" (DBnd) front and center, but not DCon. So I immediately asked myself, "What's the difference between DBnd and DCon, if any?" Rather than reinvent the wheel, I went to Google to see if anyone had already answered this question. My search for "'dynamic configuration' 'dynamic binding'" returned 352 hits. I even extended my question to ask "and what's the difference between DBnd, DCon, and 'dynamic composition' (DCom), the hot concept wrapped up with SOA?" When I searched for all three terms, I got 17 hits. So my initial search leads me to believe that no one has significantly explored the relationships among binding, configuring, and composing. I can only make several basic observations about their relationships. First, all are concepts based on modularity. Second, they are all forms of what Richard Langlois refers to as "parametric variations" as opposed to "structural variations." This is an interesting choice of words, given that one sense of binding is "binding a parameter to a value." Third, binding suggests two domains being related by the binding, while configuration suggests relationships within a domain or system. In all likelihood, these terms can be arranged along a single dimension of increasing generalization, say the scope of the parameters that can be varied. Obviously, they also range across the dynamic dimension from static to real-time (e.g. binding from one wireless protocol to another without interrupting the session). 5:31:07 PM

Wednesday, October 08, 2003

What is Music. Given my long running now-half-joking quasi-philosophical discussion of "What is Music" with my friend and former colleague Anthony Bradley, I couldn't resist (re)posting the following post from Slashdot. My favorite part is Ka-Blamo, which converts any binary file into a WAV file by putting a header in front of it! (BTW, All this started when Anthony foolishly <grin> denied that Rap was Music.) What Counts as Music and Why? Posted by michael on Friday October 03, @07:40PM from the perl-jam dept. The Importance of writes "There has been much discussion about compulsory licensing schemes. Most of the debate has been about music. But what happens when any file can easily be converted into a sound file and back again? Can shareware authors convert their software to digital music and get paid for sharing it? Can pornographers get paid for turning images into sound? Scott Matthews has written a program (Ka-Blamo) that does the conversion. LawMeme looks at some of the issues. This raises the question, what should count as music and why?" 7:10:20 AM

Saturday, October 04, 2003

Web Services Tools. Excellent survery of Web services tools. 7:13:33 AM