Once More unto the Breach

Microsoft to Standardize Office Formats in ECMA

The definitive opening post on this is Andy Updegrove.  Andy is an experienced standards lawyer and does wonderful analysis of the news to date.  You would do well to start here because he focuses on the core of the Microsoft strategy starting with the press release and positions the reporting around it perfectly.  (He also collects the relevant press articles.)

Jean Paoli (Scoble interview) and Brian Jones (blog) are proud technologists, so discount their blogs and save yourself some time if you're only interested in watching the standards strategy unfold. (This is not a snide remark -- how do you think they'd talk about their work?  They're both appropriately strong supporters of this effort.)   

The real fuss will begin once the patent license is presented (which both Jean and Brian have misunderstood in the past as well).  Ecma International requires that participants offer reasonable and non-discriminatory (RAND) terms around intellectual property with respect to their standards. 

It will likely be a royalty free license, because the current patent license around the proprietary specification is royalty free.  That patent license, however, couldn't be sublicensed, so an implementer that wanted to license their implementation under the GPL couldn't.  Indeed previous examples around the IETF SenderID standard would force users of other implementations to engage in a license with Microsoft which is a rather onerous problem for free and open source licensed software.  Microsoft will of course hide behind the "it's our property" rhetoric, and label anyone that doesn't agree as either a "thief" or an "IP socialist".  Standards can be a rhetorical game and Microsoft has a fetish about intellectual property (which may well reflect on their constant focus on competitors since IP is a vendor-to-vendor discussion).

As Stephen O'Grady points out in the ZDNet article, there is prior history here.  The C#/CLI standards were also developed at Ecma International, a royalty-free patent license offered to essential patents, and Microsoft will still make no definitive comment about the mono project thereby keeping it in an ambiguous state with respect to its potential customers.  While Novell continues to sell and ship mono, one can only presume there are lawyers from both sides snarling at each other in the ever present legal struggles between these two companies. 

While standards exist to encourage multiple implementations, Microsoft wants to have the imprimatur without the attendant loss in control of what it perceives to be "its property".

It will be fascinating to see the early technical committee meetings.  No changes will be allowed to the Microsoft formats of course, because they will claim that "changes are out of scope" due the millions of users presently using the format.  (This is certainly how the Open Software Foundation played the GUI Wars around OSF/Motif in the early 1990s.)  Indeed the point is to perfectly align the new format standard with the about to be shipped product.

The stress will be discovering how hard it is to maintain exact conformance.  One bug fix after the standard locks but before the product ships could break the link between product and standard.  Likewise, one necessary editorial correction to the document if the product ships first could render the product nonconforming with its own spec.  There could be fascinating compliance angles to be played here by the other vendors. 

This is not to be ignored.  At least one first tier UNIX vendor in the mid-1990s discovered after they shipped, that their C-language compiler was no longer compliant to U.S. government procurement standards, and had to sit there quietly through a two year ship cycle hoping no one at their competitors or in the U.S. government would notice while they fixed it.  Likewise, it would be interesting to know if Microsoft's .NET world conforms to the existing ISO standards based upon its own products.  Indeed the reaction to the question alone would be worth an interview.  (Are all conformance issues fixed in the .NET bug database?  I bet we could check in the open source world of mono.) 

The specifications review work will still (of course) be perfectly tedious. One interesting item will be the publication debates.   Should the standard in all its complexity be published as the XML schema itself?  The precedent was certainly set with the C#/CLI standards and the frameworks at Ecma International.  But that forces a different set of difficulties on reviewers, and one has to have the tools that can run presumably on more than Windows.  Which means there could be reference implementation code to leverage early, but at what cost and under what terms of use?  There are many ways to skin this cat. 

One of the more interesting Microsoft commentaries is from Jason Matusow (Microsoft Shared Source maven) .  It's good that he's getting his opinions in order.  Jason is a master of rhetoric, a brilliant presenter, and a glutton for punishment as the man from Microsoft that gets to present world-wide on Microsoft's open source and Shared Source strategies. 

I've long argued that technology standards is a classical game of diplomacy where after the receptions and tea parties are over, the job is to defend sovereign territory while expanding one's area of economic influence.  If you're really good, your best work goes unseen.  Microsoft would do well to convince Jason to take a new position as technology diplomat to sell the Office standards pitch.  It will require someone with his patience and skills and long experience presenting to governments and talking with the press and analysts to carry this forward properly. 

Let the games begin. 

Open Source and the New Microsoft Strategy

I started the week posting why I thought MSFT wouldn't be trading in 10 years.  It is a dinner wager started at last week's Open Source Business Conference.  A few friends publicly chided me, and with good reason.  It is a rather bold statement. This week's leaked memos from Ray Ozzie and Gates give me hope however that I may get dinner sooner.  Let's look at the developing discussion.

Matt Asay tackled the US$60B war chest angle, and pointed out the survivability of IBM, HP, and Apple to counter my DEC example.  Stephen O'Grady outed himself as one person on the other side of the dinner wager in his del.ico.us links for 9 November, but in the same collection includes pointers to the ZDNet story on the Ray Ozzie and Bill Gates "leaked" memos, as well as this excellent news item on the French tax agency (Direction Générale des Impôts) moving 80K desktops to OpenOffice on Windows to considerable cost savings.

I responded to Matt in comments on his post, but collect that response here:

I agree that a $60B cash reserve of shareholders money sure looks like a lot for a company with a market capitalization of $260B depending on the phase of the moon. Of course when the market starts correcting the stock price down, that $60B looks more like an asset than a war chest.

I would also point out that each of his examples survived for different reasons:

• IBM is a much older corporation that has transcended its original product brands and has rolled over multiple executive management teams, so as a corporate culture they can withstand new executives and they're not trying to tie all their new advances to the Selectric typewriter brand.
• HP is a collection of companies flying in a general direction under a brand, similar to a GE.  It can rapidly "buy" and "sell" divisions to change the mix of offerings to customers.  While this might prove a little hard on the employees in the divisions, it is a viable way for a corporation to survive. 
• Apple has focused religiously on its customers and their brand from its inception. They focused on profitability over market share.  They are laser sharp in their drive for design. Without the extreme pressures for growth, they had a chance of maintaining the cult of the customer and customer focused design within their culture. 

Microsoft has much more in common with Digital Equipment Corp. than any of those three companies. 

The two memos from Ray Ozzie and Bill Gates are actually timely demonstrations of why I think MSFT won't be a trading stock in a decade. Robert Cringely does the best job of tearing the memos down here.  Most damning was the simple business observation that:

Remember that for Gates and Ballmer to be happy, Microsoft will have to maintain $2.5 billion per month in revenue and $1.5 billion per month in profit. That's FIVE TIMES the size of Google without Google's ad expertise or ad infrastructure.  It simply won't happen.

Even Matt took umbrage with the Ray Ozzie memo's focus on advertising for the simple reason that we as customers are looking for ways to avoid advertising, not embrace more ways for it to become central to our existence to support their business.  I'm a happy TiVo user because I can watch television the way I want to watch it. 

I would point out a couple more things in the leaked memos:

• Once again Microsoft is focusing on the competition as an aggressive follower, rather than focusing on the customers to deliver new solutions to problems they want to buy. Hint: Focus on customers.  That's where the money is. 
• Ozzie paints this bright new world then says: "But in many areas we have 80% of the product and technical infrastructure already built – we just need to close the 20% gap." What's the message for the team here?  If you work just a little harder we'll win?  We might move the stock and be able to reward you the way we rewarded employees in the end of the 1990s?
• It's not about tightly integrated innovation.  I don't want to be beholden to a single software vendor.  I don't care if my TiVo doesn't run the same software base as my Linksys router.  The successful companies are disintermediating the big companies. 

The complete failure in Ozzie's memo to recognize open source as a development tool or community as a customer engagement mechanism is stunning at this point in the industry.  So many of the things he wants to enable and compete against were precisely enabled through open source community developed software. He observes:

"The rapid growth of application assembly using things such as REST, JavaScript and PHP suggests that many developers gravitate toward very rapid, lightweight ways to create and compose solutions." 

The ratios we have been unable to escape in decades of software development are how few lines of bug free code a good developer can deliver per day, how many bugs an excellent software process will still deliver (~1:1000 loc), and the cost of maintenance over the history of an application in deployment (~80%).  Every new strategy around architecture and re-use (modules -> objects -> components -> web services) and every new programming paradigm (object oriented, little languages, dynamic languages and scripting) has been focused on beating these ratios by reducing the amount of code we write and maintain.  Software (just like I talked about business execution the other day) is hard work.  Collaborative development is proving to be the best way to develop high-quality re-usable components. 

Jason Matusow gave an excellent talk on open source at Microsoft at the Open Source Business Conference last week.  Clearly Jason "gets" it.  And I know for a fact that there are others at Microsoft that also understand the business tool open source represents.  I have certainly written about their trials and lost opportunities in this regard before.  But the executives in these memos effectively just told the world that open source is not core to the mission. Their loss.  But the company also just told all their employees that open source is not core to the mission. As I observed before, executive bandwidth is one of the rarest of resources at Microsoft. Getting executive sponsorship for projects that might save the company depends upon such bandwidth. 
 
The nice thing about this dinner wager is that if I'm wrong and I lose, I won't have to pay for ten years.  On the other hand if I'm right, I'll be able to collect a lot sooner if this is the way the company is going to execute. 

MSFT will not be Trading in Ten Years

Update [11-Nov-2005, 12:32]: This post started some fun discussions (caught in the comments and trackbacks).  I've followed up with commentary here.

This began as a friendly hallway wager (for dinner) with friends at the Open Source Business Conference this past week.  The discussion has expanded, however, beyond a circle of friends so I thought it best to clearly state the "what" and "why" of the bet. To be clear, I don't wish the company ill will.  I just think they're about to discover some new realities.

Last Spring I posted an opinion on OpenOffice in a slightly different context.  Here are the bits relevent to this discussion:

• It will be OpenOffice on Windows XP that will cause a revenue slide, a few quarter on quarter losses, and the shareholders to react bluntly.
• Firefox has given Windows XP desktop users a taste of open source. The average consumer is merely downloading the binaries and discovering it's a more than good enough user experience.
• OpenOffice on Windows XP becomes an interesting next step. 
• With no licensing fee, it becomes a very inexpensive experiment to load OpenOffice on a large class of workstations in the average enterprise that already pays for Microsoft Office through its Enterprise Agreement. 
• OpenOffice reads Microsoft Office documents in a very direct manner. 
• The user interface is familiar enough.
• The ability to generate PDF for distribution in OpenOffice is also a god send.
• Once the experiment runs its course and the numbers become apparent, the enterprise procurement team can then have a different discussion with the Microsoft account exec when they arrive to negotiate the next three year enterprise agreement. I imagine they'll be able to at least call for a 20% reduction in Microsoft Office licenses.  (The sticking point for a larger percentage will likely be the Outlook email client, but even that switches nicely to Thunderbird on Windows XP.) 
  

This licensing negotiation will be even stickier if upgrading to the next version of Microsoft Office really requires Windows Vista to run. 

In the Innovator's Dilemma, Clayton Christensen talks about over delivering on functionality, but in the Innovator's Solution he gets into what actually happens in the market at that point.  When an encumbent starts over delivering, and customers can't absorb the new innovation fast enough, (and therefore are unwilling to pay for it) a call for standardization comes from the marketplace.

We have the terminology and economic models to actually see what's happening now.  We can witness forward so to speak, and not have to wait for a decade to dig through the data that would only have arrived too late to save the company anyway.

The standard has happened.  It's OASIS Open Document Format (ODF).  The vendor competition is driving this home.  Witness the speed of action from IBM and Sun supporting the Commonwealth of Massachusetts decision to require ODF for document archive purposes.  Andy Updegrove has a couple of excellent reports from last Friday's Open Document meeting in Armonk. 

Even if the political machinations that are ongoing in the Commonwealth cause the decision to be badly delayed, other governments will begin looking at the technology and route around the political minefields in their respective situations. Corporations will follow. 

Microsoft will react in very predictable ways.  (Here I'm even willing to give them the benefit of the doubt on their role in the Massachusetts political machinations.  I don't tend to ascribe to malice what can be explained by simple ignorance, or in this case the pettiness of politicians that weren't asked about an issue they felt they should have a say in.)  First, they will tell the sales force to "not lose to OpenOffice" because they believe it's about innovation and marketing, not value and solution.  They may even tweak compensation models to enable the sales teams to rapidly discount around the OpenOffice experiments.  This will accelerate their problem.  But regardless of whether an enterprise negotiates a discount based on lowered Microsoft Office licensing requirements or because the Microsoft sales team dropped the price, it's still lost revenue. 

Second, they will continue to hammer away at the message that it's about innovation, and tightly integrated innovation at that.  But the problem is that they've already innovated past the needs of most of their users in Office.  Delivering more innovation exacerbates the problem.  You can almost hear the bewilderment in Alan Yates statements around Microsoft Office formats versus ODF.  He doesn't "get" why customers would choose a "less innovative" solution. 

This situation is very different from Linux versus Windows: 

• Switching to a Linux desktop is a risk prone experiment in the eyes of the average CIO.  Here they can easily load OpenOffice beside Microsoft Office on the same machine for free. 
• Tightly integrated innovation still works in the delivery of Windows desktops.  Microsoft Office has massively over delivered and more innovation is the last thing customers really want. 
• Because of the fidelity between OpenOffice and Microsoft document formats, this situation is much more dangerous than the tough migration of Windows to Linux, and the speed that things change will likely be very very rapid. 

Microsoft Office represents a considerable amount of their revenue stream.  It won't take many OpenOffice experiments (15%?) to impact that revenue stream.  So there will be a down quarter or two and Wall Street will punish Microsoft through its stock price.  Microsoft will have to behave differently: they won't be able hire, compensate, and retain staff the same way.  They won't be able to execute in the same way.  A number of friends made reference to the war chest cash reserves Microsoft has as if this somehow proved the argument for them.  One has to remember that the cash reserve, large as it is, is the shareholders' money.  It's profits.  They won't simply be able to spend the cash reserve either. (Indeed one of the toughest things I learned in my first start-up was the hard decisions you are forced to make because of your fiduciary responsibility.) It will be interesting for the financial analysts amongst us to see how creative Microsoft becomes trying to cover for a down quarter out of the cash reserve. 

Microsoft is still on its original executive team.  When IBM stumbled and Gerstner came on board, IBM was already well past its first CEO and early brands around adding machines and later Selectric typewriters.  The Microsoft exec team still believes it's about delivering more and more value on the Windows and Office brands.  They will continue to execute on the current Windows and Office cash centers as well, because that is their high margin business, and that's what Wall Street rewards. 

Surrounded by the bricks and mortar of the Redmond campus, it's hard to think of a world without Microsoft, because there is obviously so much going on around you.  There are people everywhere.  The grounds are beautiful.  But as I sit here in Boston writing this post, I can remember a different time and a different company.  We wouldn't have believed Digital Equipment Corp. could fail either.  Yet in the space of 2 years from 1988 to 1990 they went from industry juggernaut at the top of their game to a very troubled organization thrashing about in a new market order. And a few years after that they were acquired by Compaq (at the height of their game). 

DEC had over delivered.  Customers couldn't absorb more and more innovative VAX technology each year.  Standards happened (POSIX/UNIX).  Customers migrated away from a vendor locked de facto technology to a de jure based standard with multiple implementations that were initially inferior to the de facto technology. The economic model works.  We are indeed watching history present itself. DEC employees went on to other jobs — some with Compaq — and some now with HP.  Buildings had new signs put on them.  The economy moved on and the high tech industry still calls Boston a home. 

Microsoft's hubris will be a belief that the software business is different.
Quod erat demonstrandum.
 

Microsoft, Open Source and MPICH2

This is great news (eWeek article)!  The coming Windows Server 2003 Compute Cluster Edition includes the MPICH2 library from Argonne National Labs.  N.B. This isn't Microsoft playing the Shared Source pitch — it's Microsoft using open source appropriately and participating in someone else's community.  I had coffee with Kyril on a number of occasions before I left to discuss how they might go about doing this.  It's great to see they worked it out and engaged!  Congratulations!

Now run faster! :0)

Microsoft, Massachusetts and a Standards Primer

Updated (22-Sep-2005, 23:50 PDT): It's official as of yesterday.  The Enterprise Technical Reference Model V3.5 is now published.  FAQ is here.  Official comments are here.  This post has been kindly referenced from LinuxToday, NewsForge, and Bob Sutor (IBM) on his blog. I've no idea from where it's been unkindly referenced.
Updated (19-Sep-2005, 15:00 PDT): Started collecting blogs and news at the end in a further links section, starting with Nicolas Carr's post. 

The discussion is certainly heating up around the Commonwealth of Massachusetts' decision to move to the Open Document Format standard.  Simon Phipps (Sun) has an excellent blog entry on the framing that has begun.  Bob Sutor (IBM) also provides excellent insight into the vendor positioning from his point of view.  In the past week, the official vendor comments (Sun, IBM, Corel, Adobe, Microsoft) have been flowing into the Commonwealth. 

The Microsoft response is a study in exactly this sort of framing game.  While the others are a page or two supporting the decision, the Microsoft response is a fine example of norm entrepreneurial rhetoric at 14 pages.

So let's shape that frame a little. My experience here comes from more than 15 years as a developer of standards, a user of standards, and an implementer of standards, as both consumer and vendor, and across multiple products.  My viewpoint is not constrained to narrow legal concerns with how standards impact my IP portfolio, but rather the broad market economics of standardization.

First, let's start with the market.

Christensen observed that when a technology space is immature, and the technology isn't "good enough", the vendor that controls the entire delivery process (i.e. tightly integrated innovation) wins.  Once the vendor starts to over deliver on customer needs, indeed customers can't absorb (and so don't want to pay for) new innovation, then there is a call for standardization in the market place.  At this point, the component players can get involved, and tightly integrated innovation no longer wins the day. 

You might imagine that the call for standards is a market signal that the competitors love as well.  They get a wedge to drive into the incumbents business to those over-served customers. Customers believe the standards are good (often sight unseen) because there appears growing choice in the market which at the very least puts price pressure on the incumbent.  Innovating faster along the original trajectory doesn't help the incumbent because customers are looking for solutions, not even more innovation that they can't absorb and don't want to pay for.  (Remember -- innovation is a supply side action.  The demand side is about solving a problem.) 

Geoff Moore also talks about standards from a technology point of view.  He observed that when a gorilla is forming in a market there is a call for "open systems".  He's right.  But the timing is important here.  If you call for standards before the customer is over served, the incumbent still delivers a better experience and wins.  (To be clear: I like Christensen and Moore models a lot because they reflect how I've seen businesses and customers behave as both technology provider and consumer.)

This also means the vendor that is being "standardized" is going to "fight" that standards effort with a vengeance.  (More on that in a moment.)  So everyone is behaving economically rationally.  Microsoft hates what's happening and IBM, Sun, et al love it.

Let's look at standards for a moment.  Standards exist to encourage multiple implementations.  That is their economic purpose.  It is a specification of how something works at an interface such that multiple providers can provide the functionality.  I wrote the following in a soon to be published O'Reilly essay:

*****
All standards live within a context of development and use. Many formal standards are developed by national bodies or international organizations such as ISO. These standards often define procurement policy for government organizations and large enterprises alike.  Industry and trade associations develop standards relevant to their expert and specialized constituencies. In the information technology space for example, the IEEE has a standards arm, and historically CBEMA (now NCITS) and Ecma International acted as standards development organizations in the USA and Europe respectively for IT standards. Each of these three organizations was accredited within their national and regional geographies to produce standards that could be later adopted by the relevant nationally or internationally sponsored standards organization to prevent overlapping efforts, and to build on the relevant expertise within different industry groups.

Narrowing the focus even further, consortia of vendors [e.g. W3C and OASIS] often arise within a specific area of technology within an industry to develop standards and specifications. The consortia often try to build specifications more quickly to expand a particular market, feeling that the more traditional organizations are too slow to deliver standards
....

Another categorization attempts to discuss the difference between de jure standards developed in a consensus-based process and de facto standards. A more accurate statement might be de facto technology that describes a market dominant product, rather than a specification for interoperability open to all implementers.  [One can argue that Christensen's model talks about when the market pressure turns de facto technologies into de jure specifications.]

*****

Proprietary specifications of de facto technologies are NOT standards.  They don't exist to encourage multiple implementations of the technology for the consumer, but rather to encourage multiple complements to the technology, therefore increasing the value proposition of the de facto technology for the vendor. They serve a very different economic purpose.  All of Microsoft's claims about being "the standard" are the antithesis of the standardization efforts the Commonwealth want to support. 

If a company is not drowning in their own brand or innovation Kool-Aid™, then they will play the standards game well. They can't arguable stop the standards process, so it is better to get involved and manage the process. 

This is certainly what Digital Equipment Corp. did through the 1980s and into the 1990s.  The POSIX and UNIX standards efforts weren't about UNIX per se.  DEC ruled the minicomputer market.  VAX/VMS was an incredible engineering friendly platform for rich application development.  Nothing wasn't documented for the platform so as to enable one to best develop new VMS applications.  But, they were innovating faster than customers could absorb the innovation, and trying to drive customers up the product curve.  They were also getting protectionist -- you couldn't add less expensive third party disks or memory to your VAX clusters without voiding your warranty.  Customers wanted off the upgrade carousel.  Applications (and ultimately data) are what's sticky for the customer.  So if we could standardize on a portable application migration path, then even if portability wasn't perfect, it would have to be better than being locked onto the DEC platform.  (Sound familiar?)

DEC led the POSIX standards efforts.  In my opinion, they brilliantly applied Brooks' Law to the process.  If a little standardization is good, then a lot must be better!  Soon there were 350 people attending quarterly week long meetings, and a couple of dozen projects spread across a dozen working groups.  There were co-ordination committees, and liaison points to countless other standards organizations and consortia around the world.  You don't even have to be that deliberate.  The best companies at the standards game simply invite/involve people with the right attributes.  There are certainly other games to be played as well.  The POSIX efforts went on for over a decade.  Microsoft should have perhaps played the OASIS standards effort differently. 

In the end, as customers, we chose the standards path.  We were incredibly happy with the VMS platform, but not DEC's lock on us.  We chose inferior technology (UNIX products were pretty raw in the late 1980s) and rewrote new applications for it, and began migrating existing applications.  We could get UNIX systems from a wide variety of vendors, and we certainly beat DEC down in pricing discussions in the mean time.  UNIX systems matured over time to be as robust and innovative as VMS was in its day.  We were a commercial organization.  We were following the lead of the U.S. government in its procurement policy.  The customer that was front and center in the POSIX standardization efforts was the U.S. National Institute of Standards and Technology.  (This was back before the head of NIST became a presidential appointment, and the NIST agenda became political.  In those days, NIST dealt with boring tedious initiatives like how to the government should best procure technology to protect the long term investment of tax dollars.)

The cost of migration is always painful the first time.  But then the pain is eaten and you're done.  For a period of time, in Microsoft's "Get the Facts" campaign they had a collection of public case studies demonstrating that the Total Cost of Ownership was cheaper for Windows, but these case studies invariably ignored the cost of migration to Windows in the numbers (even when they mentioned them in the executive summary.)  Amazing how the migration costs got ignored when the shoe was on the other foot. 

While Microsoft would love us to believe that this is an "unproven" standard, document format standards are actually amazingly mature at this point in history.  You have to set the optic right.  We have lots of experience with document formatting standards, all the way back to SGML (somewhat before its time), to HTML for the web, to XML. Even the early SGML work was based on proprietary mark-up and type setting languages from IBM and DEC, and early UNIX systems. We even have the experience collectively in the industry to deal with the differences between page layout (PostScript) and the structure of the information.  Like networking protocols, document formats either work or they don't and it's very direct evidence.  With the first standard in place, the space will mature to accomodate new document objects, and the product space will mature as well.

So in the end, Microsoft's commentary and framing reflect their vendor-centric, market-position threatened point of view.  And Sun's and IBM's positions are certainly to be expected.  But then so is the Commonwealth's.  The Commonwealth is making a tough choice, but ultimately the right choice for its constituents.  The rest is just framing.

Additional Links:

• Nicolas Carr hits the nail on the head in this blog post (19-Sep-2005).  (Thanks to Stephen O'Grady for the pointer.)  To speak to Mr. Carr's concern about reading state documents, I would suggest that's why the Commonwealth is also supporting the ubiquitous PDF (which is an ISO standard and freely licensed and widely implemented by multiple sources).  One can use ODF as the internal working format, and PDF as the external publishing format. 
• Boston Globe article (14-Sep-2005) "Microsoft fights bid to drop Office Software". This is the article that provoked this response. 
• My original roll-up post is here.

 

Postscript: For an excellent discussion on norm entrepreneurial rhetoric, I would encourage you to read David McGowan's "SCO What: Rhetoric, Law, and the Future of F/OSS Production". While the paper is a little long and free and open source software centric, the discussion is directly relevant.  [It gets a little legalistic in the middle, but read the rest of it aloud to appreciate the use of language.]

On Why I Left Microsoft

Updated (17-Sep-2005, 11:21 PDT): Added the second postscript pointing to the Business Week article. 
Updated (14-Sep-2005, 22:36 PDT): Added the postscript at the end. 
Updated (12-Sep-2005, 18:00 PDT): Added RobMen's blog reference and corrected a grammatical ambiguity.

Eric Raymond recently posted on a recruiting attempt on him from Microsoft and towards the end of his post mentioned:

FURTHER UPDATE: I had my serious, constructive conversation with Microsoft last year, when a midlevel exec named Steven Walli took me out to dinner at OSCON 2004 and asked, in so many words, “How can we not be evil?” And I told him — open up your file formats (including Word and multimedia), support open technical standards instead of sabotaging them, license your patents under royalty-free, paperwork-free terms.

I believe Steve Walli went back to his bosses and told them that truth. He is no longer with Microsoft, and what little he’ll say about it hints that they canned him for trying to change their culture.

It was a great dinner.  Eric was there, along with Larry Augustin, Pamela Roussos and Andi Gutmans from Zend, Rob Mensching (Microsoft — runs the WiX community), and my then partner in crime Jon Rosenberg.  I believe it might have actually been Rob that asked how Microsoft could be less evil to Eric.  Those that know me well know that's not really a "me" sort of question.  I've always felt actions were important and was trying to get Microsoft engaged in the broader open source community beyond strict sharing programs under the Shared Source banner.  I still believe that Microsoft has a huge opportunity here. 

They didn't fire me for trying to change the culture.  I've met Eric and others in the free and open source software community since leaving early in December 2004, and friends have heard the anecdotal examples of things that were deeply frustrating me. I left completely of my own accord. 

Microsoft is a Very Large Company.  It is wrestling (as all large companies do) with the very processes that enabled it to succeed and grow.  It is at a point in it's life where it needs to be defined by operational excellence as much as product development.  The single scarcest resource is (as one friend pointed out five years ago) executive bandwidth.  It takes forever to get a meeting with the very execs that can open the doors for success through their sponsorship.  And once one of those meetings is scheduled, you enter data preparation and PowerPoint hell.  For me, it had become not very fun. I simply don't have the patience of people like Danese Cooper, Simon Phipps, Bill Hilf or Bob Sutor.  When interviewing with companies like Microsoft, you will often hear about the incredible reach you'll have in the world.  That is of course predicated on the ability to be heard inside the company first, and on the ability to actually ship something.

As a vice-president of R&D at Softway, I worked with a team that delivered 11 releases of software (Interix) over our four years.  As the product unit manager for the same team and technology, I left the team two years after entering Microsoft because we still hadn't shipped technology that was basically almost ready when we were acquired.  We were essentially waiting on integration with a ship vehicle.  Tightly integrated innovation isn't always the right answer. (Look at the length of time it's taken to get the new Windows command line shell out the door.)  There was a recent article in the Seattle Post-Intelligencer on Microsoft warming up to open source.  The data points were the same three small open source projects I helped launch last year. I also have a background in industry standards, and as a gross generalization, most people at Microsoft don't care about the difference between de facto technologies and industry developed standards, and the subtleties required to engage customers while confusing competitors with standards, so I was only frustrating myself in those discussions as well. 

To be clear.  I thought I was fairly treated in the acquisition (despite the fact that Softway had essentially hit the cash wall as a startup).  I was well compensated as an employee.  The benefits were great (but that may be shifting a little as they need to deal with the operational excellence mentioned above).  I had sufficient seniority, credibility and visibility to be able to request short (~10 minute) meetings with a number of senior execs on a number of occasions.  (It was larger reviews where decisions need to be made across parts of the organization that are impossible.) Three of the best managers I've ever worked for in 25 years were at Microsoft, including my immediate manager in my last assignment.  I learned an incredible amount about business, but then I had a different set of radical business lessons running my part of Softway, so part of that may simply be a pre-disposition to learning.  I haven't met anyone outright evil, (and I've been in Bill and Steve reviews on a number of occasions).  Only a couple of people were sort of broken with respect to their thinking, and they certainly weren't execs.  I still have the utmost respect for the company and its executive team.  If you looked at my annual poll results, you would see a "happy" employee. 

I just wanted to build things and have fun again.  I believe free and open source software are incredible business tools, and a lot of fun.  I left Microsoft to work for myself after a little over five years.  I've done this twice before in my career.  I left EDS in 1991 and in 1994 I left Mortice Kern Systems (MKS).  The latter lead to the founding of Softway. I'm one of those people that would rather rock the world of 40,000 customers, than touch the lives of 40 million.  I like small growing companies. 

Early this year, Optaros convinced me I should help build their company, and it's been a great experience ever since.  Life's too short.

Postscript: I was going through my back log of reading this evening and came across Hugh MacLeod's "How to be Creative" again.  I appreciate everyone on the planet that reads blogs or themself blogs has probably come across MacLeod by now, but just in case you haven't, I've included the link.  Re-reading parts of it again summarizes so much of the way I try (and sometimes even succeed) to create over the past 15 years.  (Yes, I've known Jeff Haemer that long, and he continues to push and encourage me.)  MacLeod just captures the ideas and the work so perfectly.  If you haven't discovered "How to be Creative" — please do.

P.P.S. Business Week's cover story this week is on the Microsoft exodus.  (The article is also the start of Business Week's podcasting initiative around their cover stories.)  I have a minor quote in the article.  In discussions with friends about the article, it was observed that the type of people leaving now are the sort that question everything and aren't willing to take "no" for an answer.  Essentially, these are the sort of people that made the company what it is, and exactly the sort that can't be happy in what it's become.  This doesn't come out in the article, but explains the dichotomy in the article's view points. 

Microsoft Services for UNIX Requiem

“Just say the report of my death has been grossly exaggerated.” — Mark Twain (from Bartlett's on Bartleby)

A couple of reports came out last week on the end of Microsoft's Services for UNIX product line. 

• 1 Sept., Mary Jo Foley "It's the End of the Line for Microsoft's Services for UNIX Product" on Microsoft Watch (also on eWeek). 
• 1 Sept., Ina Fried, "Microsoft Makes UNIX Changes", on CNet.

Each makes it sound like Microsoft is wandering away from the product, emphasizing that it will hit end of life support in 2011, and with a bit of a hand wave towards some integration. 

I have a rather personal stake in this one.  In the Spring of 1995, Doug Miller and I started planning a company that would become Softway Systems and deliver the Interix UNIX-to-NT application migration technology until the acquisition by Microsoft in September 1999.  We were immediately wired into the Services for UNIX space, which had traditionally been purely an interoperability product (client and server NFS, NIS, etc.).  So I went hunting around old contacts, which lead me to Rik Wright who handles product marketing for SFU.  Here is the official statement from him:

Several platform level components of Services for UNIX will be available in Windows Server 2003 R2, but there are also components of Services for UNIX, such as NFS Gateway, that are not included in Windows Server 2003 R2.  Microsoft is responding to customer requests that some of the functionality of Services for UNIX be more tightly integrated with the Windows Server operating system.

The next generations of NFS client, NFS Server, User/Name Mapping, Telnet Server & Client, Password Sync and NIS Server components of Services for UNIX are all present in the Windows Server 2003 R2 builds. In addition, for UNIX application portability, we provide the Subsystem for UNIX-based applications which allows re-compiling of UNIX-based applications on Windows, including 64-bit support, database connectivity libraries and mixed-mode processing that is not available in SFU

The Subsystem for UNIX-based applications offers the same functionality as the Interix System in addition to 64-bit support and database connectivity enhancements. Microsoft has continued developing the Interix subsystem since we acquired it from Softway Systems, but this release is based on a completely new architecture that allows for considerable new functionality such as mixed-mode processing. The Subsystem for UNIX-based applications is backwards compatible with some Interix applications but may need recompiling of source for others.

Q. How will current SFU 3.5 customers adopt this new SFU technology?
A. Customers will have to upgrade to Windows Server 2003 R2 to take advantage of new functionality.

Q. Is the SFU functionality available in all editions of Windows Server 2003 R2?
A. Yes. This functionality is available in all editions.

Q. Will there be SFU version 4?
A. No. There are no plans for future out-of-band releases of Services for UNIX. Services for UNIX 3.5 will be supported on Windows Server 2003, SP1, Windows 2000 and Windows XP until 2011 with extended support offered until 2014.

So what's happening is that the primary parts of the product that would have been the next version are being included in the next release of the OS.   When I asked Rik where the (Interix) utilities and SDK would be available, he pointed me to here, which is similar to the free download site for SFU 3.5.

One day I'll blog the history of Softway and Interix.  For now, I'm simply happy to report that "it's not dead yet."

MORTICIAN: Bring out your dead!
      Bring out your dead!
      [clang] Bring out your dead!
      [clang] Bring out your dead!
      [clang] Bring out your dead!
      [clang] Bring out your dead!
CUSTOMER: Here's one -- nine pence.
DEAD PERSON: I'm not dead!
MORTICIAN: What?
CUSTOMER: Nothing -- here's your nine pence.
DEAD PERSON: I'm not dead!
MORTICIAN: Here -- he says he's not dead!
CUSTOMER: Yes, he is.
DEAD PERSON: I'm not!
MORTICIAN: He isn't.
CUSTOMER: Well, he will be soon, he's very ill.
DEAD PERSON: I'm getting better!
...
(Courtesy of http://www.rit.edu/~smo4215/monty.htm)

Innovation and Customers, Patents and IP Strategy

I recently posted on Massachusetts's declaration to move to open document formats for government documents. Matt Asay raises several issues (innovation and customers, my possible confusion of patents and innovation, me being anti-Microsoft).  One I've corrected inline, by bringing it here because he was correct that it really didn't fit the other post.  The other two I've brought here to hopefully clarify.

Matt calls me out on the fact that I said:
Customers don't buy innovation.  (And customers couldn't care less about nor do they buy patents.)  They buy solutions to problems.

He says (by calling out Jason Matusow's discussion on his blog):
... That Microsoft is an innovator (sometimes measured in patents and copyright), and customers should buy from it because it will consistently deliver cool, useful (i.e., innovative) new things worth buying.

Here's where I realize Matt and I are likely agreeing, but confused in each other's terminology.  He is equating innovative with useful.  I absolutely think consumers buy things because they're cool and useful.  That's the demand side act.  Innovation (to me) is the act of design and creation.  The dictionary defines it as a noun meaning "the act of introducing something new", or "something newly introduced."  It's a supply side act.  And it is different from the buyer's motivation. 

As Matt later points out in a subsequent post:
It's not that we wake up one morning and say, "My, I sure need some innovation today! Where can I buy some?" Rather, we buy products that catch our eye, that deliver results, that give us a performance or feature edge. All of this stems from innovation.

Exactly so. 

But I am not confusing innovation and patents.  (I think he's merely goading me in that statement to produce the following bit).  Intellectual property (IP) is a collection of legal tools used to protect assets (i.e. innovation), consisting of copyright, patent, trademark, and trade secret law. You need to create first (i.e. innovate in Matt's terms) before you patent. 

Patents are government-granted time-limited mini-monopolies to enable the creator to bring their work to market protected in such a way that they recover their investment costs, but published in such a way that others will be able to build eventually upon the work. But patents cost money to obtain and defend and maintain.  They are merely tickets to a negotiation as I've observed. Useful?  Depends.  Patenting is merely one choice in relation to how you use the asset created.   

From a business perspective, I would rather have a growing bed of customers, and out run and out deliver the competition, possibly even publishing like crazy to prevent those competitors from patenting in the space.  Takes less time.  Costs less money.  And if I'm a small company, having a little stack of a few patents, even brilliant essential patents, won't help me going up against a cross-license negotiation with any of the big guns (IBM, Microsoft, Intel, etc.).  In software, the speed and cost of delivery of innovation makes patenting a possibly suboptimal use of my time and money.  Indeed, I could do everything "right", researching patents, possibly applying for my own, and with the 18-24 month gap between application and award for a patent, still discover I'm infringing months after I've shipped my product to market.  I'm not the only one thinking this way.  Here's what Joi Ito had to say recently. 

Lastly, I'm not an ardent anti-Microsoft crusader.  (Yes, I was on an airplane at the time he wrote that bit.)  I spent the first decade of my career from '80 through '91 on the customer side of the fence in large IT organizations in retail and manufacturing.  I have always been unsympathetic towards any vendor that forgets that it's the customer's money, and they have to earn it.  The vendor has to care about helping me solve my problem.  Vendors aren't universally "good" or "bad" at this.  Microsoft ends up in my sights a little frequently only because of a certain depth of experience inside the walls. 

Pax.

Massachusetts, Microsoft, and Open Document Format Standards

Updated (19-Sep-2005, 15:00 PDT): Moved the discussion to a new blog entry here.  This first entry was very much a roll-up and initial thoughts on the announcement.  Now we're moving into "Phase II" and the framing debates.  I felt it needed additional context in a new entry. 
Updated (13-Sep-2005, 16:17 PDT): Added more blog and press entries to the end.
Updated (12-Sep-2005, 00:25 PDT): Added a number of other blog and press references. Cleaned up an opinion that Matt Asay pointed out was confusing. I also (hopefully) clarify that opinion here.

A lot has been reported and blogged about the recent announcement from the Commonwealth of Massachusetts concerning a move towards the recently ratified OASIS OpenDocument format. 

Essentially, the Commonwealth published their desire to move to freely available standards and specifications, choose the standards they want to use, and have requested industry comments back.  As they chose OpenDocument and Adobe's PDF, everyone is looking at this as a slap in the face to Microsoft.

There is lots of debate and commentary.  Microsoft has said it won't support OpenDocument in the next release of Microsoft Office.  They've tried to point out that OpenDocument is inferior as a format, that the Microsoft Office XML are freely licenseable, and that the next release is about compatibility with older Office formats. 

Stephen O'Grady (RedMonk) has an excellent commentary based on an interview covering the ground with Allan Yates of Microsoft.  Stephen also responds to Matt Asay's commentary.  I have to agree with Stephen on this one.  David Wheeler also has excellent commentary here. 

Microsoft's commentary is fairly predictable.  Mike Champion blog debates with Stephen here, and Brian Jones has comments on Aug 31 and September 5.  Stephen O'Grady's responses are all great. 

Brian Jones expresses surprise and confusion based on the royalty free patent license Microsoft offers on their Office XML schema.  Of course, while the patent license for the Microsoft Office XML formats is royalty free, the license is not sublicensable (preventing a free software GPL licensed implementation).  (The Groklaw commentary gets into this issue as well as David Wheeler's commentary.)

Observations:

• I first commented in May of this year that the threat to Microsoft's hegemony wasn't the Linux desktop but rather OpenOffice on Windows, that Microsoft would try to battle it by claiming superior technology and innovation, and the role the OASIS standard would play.  Clearly the Commonwealth of Massachusetts is moving in this direction, but so it appears is the City of Munich.  While they still try to struggle with their Linux desktop decision, "Some departments will start with OpenOffice on Windows, ...."  So it begins.
• Microsoft appears to remain confused about standards and patents.  Standards exist by definition to encourage multiple implementations of a technology.  This is regardless of who defines them: whether government sponsored and recognized organizations, industry associations, or even more narrowly focused technology consortia.  Patents exist to protect a single implementation.  By definition standards and patents exist to serve different functions in the economic spectrum. A patent is a tax on the standard.  No participant wants such a tax, or perhaps a better way to say this is that any participant that has a patent that reads on a standard needs to determine whether or not the tax will slow the adoption of the standard, and whether or not the economic goals served by the standard are more valuable to the patent holder. Yes I appreciate the Microsoft patent license is royalty free, but it is the Microsoft point of view that these assets still need to be strictly protected that is the problem. 
• The patent license is absolutely designed to prevent GPL implementations.  Before attributing malice here, consider the following.  If a patent is allowed to be sublicensed (enabling a GPL implementation), then when the software that implements the patent is used downstream for a different purpose, and the GPL attaches to that software (and so on), then the scope of the patent can "creep", and the patent holder can loose control of the patent and the ability to license it differently (possibly for a royalty) or prevent others using the invention in different circumstances where they might want or need to be less generous with their protected asset.  If you were a very conservative lawyer (and working for a large legal target like Microsoft appears to engender that in their legal team), then this loss of patent control is a scary thing.  I have no idea if lawyers can develop a license that would enable GPL implementations but scope them to strictly the royalty free use desired.  (IANAL.)  The GPL would actually support the downstream scoping if I am guessing correctly.  You see, the GPL says you need all the rights to distribute the software, or cease distribution.  If the patent rights you hold are strictly to implement the standard, then you can implement the standard, but if you try to use the GPL licensed software for any other purpose, the GPL prevents it. 
• I find it ironic that when Microsoft has been confronted in the past with governments trying to develop open source software preference policies, they (rightly in my opinion) argue that the government shouldn't care how the software was developed and licensed, but rather focus procurement policy on software acquisition against standards.  Now that they have a government behaving exactly as they have requested, they are claiming that governments should focus procurement policy only on standards that Microsoft already supports, or de facto technologies that they sell, i.e. the status quo. 
• Microsoft is still confused about why customers buy things. Customers don't buy innovation.  They buy solutions to problems.  The government has a problem, and determined a solution, and Microsoft could support that solution, but would rather try to tell the customer that they don't understand their own problem space, and should keep giving Microsoft money for innovations, regardless of the problems the customer wants to solve.  It's the "we know best what's right for you" defense, and Microsoft still doesn't appreciate how arrogant it sounds. 
• I pretended to be Novell the other day.  I'll pretend to be Microsoft now.  If I was Microsoft, I would publicly state that we had reconsidered the customer's position (after meeting with them), and that while we have confidence in the integrated innovation (blah blah blah ... fill in marketspeak here about great new Office features... blah blah blah), they will support ODF.  They may not be able to add it in the current time table for the next release of Office, and delaying the release for the feature would be unfriendly to their shareholders.  But it could certainly be made available shortly after in a service pack.  And then deliver it.  If Microsoft is genuinely delivering huge value for other features that solve problems, they will still out sell all the competition.  They'll appear responsive and responsible.  And the discussion will move on.

I commend the Commonwealth on its move.  It has some short term pain, but should show a lot of benefits in the long term.  This is how I would want my government to spend my tax dollars — take the long view, and do the hard work to serve my needs best. 

The reporting trail and additional blogs:

• 1 Sept., CNet's Martin LaMonica's "Massachusetts to adopt open desktop" on ZDnet.  Covers the initial announcement from the Commonwealth.
• 2 Sept., Paula Rooney "Microsoft Blasts Massachusetts' XML Policy" on Information Week.  The first reporting back from the Microsoft point-of-view.
• 2 Sept., Boston Globe coverage "State May Drop Office Software".  The local coverage. 
• 5 Sept., Tom Espiner, "Legal Worries led Massachusetts to open standards" on ZDnet.  The Commonwealth responds back to the Microsoft news cycle.
• 8 Sept., David Coursey "Massachusetts' Move to Open Format is Closed-Minded" on eWeek.  A few things he gets very incorrect.  Some good opinions however when he points out both that Microsoft should support ODF as well as open their own formats further.
• 8 Sept., Simon Phipps (Sun) blog commentary on Coursey. Groklaw added some further comments as well.
• 8 Sept., Steven J. Vaughan-Nicols, "Microsoft's Standards Are No Standard at All", on eWeek.  Debates his co-worker Coursey's comments. 
• 9 Sept., Tim Bray (Sun) blog publishes Scott McNeally's letter to Peter Quinn.
• 10 Sept., Tim Bray (Sun) blog commentary "Massachusetts Back-Room" Groklaw additions here too.
• 10 Sept., More blog commentary from Microsoft's Mike Champion on the Microsoft XML blog.  To quote Mr. Phipps (from whence I found the link): "My, but he's prolific and subtle."
• 12 Sept., Excellent blog follow-up from Sun's Mr. Phipps on the framing game that's being played out here. 
• 12 Sept., Jim Wagner's "Open Season in Massachusetts" on InternetNews.
• Just catching up with blog commentary from Bob Sutor (IBM, V.P. Standards and Open Source) who is obviously supportive:
• 13-Sep-2005, Myths of Technical Neutrality, excellent bit with a peek at how vendors frame things. Nice bookend for Simon Phipps' bit on framing. 
• 8-Sep-2005, Official IBM response to the Massachusetts Enterprise Technical Reference Model v.3.5 - Public Review Draft
• 6-Sep-2005, Massachusetts Entry of the Day discussing David Berlind's ZDnet article.
• 1-Sep-2005, More Links on the Decision.  Covers off a number of links I haven't yet linked. 

Mono, Microsoft, and Mischief

Update (17-Sep-2005, 19:00 PDT): Well the party did indeed happen, t-shirts and all!  I'm sorry I wasn't in L.A. for it.  Todd Bishop from the Seattle Post Intelligencer was covering the PDC and talked with Miguel.  He talks about mono in an article and on his blog.  Hopefully Miguel posts pictures soon.

Apparently Microsoft is again denying Miguel de Icaza a Mono BoF at this year's PDC. This is such a missed opportunity for Microsoft.  Microsoft needs Mono.  They published the CLR and C# language specifications as international standards through ECMA and ISO, but it still looks like a proprietary controlled technology to the customer base.  Standards are a message in the marketplace that encourage multiple implementations.  As long as there is only one commercial implementation, there is no standard, and the customers know it regardless of the number of trees killed to produce specifications.  Encourage Mono and then deliver the better integrated experience for writing web services on .NET.  Of course that would require the ability to ship software in a timely fashion. 

Microsoft is the company that re-branded the ".NET Server" to "Windows Server 2003", and ".NET" to "Windows .NET" to ensure they really drove home the idea that it's all about Windows.  At this point in their existence, they already own that word in customers' minds, and unfortunately customers know that really means their "desktop".  There was the opportunity to own a new word — ".NET" — and even associate it with the idea of servers, but that's gone now.  It's "Windows .NET".  IBM learned a long time ago to own new words.  They have a history of owning new words and moving along with their customers.  There was "mainframe" and "PC" and even "Selectric" at one glorious time.  Digital Equipment Corporation on the other hand ended up only owning one word.  Too bad. 

If I was Novell, I would start a nasty viral ad campaign for the PDC.  Instead of gathering in a hallway like stubborn refugees, get a meeting room in a neighbouring hotel, or take over a bar like a Dick's Last Resort.  Pass out invitation cards and put out notices in the lobbies of all the local hotels for the PDC, and on the message boards at the conference.  Set-up the screen and one of those nifty portable projectors, plug in the laptop, and let Miguel do what Miguel does so well.  Buy the beer and the food.  Hand everyone a copy of the latest SuSE distribution, and whatever Mono tools can be put together.  Hand out t-shirts to the first 100 people through the door and make them GREAT t-shirts.  (I'm thinking a brick wall background with the Mono monkey logo "spray painted" on the wall in a loud fluorescent blue or green.  No words.  No other logos.  We already know what the Novell logo looks like and it would spoil the effect.)  Encourage everyone to wear them to the PDC the next day.  Remind Microsoft that you're customers — and that it's your money.  Make the Mono BoF the best little party of the PDC.  It certainly couldn't cost as much as a booth in the vendor hall. 

Of course that's just what I'd do.