April 2008

Everything I Said Last Month was Wrong, and I'm a Lousy Prophet, Too

In this month's newsletter, I'd like to reassess the conclusions of last month's newsletter in light of some interesting interactions I had with analyst firms this month as well as the feedback I received about last month's newsletter from readers. I'd also like to examine my prediction made in December that Microsoft would fail to achieve acceptance of its OOXML specification by ISO and offer some comments on the implications of its (as of this writing) expected success in gaining ISO adoption.

Everything I Said Last Month was Wrong

You may remember that last month I commented on a study published by IDC that concluded that enterprises implementing open source software are performing the implementation themselves; this is in contrast to the heavy use of system integrators and consulting firms used for mainstream software implementations. I discussed this seeming paradox and concluded that, no matter what the reason that caused this paradox, it probably reflected the fact that true enterprise adoption of open source is still in its infancy – as more enterprises implement open source, more integrators will be used.

I had my eyes opened about analyst firms this month. Even though I had a rather jaundiced view of them in the past, I had a couple of interactions that really illustrated the shallowness of their approach and the undoubted unreliability of their research, if it actually can be spoken of with that term.

First up, I got an email from one firm wanting to research an open source topic. The email was something along the lines of “I'm interested in learning more about XXX. I'd like to interview you for 45 minutes to an hour and would like to talk to two of your customers. Please let me know when would be convenient for you to speak.” I found the assumption that I would offer a bunch of my time for free and would impose on clients all in the service of the analyst firm's business incredibly presumptuous and so naturally, though politely, declined.

The second interaction was an actual conversation. An open source company asked me to speak with another analyst firm to discuss the potential for their application from my perspective – in other words, to discuss my opinion about how their application might be used in the marketplace. I agreed and arranged a time to speak to the analyst. She called me and immediately asked me to describe my firm's business and then wanted to know how many engagements it had done in the previous year. When I demurred to discuss my business – remember, the purpose of the call was to talk about the potential for an application released by another company – the analyst got very huffy and rather frostily ended the conversation shortly thereafter. And, by the way, the insights offered by the analyst during our conversation were, to my mind, extremely puerile and unsophisticated.

It's fascinating, however, that commercial open source companies, despite their brave insistence that they are upending the bloated and irresponsible proprietary software industry, still kowtow to the traditional practices of the industry, including focusing on currying favor with analyst firms in the hope of gaining a coveted endorsement or at least recognition. This despite what last month's newsletter discussed: most open source software enters enterprises in a very different fashion than the historic patterns of proprietary products. That's why the traditional handmaidens of proprietary software, the big system integrators, have heretofore played no role in the spread of open source.

Nevertheless, notwithstanding the lack of insight reflective of the industry's practices, the IDC study referenced in last month's newsletter, and the discussion regarding it, struck a chord with readers. I received more comment on last month's newsletter than any other in its three-year history. And most of them agreed with the conclusion that enterprises are, for the most part, early in the open source adoption game. So, despite everything, it seems that IDC might have stumbled into the truth when it concluded that enterprises are doing most open source implementation internally and have not yet begun to focus on using system integrators.

My Crystal Ball is Cloudy

My December newsletter confidently predicted that OOXML, Microsoft's initiative to get its proprietary document format accepted as an ISO international standard, would fail. After all, with the Microsoft activities exposed as sailing very near the wind, and the further identification of technical shortcomings in the specification itself, it seemed to me that the March vote would be likely to be even less positive regarding an OOXML adoption.

It seems, though, that Microsoft is going to be able to squeeze out a victory; a tribute, I guess, to backroom deals and cupidity.

One must ask: what will Microsoft win? This question was brought to my mind at last week's OSBC conference where its head lawyer, Brad Smith, came and spoke in a keynote panel. A question posed to Smith several times regarding Microsoft's repeated insistence on forcing licensing upon companies wishing to interoperate with Microsoft products (remember, Microsoft characterizes this patent-based licensing as “building bridges to the open source community”) was this: “The actual revenue Microsoft attains through this licensing activity is very small in comparison to its overall revenues – so why does Microsoft bother?” Of course, when one considers the top management attention used in this low payoff licensing activity, the financial pointlessness of pursuing it becomes even clearer.

Smith's response, stated several times, was along the lines of “We believe our intellectual property has value and we should be compensated for its use.” Most observers regarded this rejoinder as a smokescreen, a cover up for some imagined even-more-nefarious Microsoft motive. I took a different view. I think Smith was fairly presenting Microsoft's position: It created the IP and if other people want to interoperate with Microsoft's products, they should pay for the privilege – it's only right. This attitude is so embedded in Microsoft's zeitgeist that it is willing to pursue it well beyond the bounds of economic rationality; in defense of the principle, Microsoft is willing to spend a dollar to gain a penny.

However, as I later observed to someone, that position only holds water when the licensor has the upper hand – when the licensor's products hold the dominant position and other products must interoperate if they wish to gain customer acceptance. If there is a critical mass of those other products, all of a sudden the shoe is on the other foot, so to speak: it would be in Microsoft's interests to get its products able to interoperate with the others; in fact, failing to offer good interoperability with those products would handicap the acceptance and adoption of Microsoft's products.

That got me to thinking about what the (expected) victory of OOXML would mean. It's true that Microsoft dominates the desktop office market and therefore dominates the resulting document formats. When I look at the OOXML votes, they seem to break down along a developed/developing nation split, with the former falling into line with Microsoft's desires (to be expected for sure, since there is more money at stake in a developed market and therefore more cupidity on tap) and the latter hewing to rejecting OOXML.

My speculation as to the motive for the developing nations taking this position is that they are adopting it for sovereignty reasons. Simply put, they don't want their computing future dictated (and that is what OOXML is, notwithstanding the ISO standard fig leaf) by an American company.

These nations will insist on using ODF. And, while the developed world currently represents the vast majority of computing use, the future is all with the developing world. Which calls into question the assumption that if the developed world uses OOXML, it will be forced onto everyone else. Remember, the power of commercial relationship lies with whoever represents the majority of the market, as the licensing discussion noted. If a couple of billion people in the developing world insist on using ODF, everyone else is likely to find it incumbent upon them to find a way to communicate with them. So Microsoft's OOXML initiative may end up, in the long run, looking like a Pyrrhic victory. We'll see. My crystal ball may just clear up some day.

Navica News

Speaking Engagements
May 1, 2008: Open Source Governance Workshop, Boston, MA. Contact Navica if interested in attending.

If you are interested in having me speak at your organization, contact me directly via email.