Kicking it up a notch: Microsoft’s open source message gains subtlety

50

Author: Bruce Byfield

VANCOUVER, B.C. — Barnaby Jeans, an IT Pro Advisor at Microsoft Canada, has a new approach to
Microsoft advocacy. In the past, Microsoft executives have been reported as
describing GNU/Linux as a
cancer
or communism.
Other reports include warnings that corporate users risk
being sued,
and that the “Pac-Man-like nature
of its licenses may make interoperability between applications harder.
The Get the Facts
Web site continues these direct attacks, although
with greater detail. However, Jeans has a subtler message for corporate customers.
In fact, his message is less in what he says than how he says it.

I attended Jean’s presentation, “Windows in an Open Source World” at the
TechVibes Massive 2005 event on March 30 in Vancouver, and interviewed him afterward. His approach
does contain traces of older Microsoft messages, including references to the Get the Fact site.
Overall, though, he gives an impression of not being
against anything at all and of being interested only in open exploration of the
differences between Microsoft and open source software. Emphasizing the changes in
Microsoft’s products and policies, he uses several basic rhetorical devices, relying
heavily on his own knowledge of open source and his presentation of himself as a
reasonable man. The result is a much more sophisticated and effective defense of Microsoft than
consumers usually see.


The background of an evangelist

Jeans joined Microsoft’s Developer Platform Evangelism team — the IT Pro
Advisors, as they are known externally — five months ago. Previously, he
spent seven years at Oracle. His last position before joining Microsoft was in
technical sales at Red Hat Canada. Jeans describes Red Hat “a great place
to work,” but joined Microsoft because he believed the company offered more
chances for career advancement. Besides, he welcomed the chance to
work directly with customers.

Jean’s role can be described as putting a human face on one of the world’s
largest corporations. Speaking at trade fairs, corporate seminars and user
groups, Jeans offers his listeners “a direct contact into Microsoft” that
includes passing along information and relaying feedback to product managers.
He also writes a blog
that helps him keep in contact with his audience. Referring to the entire
evangelism team, he stresses that “our role is not a sales role in any way.”

Although his audience often includes members of the Free and Open Source
Software communities, Jeans reports that his reception is usually
“incredibly good.” Most people, he says, are surprised that Microsoft would
make the effort to contact its customers directly.


Speaking from experience

Jeans believes that his experience at Red Hat gives him an advantage as a Microsoft
evangelist. “I have a level of credibility having worked on both sides,” he noted.
Although his claim to have been working with open source software “almost
since DOS 3.3” seems an obvious exaggeration, Jeans frequently mentions
the names of leading distributions and open source applications with an ease that
reinforces the impression of expertise. While he does not actually refer to
“GNU/Linux,” he does point out that “Linux” properly refers only to the kernel.

Jean’s familiarity with open source allows him to make arguments that convince
less because of their logic than because of their detail. For example, at the start
of his presentation, he divides
distributions into three categories: Commercial, Community-based, and Roll
Your Own. Warning that “not all distributions are created equal or even compatible”
without giving specifics, he goes on to say that most of the highly customized
distributions fall into the Roll Your Own category that most corporations lack
“the skill or the money to maintain.” The implication is that the advantages of
having access to the source code is overrated.

Similarly, soon after, Jeans presented a slide based on contributions to the Linux kernel during
September 2004 to prove that 12 developers are responsible for 44 percent of the
patches. Of these 12, he claims, 10 are employed by companies. Here, the
implication is that Linux is less community-based than most people imagine, and
that customizing Linux is beyond the reach of the average company.

In both cases, the implications are questionable. For instance, the bulk of corporate customers
for any operating system are probably small- to medium-sized businesses that need no
customization. In the same way, to include Linux Torvalds and Andrew Morton, both of
whom work for the Open Source Development Lab,
a non-profit consortium, as being employed by a commercial company seems to be stretching the
point.

However, in such cases, to focus on the arguments misses the point. What matters is not the structure of
the argument so much as the impression that it leaves. Because Jeans demonstrates a knowledge of his
subject, the implications of his argument are likely to be accepted without question, except by audience
members who already know something about the subject. The rest seem to nod in acceptance. Moreover,
because these arguments are made at the start of the presentation, when Jeans claims to be simply defining
terminology, they set the perspective for rest of the presentation. The fact that the
implications are left for each member of the audience to make for themselves only makes them more credible.


Granting limited validity to the opposition

Jeans also draws on the time-honored debating tactic of acknowledging that opposing views contain
some truth before going on to refute them.

Yes, Jeans admits, NT Server and Windows 2000 had security problems. So, too, in the past, did
Windows XP and Internet Explorer. It might even be true, he adds, that the present generation of
Microsoft products could be improved. He goes on to
describe Windows 2003 and Windows XP Service Pack 2 as major improvements on past products.
“We love to have these discussions,” Jeans said after his presentation, “Because in a lot of cases we
actually find that some of the things that people bring to these discussions, some of the questions they
have, are based on misconceptions or old information.”

Jeans uses the same tactic in his discussion of openness. On the one hand, he downplays the importance of
having the source code and blurs the distinction between open standards and open source. However, on
the other hand, Jeans acknowledges early in his presentation that a trend toward openness exists in
modern software development. Microsoft, he admits, has not been a leader in openness, but he points to
the Shared Source Initiative,
in which governments and larger companies can view the source code of Microsoft products and the
Windows XML Installer open source project as a sign
that the company is changing. He does not exaggerate these changes; he acknowledges, for example,
that the Shared Source Initiative is a “transparency mechanism” more than an actual benefit, yet he
does suggest that they are proof that the differences between proprietary and open source development
models are narrowing over time. At any rate, he believes, “at the end of the day, what everyone is
looking out for is the best interest of their customers” so that the differences are largely irrelevant.

Asked directly, Jeans stops short of admitting that these changes are in response to any threat to
Microsoft. However, he does say that “Microsoft acknowledges that there is some competition out
there” and notes that “people think that open source is more accessible.” According to Jeans, one reason
for his own position is that “Microsoft has learned that community involvement is important.” He adds
that “this is something that Microsoft is still learning.”

Used by anyone, this tactic creates an impression of honesty and thoughtfulness. It also has the effect
of anticipating counter-arguments and refuting them beforehand. Opponents can still try to answer
the argument but are likely to appear small-minded and ungenerous by comparison. Used by the
representative of a large corporation — especially one like Microsoft, that people love to hate — the
tactic is probably even more effective than usual, simply because it is unexpected.


Understatement and an open-ended conclusion

Throughout his presentation, Jeans words his key points as questions. “Does having the source
code really help?” he asks at one point. At another, he asks, “How do you select a [Linux kernel]
patch?” Instead of hammering home his point with direct statements, he leaves these questions
unanswered. However the simple fact that he asks them tends to raise doubts.

Even at the end of his presentation, Jeans avoided any emphatic statement of his position. All he wants
he says, is for the audience to go away convinced of three points:

  • that Microsoft had shifted toward openness with the Shared Source Initiatives,
  • that any software must meet specific requirements to be useful, and
  • that any solution has costs that go beyond the purchase price, such as maintenance
    and training.

    None of these points sound controversial. When he mentioned them at the start of his presentation,
    nobody answered when he asked, “Does anyone have a fundamental problem with these statements?”
    Nor did anyone question them when they were repeated at the end. However, after the rest of the
    presentation, these points seemed to apply to open source software more than Microsoft products.
    Having accepted the points from the beginning, the non-open source people in the audience to whom
    I talked to also seemed predisposed to accept the implications attached to the questions by the end
    of the presentation. At the very least, they seemed responsible concerns raised by an expert who is
    genuinely interested in helping people.


    The importance of personality

    None of this discussion of tactics is meant to present Barnaby Jeans as a diabolical shill. In fact, nothing
    seems farther from the truth. Jeans comes across as an intelligent man presenting a reasonable and
    informed opinion.

    This impression is reinforced by his on-stage persona — a persona, I should add, that
    shows no signs of being based on anything except his natural tendencies. Faced with
    persistent questions from open source users in the audience, Jeans consistently
    shows politeness and a willingness to listen. When pressed, he showed a reluctance to
    argue — partly, perhaps, out of an awareness of his limited time. Nor, when he lacked information,
    did he hesitate for very long before admitting the fact. Time and time again, he returned to the
    practical question of what business customers want, constantly repeating, “at the end of the
    day …” as a way of cutting off digressions without annoying anyone.

    In the end, Jean’s persona is perhaps his most successful tactic of all. Admittedly, his comment later that
    “It’s great to hear everyone’s point of view and that we can all agree to disagree on
    certain aspects” sounded a little rehearsed. Yet, at the same time, it sounded as though, on some
    level, he meant it.

    The effectiveness of Jeans’ self-presentation can be judged by the fact that, as I waited to interview him
    after the seminar, the people who approached him were the open source users in the audience. Without
    exception, they mirrored his politeness. One man even asked Jeans’ help in getting in to LinuxWorld Canada.
    Although their opinions were unchanged, even these informed people seemed to respect the Microsoft
    perspective in a way that they had not at the start of the presentation. The impression that people had of
    Jeans seems to have been transferred successfully to the company that he promotes.

    To what extent Barnaby Jeans is typical of other developer platform evangelists at Microsoft is unclear.
    According to Jeans, all Microsoft evangelists have considerable freedom in choosing how to approach their work.
    However, Jeans’ ability after the presentation to quickly name a counterpart in Australia suggests some pooling
    of efforts. Moreover, the recent
    comment by Jason Matusow of the Shared
    Source Initiative that there is a “movement to the middle” between proprietary and open source comments
    is a close enough echo of Jeans’ comments to suggest that his approach may be due as much to Microsoft
    policy as to his own initiative. Possibly, Jeans is simply more skilled than most at presenting the new approach.

    Yet for members of the FOSS communities who want to refute Microsoft’s evangelists, it may not matter whether
    Jeans is currently typical or not. What matters is that Jeans demonstrates the effectiveness of an alternative
    approach. Many people in the FOSS
    communities are used to the Microsoft response to open source being crude and hysterical. What Jeans proves
    is that it can just as easily be subtle and sound reasonable. Even more importantly, he shows that the rules of engagement
    can change at any time — and that the FOSS Community had better be ready when they do.

    Bruce Byfield is a freelance course designer and instructor, a technical journalist, and a NewsForge and ITMJ regular contributor.