Microsoft TechEd Trip Report: Monday, June 18th

by Brian Jepson

Monday began with a keynote from Paul Flessner, Senior VP of .NET
Enterprise Servers. He talked about using XML web services in the href="http://foldoc.doc.ic.ac.uk/foldoc/foldoc.cgi?query=enterprise">enterprise.
He first put forth a definition of a dream enterprise, which mirrored
the story Microsoft has been telling us for some time now: all your
customers, partners, and employees are connected through href="http://msdn.microsoft.com/library/techart/windnadesign_guide.htm">common
application services and a common data model. Then, he put forth
his definition of reality: a bunch of heterogeneous stuff (different
platforms, languages, etc.). For me, the dream enterprise is the
promise that Microsoft didn't deliver on. I'm glad they didn't, since
I believe that diversity of hardware and operating systems fuels
innovation (and makes life more interesting). I'm twice as happy to
see Microsoft accepting the world as it is.



Flessner went on to explain the various ways in which web services
will save the day. There's nothing new in that - we've been hearing
how XML and web services are going to unify href="http://directory.google.com/Top/Computers/Programming/Component_Frameworks/Comparison_and_Critique/">that
which could never be unified. Within the last year and
a half, I developed two href="http://www.tuxedo.org/jargon/html/entry/nontrivial.html">nontrivial
XML-based applications, and I think the industry
might have a sensible solution in XML.



We heard a customer success story, and then Flessner went back to his
slides. In a slide about .NET, Web Services were represented as pink
triangles. I hope that catches on - I wonder if the Visio 2002
templates will use pink triangles for web services.



A Few More Success Stories



I've never gotten the value of these customer success stories. I
guess it depends on your audience - when I learn a new technology, I
want to drink of its technical details and understand every aspect of
it. Then, I form a judgment as to whether it's something I want to
use. For example, I think href="http://www.apple.com/enterprise/">NEXTSTEP is a great
platform, even measured against the capabilities of current operating
system/object model/development environment
combinations. I think
these customer success stories provide comfort with a little dose of
peer pressure - this guy bet his enterprise on it, why don't you?



If you want to get developers excited about this stuff, don't focus
on how fast it is, how many seats it supports (this stuff is
important nonetheless), but tell us how much fun it was to
develop. Maybe that's the difference between TechEd and the PDC (this is my first TechEd, and I've never been to the
PDC).



Benchmarks



Then he got into SQL Server benchmarks. I'm jaded and I don't like
benchmarks. If you want to do benchmarks that I'll be interested in,
set it up in an arena, gladiator-style, for all to see. Form tag teams
made up of the best href="http://www.microsoft.com/sql/default.asp">SQL Server, href="http://www.sybase.com/">Sybase, href="http://www.postgresql.org/">PostgreSQL, href="http://www.oracle.com/">Oracle, href="http://www.mysql.com/">MySQL, and href="http://www-4.ibm.com/software/data/db2/">DB2 data base
administrators (DBAs). Have them go at it, tuning their machines,
sweating furiously under the floodlights, until one emerges
victorious! Lather, rinse, repeat - hold the games every year. Better
yet, how about a database Olympics - each database will earn the gold
in different areas, and the users will know the strengths and
weaknesses. But I get sick of benchmark wars, be it operating
systems, databases, or anything.



Mobile information server



Microsoft announced the href="http://www.microsoft.com/miserver/default.asp">Mobile
Information Server today. They showed us an interesting mobile
application where a task was assigned to a developer and relayed to
his mobile phone. Although there was a glitch in the demo (some
trouble refreshing a web page), it showed how the developer was able
to complete the task, update the task management system from his
phone, and have that completed task propagate to the desktop
application.



I like the idea of having certain types of information
centralized. For example, I'd have no problem with Microsoft or
Netscape storing my bookmarks on their server. I don't want them
stored on my computers - it's a total hassle to keep moving bookmarks
between machines.



Do your href="http://www.microsoft.com/net/hailstorm.asp">Hailstorm-esque
centralization with something simple first! You can't do this with my
TODO list, because you can't guarantee that I'll be online when I edit
it. If I'm using my bookmarks, you can assume that I'm online
somehow.



Make it so my bookmarks find me when I fire up any computer: my
laptop, my Palm, Power Mac, or desktop PC. If I add a bookmark on my
Palm, have it instantly available on my desktop. It shouldn't matter
whether I'm using Mozilla, href="http://www.opera.com/">Opera, href="http://home.netscape.com/browsers/index.html">Netscape, href="http://www.microsoft.com/windows/ie/default.htm">Internet
Explorer, or Konqueror! If
you could pull this off, I'd be impressed. Then, maybe users would trust you
to centralize everything else, like our medical records and high
school transcripts.



Read about day one here.