to .net or to .not

Recently, Richard Grimes wrote an article in Dr. Dobb's about what's wrong with .Net ... more recently, there's been a very active discussion going on at Dan Fernandez's blog where Richard and Dan have been going at it ... throwing in my 2 rupees below, in support of Richard's claim that a large part of the .Net platform is still holding onto a win32 ancestry ... My biggest gripe is COM+ ... SWC is a step forward from the ugliness of vanilla COM+ deployment, but .Net certainly has a ways to go to catch the J2EE stack. If objectspaces was still happening, you could argue there was a competitive MS entry into ORM, but hopefully NHibernate will help fill the gap.

my comments at Dan's blog...
which clr are we even talking about ... due to SP1 changes,
presumably around the SWC functionality (services without components),
the CLR is now forked AFAIK ... there is CLR SP1 for .Net and there is
CLR SP1 for 2003 Server ... this certainly underscores Richard's point
on wrapping native libraries ... on that subject, when will COM+ die
and be replaced with a .Net-based component transaction manager? This
is an area where the J2EE stack is stomping the MS technology stack
... look at what JBoss & Hibernate are doing ...
server 2003
7982 KB

other os's
10453 KB


Time to toss that WEP router




Technology Predictions for 2005

my silly predictions for 2005-ish
  1. BEA dies or gets absorbed, maybe by IBM
  2. JBoss becomes the leading J2ee app server in high-profile deployments
  3. RSS / Atom start to show up in consumer devices
  4. Telecom companies start getting REALLY pinched by VOIP, mergers become essential to survival
  5. 1st generation of something better than iPod, smartphones, or PDAs comes out, but too expensive for anyone to buy (yet) ... like a pocket secretary maybe, speech-based interface might be the enabler ... MS might play well here as they've done the most R&D
  6. Security and Exploits both become big business
  7. Microsoft Re-wraps COM+ (again) for the longhorn transaction manager
  8. longhorn (server) and yukon both are unusable this year
  9. Oracle CPU pricing matches or is less than SQL server by year's end
  10. MS gets more involved in hardware