ObjectSpaces Showdown...

Published 30 January 04 05:51 PM | despos

I think that really interesting statements and comments have been posted in the various threads about ObjectSpaces. A monumental THANK YOU to you all.

Kathleen Dollard said "You'll accept a 30% or greater perf hit for simplicity in programming? Sorry, but I think that's nuts." I've been brought up thinking that scaling hardware up fixes many things... <g>

Paul Wilson pointed out that "Actually, the 30% extra overhead comment that's been circulating was widely misunderstood.  They were stating what they think ObjectSpaces can be compared to DataReaders, not DataSets. When they clarified this it was clear that they don't think there should be much, if any, overhead when comparing to DataSets."

It's interesting to see how to chase my thoughts I was completely losing the perspective of a thing like ObjectSpaces.

Here's (what I believe is) a (to date) summary of the ObjectSpaces stuff.

  • It is not that you create an object model for the sake of ObjectSpaces; you use ObjectSpaces because you have an object model that you want to persist.
  • You must have a strong business layer to use ObjectSpaces; it is not a matter of having a large or complex application
  • If you need maximum speed and flexibility use ADO.NET; if you have a strong business logic layer use ObjectSpaces; if you're writing a data driven application use perhaps the XML/ADO.NET API

 

Comments

# Paul Wilson said on January 30, 2004 03:05 PM:

Here's the blog entry from Andrew Conrad on the ObjectSpaces team that attempts to clarify the performance myth: http://blogs.gotdotnet.com/aconrad/PermaLink.aspx/5d3bf937-16a2-4c3c-a59d-13f1e583de0b.

# TrackBack said on January 31, 2004 12:11 PM:

Leave a Comment

(required) 
(required) 
(optional)
(required)