There seems to be a lot of noise lately about
“ROX”.  Relations, Objects,
and XML.  Several
interesting opinions and thoughts
and even some research.  "urn:schemas-microsoft-com:office:office" />


Working for the Microsoft data access team
(first MDAC, then SqlXml, and now ObjectSpaces) for several years now, I can tell
you we have been doing a lot of thinking about world of ROX.  PDC
2003 is all about showing the cool new data access frameworks, but more importantly
– it is all about explaining and reconciling this, this, this, and this.  With
flexibility and extensibility comes complexity, but I believe PDC we have a good start
at providing a data access roadmap.  But
that will only be a start…


In my opinion we are far from being done.  As
a developer using data access, I think one really needs the option to use any part
or combination of ROX data access in a given application to solve a diverse set of
development problems.  For example Xml
for the protocol layer, Objects for complicated business logic, and relational access
for performance and control. 

Completely solving this problem means consistent
patterns and APIs, interoperability between data sources, the discovery of domain
specific features that are really not (think XPath over Objects, using SqlXml as a
object persistence layer), and not getting hung up on the domain specific eccentricities
(think processing instructions/ comments in Xml) that can’t be mapped to other domains.  Are
we there yet?  No way.  Are
we going in the right direction?  I believe

Comments (0)