- From: Mark Baker <distobj@acm.org>
- Date: Tue, 3 Apr 2007 00:26:24 -0400
- To: "Dave Pawson" <dave.pawson@gmail.com>
- Cc: www-tag@w3.org
On 4/1/07, Dave Pawson <dave.pawson@gmail.com> wrote: > Downstream processing of xml content requires validation and hence > versioning to assure the processor that the content being worked > is as expected. Processing content requires only that the recipient be able to understand it. Validation plays no role in that. At best it's just one way that a document recipient can identify content that might not be understood. But even without validation the content would certainly be found to be "invalid" eventually, as processing is attempted. IME, if you define a decent extensibility model such that future documents can be processed by old software, not only do you not need validation, but it actually gets in the way of building evolvable systems. I've written about this recently; http://www.coactus.com/blog/2006/12/validation-considered-harmful/ I agree completely with Lachlan's comments. Mark.
Received on Tuesday, 3 April 2007 04:26:27 UTC