- From: Michael Kay <mike@saxonica.com>
- Date: Mon, 12 Jan 2015 21:37:06 +0000
- To: Stefan Wachter <stefan.wachter@gmx.de>
- Cc: xmlschema-dev@w3.org
> Many thanks for that information. I will definitely try to implement the algorithm described by Mr. Thompson. He already was so kind to provide me the mentioned article. In addition I will implement the runtime check for the undecidable cases. > > Is there a specific reason why you decided that Saxon does not support the "runtime-check" solution? > The schemas where testing for restriction is hard are the same as the schemas where validation is difficult and slow, and they don't come up very often. To handle these schemas, I would first need to implement a validation algorithm that uses backtracking rather than DFSAs, and that is disruptive because it's not streamable. It's only worth doing the validation-time restriction checking after I've done that. Fact is, users don't complain about Saxon's limits often enough to justify the effort. Michael Kay Saxonica
Received on Monday, 12 January 2015 21:37:31 UTC