- From: Kay, Michael <Michael.Kay@softwareag.com>
- Date: Fri, 16 May 2003 06:25:44 +0200
- To: Mike Haarman <mhaarman@infinitecampus.org>, "Kay, Michael" <Michael.Kay@softwareag.com>, public-qt-comments@w3.org
Mike Haarman wrote: > > In the current toolset, if I ask for substring($date, 4, 1) I > get it whether it is junk or not. The stylesheet doesn't > fail. I appreciate why many consider this a problem. I feel > that it is a significant factor in XSL's success. There are indeed a range of views on this. This is reflected in the frequent use in XSLT 1.0 (still retained to a lesser extent in 2.0) of the phrase "it is an error ... the processor may signal the error, or may recover...". My understanding is that the original intent of making errors silently recoverable was to meet the requirements of client-side transformation. This has been less prominent in the requirements this time around, and the emphasis has switched to a "safer" mode of operation, in which errors get reported so that they can be corrected. I think that the silent recovery from errors makes it difficult for users to debug their stylesheets. For example, use of xsl:value-of with a node-set, where nodes after the first are ignored, is a common and very bewildering mistake for novices. Although our error handling facilities still leave something to be desired, I think the less permissive approach in XSLT 2.0 will make it easier for users to develop correct code. Thanks for your comments and responses. Michael Kay
Received on Friday, 16 May 2003 00:27:27 UTC