Subject: Re: [xsl] XPath that returns nothing From: Andrew Welch <andrew.j.welch@xxxxxxxxx> Date: Thu, 29 Sep 2011 12:27:26 +0100 |
> If overnight, every night, you are making thousands of documents and in > one 40 page document a constraint on one parameter is missing. > it's much easier to point the developer at the document and say please > supply the information required to make the red bits go away, than to > say, your document failed to build last night as an internal xpath > selection failed a type assertion. Ah ok so it's the situation where 98% correct output is far better than no output at all... and that sounds like it's because the turnaround time is a day - if it was an hour or two would the same still apply or could you skip the red bits altogether? If so, then it leads on to the fun part of trying to get the total transform time down... (optimising, compiling, caching, multiple threads, better hardware etc) but also being able to just re-run a single file or step x onwards really gets the time down and means you can go through the loop a few times. I guess that sums it up - run-to-the-end and 'never fail' is needed if there is there a long turnaround, but ideally you want as short possible turnaround and then to get the short feedback loop you can fail-early. -- Andrew Welch http://andrewjwelch.com
Current Thread |
---|
|
<- Previous | Index | Next -> |
---|---|---|
Re: [xsl] XPath that returns nothin, David Carlisle | Thread | Re: [xsl] XPath that returns nothin, Wendell Piez |
Re: [xsl] Identifying unique attrib, G. Ken Holman | Date | Re: [xsl] Identifying unique attrib, Mark |
Month |