Re: [xsl] When to use text()

Subject: Re: [xsl] When to use text()
From: Martin Holmes <mholmes@xxxxxxx>
Date: Fri, 21 Mar 2014 11:21:39 -0700
On 14-03-21 11:06 AM, Ihe Onwuka wrote:
On Fri, Mar 21, 2014 at 6:03 PM, Michael Kay <mike@xxxxxxxxxxxx> wrote:

I am arguing against RTFM as a stock response.

There are times when it is definitely appropriate, but  I don't think
this is one of them.

Heiko Niemann started this thread by asking for advice on how to use a particular construct, and various people offered advice that addressed his question.

You, Ihe, then raised a question about who was to blame if programmers got it wrong, the programmers or the spec writers?

We have an honourable and valuable tradition in the programming profession that we try to solve problems without apportioning blame.

If there are usability problems, it can certainly be instructive to study how they arose, and how the user's expectations of the system came to be different from its actual behaviour. But you don't seem to be engaging in a constructive discussion of that question. Discussing "who is to blame" is almost never useful, because usually, no-one is to blame; the cause of usability problems is generally that the designer didn't understand the mindset of the user, and vice versa, and neither can be blamed for that in a world where the the user was not available to be consulted at the time the design was done.

As it happens, no one responded by saying "read the F manual". On the contrary, several people contributed information derived from the F manual, to spare Heiko the trouble of reading it.

Martin Holmes mholmes@xxxxxxx via

3:28 PM (2 hours ago)
to xsl-list
text()..... as in text()...... not text() interleaved with comment nodes.

I must have missed that version of the spec. :-)

To be fair, that was in answer to your claim that text() doesn't do "what it says on the tin". I assumed that your tin was the specification; I was honestly asking what part of the specification you read that had caused you to believe that text() would behave differently than it does. Only after the thread proceeded for a while did I realize that you meant that you had deduced how it "should" behave from the word "text" itself. That had seemed implausible to me because "text" is such a generic term that it could mean almost anything; I'd assume that anyone encountering something like this who hadn't come across it before would look it up somewhere rather than trying to guess how it works (or perhaps write test cases to deduce how it works, if they're averse to spec documents and tutorials).


Cheers, Martin

Current Thread