- From: Tony Rogers <tony@gonk.net>
- Date: Sat, 18 Dec 2010 18:46:53 -0500
- To: XProc Dev <xproc-dev@w3.org>
- Message-Id: <4D74583F-91E1-4F14-A120-3039B204FF59@gonk.net>
Something has been bothering me for quite some time. I can confidently say XProc is powerful XSLT 2 is powerful XPath 2 is *unbelievably* powerful (and can be used independent of the previous languages) … However…is it just me, or um—is it really, *really* weird just how little control I have over the formatting of XML created using XProc/XSLT? (And by “formatting”, I mean “coding style”. If you’ve ever joined a project with a document titled something like “Coding Style Guidelines”, that’s what I’m talking about.) … I mean, sure—there’s @indent = “yes | no”. And @xml:space, strip-space, and preserve-space. But hasn’t anyone ever put a lot of time into really, REALLY using these everywhere possible, and then looked at the generated XML and thought “God, that’s just hideous!” ? I mean, it’s not uncommon for me to find that generated markup is irritating as hell to read. … I constantly find myself arriving at the following questions: What if I want to specify certain elements have 1 attribute per line each? What if I want to apply that to certain elements but not others? What if I want to apply it to an element only SOMETIMES—depending on arbitrary conditions? … I see these amazingly powerful tools, and I see that they recognize whitespace is significant. But it’s significant beyond CDATA and snippets of code. There is real value in producing XML that is *easy for human eyes to read*. (And if that weren’t true, then obviously programmers would never, ever argue about coding style…or have a reason to publish documents titled “Code Style Guidelines”. Am I right? ☺) … So, is this ever a problem for anybody else? Is there a solution to this I haven’t heard about yet? Any current ongoing work on this issue? … Thanks one and all, —Tony
Received on Saturday, 18 December 2010 23:47:30 UTC