- From: Erik Bruchez <erik@bruchez.org>
- Date: Thu, 16 Oct 2003 16:16:36 -0700
- To: Liam Quin <liam@w3.org>, public-qt-comments@w3.org
> speaking personally (i.e. without any W3C hats on!) I find this > use case compelling -- I have always found it frustrating when > languages add simple regexp matching without this, especially as > every regexp library I've ever used at the programming language > level provides the functionality. We are 100% in agreement about this one. > On decoding URLs and a QUERY_SRING or QUERY_PATH, I suspect you'd > also want access to environment variables Absolutely not! I am simply talking about transforming or querying documents containing encoded URLs. URL encoding and decoding functions fit perfectly in a functional language as they do not have any side effect. This is true for the current fn:escape-url(), and would be true for fn:unescape-url() (or whatever it is called). This is a scenario I hit recently: A Web form based on XForms allows the user to enter an URI, possibly with encoded query parameters, for example: http://example.org/rss?url=http%3A%2F%2Fwww.theserverside.com\ %2Fhome%2Fthread.jsp%3Fthread_id%3D21921 That form is submitted. An XForms instance containing that URI is created, and looks like this: <form> <my-uri>http://example.org/rss?url=http%3A%2F%2Fwww.theserverside.com\ %2Fhome%2Fthread.jsp%3Fthread_id%3D21921</my-uri> </form> Now, assume you want to process the XForms instance using XSLT or XQuery (that's what we typically do in the product my company is developing), and that you would like to decode the URL. You can easily parse the string with regexps, with fn:tokenize() for example, and obtain the value of the "url" parameter: http%3A%2F%2Fwww.theserverside.com%2Fhome%2Fthread.jsp%3Fthread_id%3D21921 Then you need a function to decode it and get the original URL: http://www.theserverside.com/home/thread.jsp/thread_id=21921 Somebody must have argued for the presence of fn:escape-url(). I argue for the presence of the inverse function, both for reasons of usefulness and of symmetry. -Erik
Received on Thursday, 16 October 2003 19:16:43 UTC