- From: <bugzilla@wiggum.w3.org>
- Date: Fri, 18 Jan 2008 19:33:34 +0000
- To: public-qt-comments@w3.org
- CC:
http://www.w3.org/Bugs/Public/show_bug.cgi?id=4844 ------- Comment #2 from andrew.eisenberg@us.ibm.com 2008-01-18 19:33 ------- While cleaning up some of my email, I noticed that the author of this request later provided an explanation of his need. He said, in part, "Sometimes you want to hide the logic behind the ordering for a list. Without any explicit ordering and depending on how items where added to the database, a resultset could reveal how products differ in price, or freshness. If this is sensitive information, you could sort by a nonsensitive field, like a name or something. But sometimes you want to avoid the strict appearance that comes with alphabetical ordering, or there may not even be any textual data to sort by. Any type of lottery equivalent. Like a competition where highest score is shared by multiple contestants, yet there can be only (n) official winners. Those need to be picked so it doesn't depend on application date, contestant's name etc. So you sort by random and pick the uppermost (n).Similarly, you want to show a selection of 10 out of a larger set. It could be a highlight of products, or a list of online users. Where you don't visitors to say "I've already seen this exact presentation". Anyway, most RDBMS supports random ordering, being standard or not. So it's likely that many developers are already relying on it, and they would expect it to be available in XQuery also, once they make the transition to xml. By the way, if this feature gets adopted, maybe it should be discussed for xsl:sort as well?"
Received on Friday, 18 January 2008 19:33:41 UTC