- From: Dave Singer <singer@apple.com>
- Date: Wed, 23 Jan 2008 19:18:07 +0100
At 17:33 +0000 23/01/08, Philip Parker wrote: >What about having it render as a standard unordered list ( ie, >bulletpoints ) until the entire set of items has been received - and >then re-rendering the list as a numbered type, all properly >calculated how about assuming that if the source wants it numbered, in reverse order, it knows what it is doing, and can tell the browser what number to start at? it still seems the simplest; an attribute that gives the starting number (default 1) and an attribute that gives the direction (increasing or decreasing, default increasing). > >James Graham wrote: >>Siemova wrote: >>>On Jan 23, 2008 10:54 AM, David Walbert <dwalbert at learnnc.org >>><mailto:dwalbert at learnnc.org>> wrote: >>> >>> >>> It's not that simple -- the last line should be >>> >>> start = 1 + ( (number of items - 1) * step) >>> >>> if it's assumed that the last item of the list is numbered one by >>> default. >>> >>> >>>Alas, we see the ill effects of my hastiness today! I stand >>>happily corrected. In that case, it's even simpler: >>> >>>if start is not specified >>>start = 1 >>>if reverse >>>start += (number of items - 1) * step >> >>The problem that Jonas originally pointed out is that, given >>browsers do incremental rendering "number of items" is not a known >>quantity when the list is first rendered. For a pathological >>example of why this is a problem, imagine a cgi script that just >>kept spewing out reverse numbered list items, one per second, >>indefinitely. >> >>It may be that in practice lists are short enough that they are >>typically rendered all in one go so this wouldn't be a problem. I >>don't think that's obvious, however. -- David Singer Apple/QuickTime
Received on Wednesday, 23 January 2008 10:18:07 UTC