- From: William F Hammond <hammond@csc.albany.edu>
- Date: 11 Aug 2002 16:21:40 -0400
- To: Chris Croome <chris@webarchitects.co.uk>
- Cc: www-html@w3.org
Chris Croome <chris@webarchitects.co.uk> writes: > On Sun 11-Aug-2002 at 01:46:32PM +0200, Jan Roland Eriksson wrote: > > > > Any way, going for the use of PI's to reference non structural > > external info would "pave the way" for a removal of the 'link' > > element too :) > > There was quite a bit of discussion of the TAG about the use of PIs > and the consensus seemed to be that new uses of them would not be a > good idea, nor would getting rid of them. Sounds like a reasonable TAG conclusion. If "script" is going to survive at all, I think the PI approach is probably the best option since it involves no specific expectation for either markup or client behavior. But the whole idea of client-side processing tied to publicly-served default-format web pages [as indicated by mimetype :-)] has presented unacceptable network security risks from the start. In the context of <script> much extant practice is unnecessary, and none of it is acceptable from the viewpoint of network security as long as there is a significant presence of platforms without both (1) adequate protection from memory over-writing and (2) a reliable well-tested (also IMHO "open source") library base. -- Bill
Received on Sunday, 11 August 2002 16:21:43 UTC