Transition from Legacy to Native rendering - (was Re: [whatwg] repetition model)

On Sat, 26 Jun 2004 02:07:52 +0100, Dean Edwards <dean at edwards.name> wrote:
> Jim Ley wrote:
> > No I can't, they will not meet my QA teams requirements (HTC's are
> > outlawed, they're SHITE!) they'd also need to be available in a
> > suitable licence and up to sufficient code quality to satisfy me, I
> > really can't see that happening, especially as it's likely to be a
> > considerable bloat to the size of each page impacting on rendering
> > time and download times.
> > 
> can you outline your objections to HTCs? i am in touch with the IT
> department of a huge company that has invested heavily in them (100,000
> user intranet). although they have had some problems, it's my
> understanding that they are not insurmountable. 

Where I'm am at the moment before I arrived they'd invested heavily in
them, but were having problems (many browsers didn't cache, and
couldn't be made to cache even in new fully patched IE's a return of
the old IE5.5 problems with caching of them for each element in the
document the behavior applies to which I believe are well documented,
but only occuring if you doc.write the content in an iframe by the
looks of things.) they leaked a lot of memory, many orders more than
IE/jscript does other times.   They got some MS dev guys in to solve
it, they recommended they remove them, for similar reasons.

Most of these are only noticeably once you get a number of elements
acting on it, it probably is an acceptable solution for a page that
reloads regularly, and only has the behavior attached to a couple of
elements (perhaps just attach it to FORM, rather than each INPUT
element.) but you're just working around the poor implementation
rather than anything else.

> page bloat can be avoided by implementing correct caching. 

Caching doesn't solve page bloat! it just helps it out for the repeat
visitor, not the initial one who is often more important - and anyway
that script still needs to be compiled each time too.

Jim.

Received on Friday, 25 June 2004 18:33:54 UTC