- From: Craig Brozefsky <craig@onshore.com>
- Date: 31 Jul 1998 22:19:05 -0500
- To: "Les Cuff" <lez@fastfwd.com>
- Cc: "w3www-dom" <www-dom@w3.org>
"Les Cuff" <lez@fastfwd.com> writes: > Would it be off base to suggest that (when the time comes) Frozen Document > Objects be hoerked out by the servers; into the waiting maw of the DOM aware > clients? That sounds like, at last, the 'Compiled Hypertext' I have been > seeking! All the world is not a DOM implementation. Of course it's not difficult to have preferred response types and the like make this only happen when a compatible DOM impl. is on the requesting end. > An aside from the bandwidth junkie that lurks within me: can we PLEASE > settle on a compressed transmission protocol? XML seems like teXtML whereas > I want gzipML. That's a transport issue. And rememeber that whatever you shoot over the wire, the more human readable it is, the easier it is to debug, the longer life it will enjoy. The internet protocls like HTTP SMTP NTTP and others are prime examples. > In addition to resolving & into &, the text could be sanitized for white > space by reducing all white space runs into the single ' '. Analysers might > benefit from the guarantee. Back to the issue of human readability. What if I wanted that whitespace there for my own formatting purposes when generating, debugging, changing the document. Do not ever silently change that which has been given to you. I don't mean to respond only to the ones I disagree with. Your other suggestions seemed like interesting ideas. I'm just trying to make the point that being freindly to humans is good on the wire, and whenver you throw data around. The issues of on the wire human readability, simple largely stateless protocols which are easy to simulate; is something I think should be kept in everyone's mind when making these DOM mechanisms, since that is something I feel has led to the spread of other succesful protocols like HTTP and SMTP.
Received on Friday, 31 July 1998 23:11:47 UTC