Date: Fri, 20 Nov 92 21:18:36 -0800 From: marca@ncsa.uiuc.edu (Marc Andreessen) Message-Id: <9211210518.AA14876@wintermute.ncsa.uiuc.edu> To: Dan Connolly <connolly@pixel.convex.com> Cc: marca@ncsa.uiuc.edu (Marc Andreessen), www-talk@nxoc01.cern.ch Subject: Re: Annotation & asynch collaboration In-Reply-To: <9211210309.AA13411@pixel.convex.com> <9211210500.AA14826@wintermute.ncsa.uiuc.edu> Dan Connolly writes: > >ANNOTATE /file/to/annotate?/file/of/annotation?Text describing annotation. > > I'd suggest > > APPEND /file/to/annotate > <a HREF="file/of/annotation">Text describing annotation</a> > . > > The server could impose limits on the length of the annotation, > and it should somehow validate the syntax, I suppose. The point behind *not* doing general appends is that they aren't a good idea since document integrity is breached (a link is immediately obvious as just that and will cause little confusion [especially if clients always prepend a fullname], but if arbitrary text can be added to a document, you'll never know what's going on). To be practical, I know *I'm* not going to run an httpd server that allows anyone to append any text they want to my documents. And if general appends aren't going to be allowed (through server validation of the <A...> syntax or whatever), then why not just let the server construct the link append itself? Marc