auth-tools.html 3.45 KB
<HTML>
<HEAD>
  <TITLE>WWW4 Breakout Session on Authoring Tools</TITLE>
</HEAD>
<BODY>
<P>
<A HREF="../../../Conferences/WWW4/"><IMG SRC="../../Icons/WWW95/WWW4" ALT=""
    WIDTH="120" HEIGHT="120"></A>
<H1>
  Authoring Tools Breakout Session
</H1>
<P>
<B>Update (Jun 96)</B>: W3C is mustering resources to carry this work forward.
See: <A HREF="http://www.ics.uci.edu/~ejw/authoring/">WWW Distributed Authoring
Working Group Home Page</A>
<P>
These are my notes from the break-out session following the "development
tools" session on <A HREF="../../../Conferences/WWW4/Thursday">developer's
day</A>.
<H2>
  Attendees
</H2>
<UL>
  <LI>
    Dan Connolly, W3C
  <LI>
    Tad Coburn, Vermeer &lt;tcoburn@vermeer.com&gt;
  <LI>
    Dave Long, Navisoft &lt;dave@navisoft.com&gt;
  <LI>
    Ed Burns, SGI &lt;edburns@sgi.com&gt;
  <LI>
    ... others? (not sure)
</UL>
<H2>
  Focus
</H2>
<P>
Interoperability among distributed (i.e. client/server) web authoring tools.
<H2>
  Issues
</H2>
<H3>
  Strong Authentication
</H3>
<P>
We agreed this was a requirement. Either
<A HREF="http://www.ics.uci.edu/pub/ietf/http/draft-ietf-http-digest-aa-01.txt">md5
auth</A> or a combination of SSL and basic auth would suffice. But no product
should restrict folks to basic auth.
<H3>
  How to get "raw HTML," i.e. before server-side includes processing?
</H3>
<P>
Accept: text/plain (or text/x-html-ssi, or whatever NCSA made up)
<H3>
  <A HREF="../Locking">Lost update problem</A>
</H3>
<P>
Four levels of support:
<OL>
  <LI>
    no support. Lost updates go undetected
  <LI>
    dectect lost update (timestamp, version/derived-from)
  <LI>
    prevent lost update (locking)
  <LI>
    provide fairness (long story... not really discussed at the meeting)
</OL>
<P>
level 1 is not really acceptable. Need to find out/influence the HTTP 1.1
spec with respect to Version:, Derived-From:, etc.
<P>
Creating files and directories raises some issues.
<H3>
  Common Access Control Model
</H3>
<P>
We decided this isn't a goal worth persuing. Everybody will implement their
own forms-based access control model.
<H3>
  URL mapping
</H3>
<P>
Again, interoperability here isn't likely.
<H3>
  What HTML DTD?
</H3>
<P>
Please label document with the DTD that your product supports, e.g.
<PRE>&lt;!doctype html system "http://www.vermeer.com/public-text/html-951224.dtd"&gt;
</PRE>
<H3>
  New HTTP Methods: BROWSE, MKDIR
</H3>
<P>
These could be implemented in terms of GET and POST, but since they are likely
to have different access control policies, they deserve their own methods.
<P>
What about a CREATE method?
<H3>
  Line Breaking
</H3>
<P>
To pretty-print, or not to pretty-print. That is the question. Two issues:
<UL>
  <LI>
    newline representation: CR, LF, or CRLF? Try to detect what it was and leave
    it that way
  <LI>
    where to put newlines? Preserve old formatting, or normalize?
</UL>
<H3>
  Relative Links
</H3>
<P>
Hmm... what was the issue here? I know NaviPress gets this wrong sometimes
(e.g. when you save as, it doesn't fix relative links! And it goofs up images
sometimes too.)
<H3>
  Server-Side API
</H3>
<P>
Purveyor/MS ISAPI? NSAPI? CORBA-based access control?
<H3>
  Editing Variants
</H3>
<P>
How to edit the french version of foo.html? Get the server to tell you it's
at foo.html.fr and edit that.
<P>
  <HR>
<ADDRESS>
  <A HREF="../../"><IMG SRC="../../Icons/WWW/w3c_48x48" WIDTH="48" HEIGHT="48"></A><BR>
  Connolly<BR>
  last update by $Author: connolly $ on $Date: 1998/05/21 17:40:09 $
</ADDRESS>
</BODY></HTML>