- From: Daniel B. Austin <daniela@cnet.com>
- Date: Tue, 26 May 1998 12:43:56 -0700
- To: Chris Lilley <chris@w3.org>
- Cc: html-future@w3.org
Hi Chris, At 06:50 PM 5/26/98 +0200, you wrote: >> Ashvil D'Costa wrote: >> >> One important feature that I would like to see in the next version of >> HTML would be to cover relationships between different XML >> applications. > >Have you already considered and rejected the namespace mechanism (if so, >please tell us why) or were you unaware of it? > >http://www.w3.org/TR/WD-xml-names > >It doesn't solve all the issues you raise in your paper, but it does >address many of them. One thing I noticed while reading the latest namespaces working draft (05-18-98) is that the example given in section 3.1 would be rather difficult to code correctly by hand, and that the longer the document becomes, the more difficult and error prone hand construction is going to be. Another point is that error checking and corrrection also become correspondingly more difficult. In a hypothetical future HTML document we might see markup like this: ... <block:p> <block:iframe att1="" att2=""> <text:span>Enter Your Name</text:span> <forms:input type="text" att1=""></forms:input> <xlink:simple href="help.doc">click here for help</xlink:simple> </block:iframe> <block:p> ... (presumably styles would be attached thru class and id and so on, this is not intended to be a rigourous example but to be illustrative.) My problem with this is that this markup is far too complex to be written well (or cost effectively) by hand, difficult to check (it will require sophisticated authoring tools) and that this problem increases in a way proportionate to the length of the document. To me, this seems like an excellent supporting argument to one of my ongoing proposals for the future of HTML - that HTML be designed specifically for machine generation without human involvement. Regards, D-
Received on Tuesday, 26 May 1998 15:49:05 UTC