- From: Tod Matola <tmatola@columbus.rr.com>
- Date: 16 Apr 2001 23:19:13 -0400
- To: Murray Altheim <altheim@eng.sun.com>
- Cc: Dan Brickley <danbri@w3.org>, Charles "F." Munat <chas@munat.com>, RDF Interest <www-rdf-interest@w3.org>
On 16 Apr 2001 17:00:50 -0700, Murray Altheim wrote: > Dan Brickley wrote: > [...] > > It doesn't follow from the use of LINK REL that you have to manage > > metadata on a file-per-document basis. You might point into a CGI or > > servlet (hopefully disguising this to the outside world), or autogenerate > > those files from a database managed elsewhere. I do share your bias > > towards (at least some) embedded metadata though. > > Ideally, a good document management system should manage revision cycles, > author info, etc. for an author, but absent the $100K solution it's nice > to figure a way to allow a single author (or a married one for that matter) > to add metadata directly into a document using vi. > I would second the approach that the metadata needs to live in-line in the document. If we claim that magic happens and out scrapes your nice metadata we are slipping backing into not having a way to do this. With autogeneration (CGI/servlet) of metadata, you have a whole new set of problems (e.g., if it is static then it can be out of sync, if it is automatic there is sometimes you needed to say this one doesn't need generating). Not to mention the storage of data that is already marked up (say the abstact, or subject headings), should I cache or shouldn't I cache. You need reuse the real content in several different lights or depending on what pair of goggles you look at the content with. I think the linking is fine for small sets, but easily gets out of hand. And I like the test that you should be able to add it with vi (err let's say emacs). Cheers Tod Matola EOR Developer
Received on Monday, 16 April 2001 23:18:11 UTC