Re: [Bug-gnupedia] A Detailed Proposal - Mk I

[Bob Dodd]

Bryce Harrington wrote...
[snip]
> In playing with a software system called Wiki, wherein documents of
> arbitrary topics are collected and organized, we learned that an
> approach is to keep the hierarchical ordering independent of the
> article.  The article is stored in a "flat" collection, with a unique
> but essentially arbitrary name.  Indexes, TOC's, link-lists, or
> whatever
> other preferred hierarchical scheme can then be generated, with
> (hyper)links to the article.  A side benefit is that you have an
> unlimited number of potential alternate ways to organize...  If I'm
> interested in ancient history, the link will appear under
> greek/architecture.  If I'm an architecture buff and want a list
> focused
> to that, then history/greek.  Etc.  The same article is available
> from
> each direction, but is not physically stored in any particular
> hierarchical system.

In essence , that is what I've been suggesting: you need to separate
content from presentation. 

First you have a "core" data model (your flat collection) holding the
basic information e.g. all the tags for name, subject, version etc. we
have been discussing.

Second you have multiple "views" (or catalogs, depending on your
terminology) of that information where each view has it's own
organisation strategy and content rating system. And, of course, that
organisation strategy is implement through indexing, TOCs, hyperlinks
etc, just like you were doing with your wiki.

Since we have been talking DTDs and schemas for the past couple of
days, I'd also point out that you'd really want (at least) two levels
of XML DTD/Schema: one as a wrapper for the "core" content, and one as
a wrapper for index entries in the "view" which would extend the tags
to include rating information, additional cross-linking etc.

/Bob Dodd


__________________________________________________
Do You Yahoo!?
Yahoo! Auctions - Buy the things you want at great prices. 
http://auctions.yahoo.com/