RE: Strategies in XML storage and display interaction

[Leigh Dodds <ldodds@ingenta.com>]

-----Original Message-----
From: Jonathan Asbell [mailto:jasbell@i-2000.com]
Sent: 19 April 2001 06:16
To: xml-dev@lists.xml.org
Subject: Strategies in XML storage and display interaction


>I work for a major publishing company and we are storing content as an xml
clob, which we are indexing with oracle 8i.
>The clob is called from the application and xsl is used to display it.  My
questions are as follows:

>1) Is this a good strategy

Its one of the 3 ways that Oracle suggest to store XML with Oracle. The
other two are: to
store references to the XML files on the file system (i.e. BFILEs) which are
still indexable
using Intermedia; 'break-up' the XML and store the data as you would any
other data.

As you're in a publishing company you're data is likely very document
centric. So the
latter option is probably less useful. (Although even so there's often
document metadata
that might be useful to store separately - we found this useful as it was
less effort
than jumping through some hoops with Intermedia).

> 2) Should character references be stored for special characters or should
it just be pure
> characters (ie. should we store — for an emdash character or should
we store actual utf-8
> encoded characters only)

We actually store the original entity references, as it seems to provide the
most flexibility.
For applications that just want the characters then they can resolve the
entities to the
unicode references, however in the presentation layer you often want to
resolve these to
images that'll display correctly in a browser. Netscape and IE are wildly
different in what
they're usefully display, especially when you venture into the realms of
various
scientific characters. (Roll on Unicode enabled browsers).

Cheers,

L.


--
Leigh Dodds, Systems Architect       | "Pluralitas non est ponenda
http://weblogs.userland.com/eclectic |    sine necessitate"
http://www.xml.com/pub/xmldeviant    |     -- William of Ockham