[XML-DEV Mailing List Archive Home] [By Thread] [By Date] [Recent Entries] [Reply To This Message]

Re: XML aggregation question?

• From: Tatu Saloranta <cowtowncoder@xxxxxxxxx>
• To: xml dev <xml-dev@xxxxxxxxxxxxx>
• Date: Sun, 27 Aug 2006 14:12:25 -0700 (PDT)

--- Robert Koberg <rob@k...> wrote:

> Since this an XML list, I will assume we are talking
> about XML. I will 
> also assume the main benefit of a DB is transactions

No, I think the first reason often is actually
performance, ie. indexing. Although (just like with
relational DBs) one can start structuring files and
directories based on way information is usually
accessed (to optimize access speed), this directly
leads to "didn't start with a DB, am writing one"
development. Likewise, DBs are good at caching
commonly needed pieces; figuring out optimal access
strategies and so on.

Transactions (and rest of ACIDity) follow closely,
along with access restrictions.

> (I don't think 
> eXist has that yet). But that can be handled by code
> at the level of a 
> wellformedness check and a validity check.

Hmmh? No, that's not how you do transactions (or is
even related as far as I can see). That's how you can
maintain schema restrictions, if you want those (most
XML DBs are better off without them at DB, and indeed
are better off maintain restrictions at app level).
Transactions have to do with locking, and parallel
views for distinct transactions.

> How does using a (XML)DB from the start compare?

Consider case of using bunch of Excel files, to that
of using an RDBMS: parallels should be clear. First
one (like having bunch of XML files on disk, or in
subversion etc) has its benefits for small projects,
prototyping, but also limits one's options for bigger
projects.

-+ Tatu +-


__________________________________________________
Do You Yahoo!?
Tired of spam?  Yahoo! Mail has the best spam protection around 
http://mail.yahoo.com 

• References:
  • Re: XML aggregation question?
    • From: Robert Koberg