Massive XML docs

[Ruslan Zasukhin]

On 2/8/06 8:59 PM, "Mark Wieder" <mwieder at ahsoftware.net> wrote:

Hi Mark, 
Hi Todd,

> Todd-
> 
> I'm with Ruslan on this one. DOM is the fastest way to access xml information,
> but the tree size will be huge (are you talking about a single 100MB file
> rather than multiple files adding up to 100MB?).

> You can write a SAX processor, in which case you won't have the whole tree in
> memory, but your speed will drop noticeably.

SAX is not good for many iterations. Usually it is used to parse all data
and e,g, transform them or store into DBMS.
 
> XML is nice in that it's human-parseable as well as machine-parseable (notice
> I didn't say readable), but it's quite a bloated format for both storage and
> searching. 

right

> Native xml databases provide speed for data storage and retrieval but lack in
> querying speed unless they use significant resources in indexing.

If cover Native XML databases.

* they tend to use Xpath and/or Xquery as query language.
* they do indexing like DBMS do BUT ...

I can bet that quite soon they will go into history.
I have see similar opinions on oracle.com :-)
Oracle guys point that last 2 years vendors of Native XML dbs almost do not
take part in development of Xquery standard and so on. This job do mainly
Oracle, IBM, MS.

Again, IMHO, this was just a fashion stream. But Text format is text format.
In 1970th programmers have invent dbs exactly to run away from text formats.

We have made a lots of discussion last time in team, with other developers
who heavy use e.g. MS, with our university guys about what advantages gives
XML and where it should live.

Our resume: at least in areas:
    - WEB
    - data transfer
    
One developer have told me how is happy that he have build his C# .NET
APPLICATION (accounting for Germany customers) with XML middle-layer. So
Application code do not depend directly on database structure. Interesting
point of view I should say.

-- 
Best regards,

Ruslan Zasukhin
VP Engineering and New Technology
Paradigma Software, Inc

Valentina - Joining Worlds of Information
http://www.paradigmasoft.com

[I feel the need: the need for speed]