Re: teamup for the great ideas (xindice pros & cons)

15 Oct 2003 19:54:54 -0000

A little bit off topic...
anyway,I had some experience using axis, the apache soap. Serializatoin by Beanmapping in axis is a little bit complicated in axis,when it comes to complex user defined data type/class.
 
soap, as a messaging protocol, is normally used when there is a need to transport the data over the network. maybe voiding axis, if you just need to serialize an xml object... Or you can use oo xml database.
 
does it help in any sense?

Henrik Vendelbo <[EMAIL PROTECTED]> wrote:
How about supported binding/serialisation/marshalling, whatever you call it.
----- Original Message -----
From: chasee
Sent: Friday, October 10, 2003 8:57 PM
Subject: teamup for the great ideas (xindice pros & cons)

To all:

it's nice that this topic finally leads to a "theoretical and experimental analysis of current xml db technology" type of thing.

I have summarize the great ideas you guys have proposed for the analysis and added in my own understanding and thoughts. Indeed it seems something that could contribute to the maturity of the xml db technology, if the we have more theoretical stuff stuck in. Shall we team up for a research topic if not enough for a project as steve suggested? Send email to me or steve if you vote for this idea...

** "theoretical and experimental analysis of current xml db technology"**

Z)Theoretical comparison
1* index structure and indexing behavior
2* memeory leaking handling
3* recovery strategy
4* XUpdate semantics


A) XML:DB tests ---------------
1* access times for adding documents graphed against total resource count
2* access times for deleting documents graphed against total resource count
3* access times for updating documents graphed against total resource count


B) XML:DB:XPATH tests ------------------------
1* access times for querying documents graphed against total resource count -for single document match -for multiple document match
2* perform more and more general query to find the maximum ResourceSet size which can be managed


C) Stress & Robustness tests --------------------------
1* repeatedly add a single smallish document with different id to get max resource count possible within a collection
2* update a single document repeatedly, while increasing it's size, to get max single resource size
3. increasing the collection size with a constant load util the instability of the db is observed.



============
thanks,
chasee

Do you Yahoo!?
The New Yahoo! Shopping - with improved product search


============
thanks,
chasee

Do you Yahoo!?
The New Yahoo! Shopping - with improved product search

Reply via email to