Re: [xml-dev] XML aggregation question?

[peter murray-rust <pm286@xxxxxxxxx>]

At 12:26 27/08/2006, Robert Koberg wrote:
>Hi,

This is very useful and it's good that there are members of XML-DEV 
who have had experience.

>On eXist:
>
>the last time I tried eXist, while testing by adding/deleting 
>documents/nodes and frequently restarting the (tomcat) webapp that 
>eXist was embedded in, somehow (???) caused thousands of 
>unidentifiable fragments to be created and added to the DB. Perhaps 
>because it was trying to save some kind of context/session info - I 
>don't know. It was discussed on the list

I assume the eXist list, not here

>  by more than one person who experienced it, but could not be 
> reproduced by the developers.
>
>I also experienced hangs in the response from eXist that was also 
>reported by other people. Also not able to be reproduced by the developers.

I think we also have similar concerns but at this stage it's a 
question of experimenting

>I lost confidence for a specific need I had at the time. I still 
>have hope for it - it does seem to be the best bet out there.
>
>>>(Why I would want to see another Open Source CMS is kind of like 
>>>needing to see another lawn mowing pattern.)
>>Unless we hear otherwise there are only two Open candidates - eXist 
>>and XML-DB.
>
>For your situation eXist and XMLDB are both open and free (as in 
>price). For someone who does not wish to open their source, eXist is 
>the only one that is free (as in price). I don't mind paying for 
>something, but XML databases seem to be more of a 'cross you fingers 
>and close your eyes' kind of thing. Tamino just seems way to high 
>priced - it seems like alot of money goes into their salesforce. 
>XMLDB just pissed me off, but still too high priced.

The string "XMLDB" is causing problems, I think. I have found:

http://xmldb.sourceforge.net/

"Currently, it provides an xml persistence layer using the postgresql 
database. It is still by far the fastest library for xml data 
retrieval and xsl translation. However, it currently has limited 
search capabilities. "

C 2001-2006: This looks like a single-author effort. Latest sf jar is 2002

http://xmldb-org.sourceforge.net/

"XML:DB is an industry initiative formed by 
<http://www.smb-tec.com/>SMB GmbH, <http://www.dbxmlgroup.com/>the 
dbXML Group L.L.C and the <http://www.openhealth.org/>OpenHealth Care 
Group. XML:DB is also supported by a 
<http://xmldb-org.sourceforge.net/credits.html>growing list of 
organizations with interest in XML and XML databases. XML:DB provides 
a community for collaborative development of specifications for XML 
databases and data manipulation technologies. Along with each 
specification an open source reference implementation will be 
developed to validate the ideas put forth in the specification and to 
more rapidly drive acceptance of the specification in real products. "

There is no update to the page since 2003.

and:
http://sourceforge.net/projects/dbxml

which has no updates since 2004 and announces on
http://dbxml.sourceforge.net/


"Berkeley DB XML is an application-specific native XML data manager 
built on Berkeley DB, the world's most widely deployed data 
management engine. Berkeley DB XML provides fast, reliable, scalable 
and cost-effective storage and retrieval for native XML data and 
semi-structured data."

"dbXML is a Native XML Database (NXD). NXDs are databases that store 
XML using an internalized format for faster overall processing. dbXML 
was developed using the Java 2 Standard Edition version 1.4. "

This is very closely linked to SleepyCat so I begin to understand. It 
seems there may be an unmaintained, potentially out-of-date, engine 
on SF which just about meets the letter of the Open Source license, 
but is of little use and which encourages people to buy a commercial product.

...
I may have confused these various databases in my mails - the names 
are sufficiently similar


>So I have been sticking with the filesystem,

Yes - I am a great believer in the filesystem as a simple database engine.

>Apache's Lucene for indexing

Yes

>  and CVS or Subversion for version control.

Yes

>If I need a database, I hire a RDB consultant.

Seems a good strategy for you, but impossible for me. I have to 
deliver XML software at zero cost to chemists I have never seen and 
who don't have any money for the process. I am not without hope.

It looks like eXist is the only choice. This has been a valuable 
discussion - thank you.

>P

Peter Murray-Rust
Unilever Centre for Molecular Sciences Informatics
University of Cambridge,
Lensfield Road,  Cambridge CB2 1EW, UK
+44-1223-763069