Skip to Content.
Sympa Menu

xom-interest - re: re: [XOM-interest] Serializing objects that contain an instance of nu.xom.Document, part II

xom-interest AT lists.ibiblio.org

Subject: XOM API for Processing XML with Java

List archive

Chronological Thread  
  • From: "Brian Sam-Bodden" <bsbodden AT integrallis.com>
  • To: xom-interest AT lists.ibiblio.org
  • Subject: re: re: [XOM-interest] Serializing objects that contain an instance of nu.xom.Document, part II
  • Date: Thu, 19 Feb 2004 13:55:36 -0800 (PST)

>>Nils wrote:
>>Serializable is mostly used in distributed computing,
>>and since you're
>>using JMS, that's what you're doing too.

Exactly

>>In any kind of distributed environment, you should
want to limit you
>>dependencies and increase your interoperability.
>>Forcing the receiver to
>>use XOM is, IMHO, bad distributed design, considering
>>the nature of XML.

I agree, but it this case the receiver is part of the
same application. We are processing a couple million
docs a day and they all go through sort of pipeline or
assembly line if you wish. At each step (some step are
synchronous and some are asynchronous) specific pieces
of information in the doc are required. Some of these
steps for a given document can all happen in a matter
of milliseconds while some steps might involved hours
(Documents sometimes arrive in sections from different
external parties). All steps are synchronized using
message passing, for those steps close in time, not
having to parse the document again provides a
significant boost in performance. For those steps which
are separated in time by any significant amount of
time, say 15 min. we reparse the document. So, as you
can see it is not as clear cut as just saying passing a
DOM object or XOM object around is bad distributed
design. I some places we only pass a identifier that
allows a process to get its document from a distributed
cache, so even if communications don't involve passing
around serializable objects, they still need to be
serializable to be placed in a distributed cache.

>Is parsing it twice really a performance issue? More
>than the network
>latency itself? Do yourself a favor and lose the XOM
>dependency.

Yes, it is and see above.

>Nils

Thanks for the input Nils.

Regards,
Brian

Brian Sam-Bodden
President & Chief Software Architect
Integrallis Software, LLC.
http://www.integrallis.com




Archive powered by MHonArc 2.6.24.

Top of Page