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

RE: An alternative formulation of thedocument-centric/data-cen

• To: Sean McGrath <sean.mcgrath@p...>, xml-dev@l...
• Subject: RE: An alternative formulation of thedocument-centric/data-centric XML divide
• From: "Kirkham, Pete (UK)" <pete.kirkham@b...>
• Date: Thu, 03 Jun 2004 12:57:52 +0100
• Thread-index: AcRJUgqJxM+eUX7IRjmqLbd1hm/f7AABXPBQ
• Thread-topic: An alternative formulation of thedocument-centric/data-centric XML divide

I would expect most data XML to be approx. power law too, as you are have more nested elements under each element (in the simplest case: db:1, table:t, row:t*(rows per table), field:rows*(f=fields per table)), and typically t and f are small (say 7), but the number of rows is large, so for ~n elements root:1, level1: ~7, level 2:n/7, level3:n. So perfect power law for n~=350. If the tag names for fields differ, then the frequency of each tag ends up more evenly spread out between tags on the same level, but the cumulative distribution should still be approximatly a power law between levels.

For more structured data then similar distributions can be seen- for each element that is a container, there will be a few properties, for each property that is multivalued, many child elements. Mixed text abbreviates away some of the multivalued property tags, but still the container-contained 1:many relationship should hold.

Doing a quick count on two data oriented XML files of the same data sets (mtbf of a fuel system)- one derived straight from a database, the other from an XMI instance of the same data (which uses attributes instead of nesting elements for some properties)- give the frequencies, which look similar:

XMI data (single value properties as attributes)
1,1,1,1,1,4,4,5,5,8,10,14,16,18,18,19,21,22,24,29,51,57,59,59,64,64,83,94,95,123,125,148,190,210,242,277,447,

XML data (properties as child elements, elements named after properties)
1,1,1,1,1,1,1,1,1,1,2,2,4,4,4,4,5,5,5,6,7,8,9,10,14,16,18,18,19,21,22,24,24,31,51,57,59,64,64,64,64,83,83,94,95,95,116,116,116,116,120,123,123,125,134,147,210,211,238,242,277,385,447,646,

And for a different data set (XMI model for an application):
1,1,1,1,1,1,1,1,11,14,14,14,30,30,30,43,74,74,101,117,132,148,148,169,169,447,447,595,595,595,652,680,


Pete

********************************************************************
This email and any attachments are confidential to the intended
recipient and may also be privileged. If you are not the intended
recipient please delete it from your system and notify the sender.
You should not copy it or use it for any purpose nor disclose or
distribute its contents to any other person.
********************************************************************

• Prev by Date: Re: An alternative formulation of the document-centric/data-centric XML divide
• Next by Date: Re: An alternative formulation of the document-centric/data-centric XML divide
• Previous by thread: An alternative formulation of the document-centric/data-centric XML divide
• Next by thread: RE: An alternative formulation of thedocument-centric/data-centric XML divide
• Index(es):
  • Date
  • Thread