[
Lists Home |
Date Index |
Thread Index
]
[Miles Sabin]
>I see the graphs for doc-centric XML, but have you got any publishable
>evidence to back up your claim that data-centric XML has uniform
>element distributions?
I don't use XML too much for data-centric work (Excel, MySQL etc. are soooo
good at that.).
However, I have to hand an XML file of the form:
<Records>
<Record>
<ID>[number]</ID>
<ST>[date]</SD>
<SL>[text]</SL>
<LT>[text]</LT>
<LL>[text]</LL>
</Record>
...
</Records>
The TagShare analysis is pretty predictable:
Element Occurence
------- ---------
ST 842
SL 842
Record 842
LT 842
LL 842
ID 842
Records 1
Things get more interesting when you look at de-normalized tables, when you
"un-flatten" the model you create for RDB work. I suspect when you do that,
you get closer to a power law distribution.
Being an insufferable doc-head, I would of course argue that this points to
documents - not inter-twingled tables - being the true Tao of information:-)
Sean
|