Document vs. database

Jim Croft jrc at ANBG.GOV.AU
Fri Nov 30 08:23:22 CET 2001


> Am I missing something, but why do we keep on talking about states?  This
> is the last thing I think is interesting (double entendre here...)  States
> are frosting, observations - measurements, photos - make up the cake itself.

Is this a real issue or one of semantics?  When I write "leaf shape
ovate", does it mean I have observed the leaf and its shape most closely
matches that of ovate, or that I have observed the shape character leaf
and consider it to be in a state of ovateness?

The end result is probably the same...

But I think the issue is worth pondering... Perhaps Richard did not have
it quite right and perhaps we should be recording zero or more
observations which are made of a feature that has a certain character
one of the attributes (rather than the state) of which is present/
absent/misinterpreted, etc.  Maybe something like:

<observation sequence="3" basis="specimen" refereence="Stevens in NGF67432">
  <feature name="leaf">
    <character name="margin" value="serrate" state="present"/>
  </feature>
</observation>
<observation sequence="4" basis="literature" reference="Stevens 1998a">
etc...

or maybe it would be better to invert it and define a feature up front
and make one or more observations about it:

<feature name="leaf">
  <character name="margin">
    <observation value="serrate" modification="obscurely" state="present"
     basis="specimen" reference="Stevens in NGF67432"/>
    <observation ..... />
    <observation ..... />
  </character>
</feature>

There are numberous ways to moddel this sort of stuff, and has been
repeatedly pointed out there is no right way to do it, there are only
ways.  The challenge is to find one that is functional, comprehensive,
flexible, extensible, pragmatic and elegant.  And one that we can all
agree on...

> Probably getting old.

as are we all...  and we have to get this damn thing working before
our use-by dates are up...  :)

jim




More information about the tdwg-content mailing list