Re: [Animl-develop] Update to JCAMP Converter

[Amanda C. <af...@ab...>]

> In JCAMP-DX there is a distinction made between TEXT and STRING
> values.
> TEXT essentially says it can be anything
> STRING represents pre-defined key words
> 
> so ##TITLE= is text
> ##DATA TYPE is string with for example INFRARED SPECTRUM pre-defined and
> published in the protocol.
> 
> I think things like SAMPLEID should have a pre-defined string
> to say this is the SAMPLE or BACKGROUND or REFERENCE or BLANK
> whatever to aid ease of processing.
> 
> For an IR:
> Things like SampleState with text as solid is probably not sufficient
> whereas if it was some defined string that might be more meaningful.


For all of the terms in any XML-based document we should be using the 
principles of the Semantic Web, which have been around for years now. 
That is, the terms you use within an AnIML document should be defined in 
an ontology, and which ontology you've used for any term should be 
stated. This gives the flexibility to use any term that you need, from 
any accessible ontology, but the term is still well defined, so that 
processing software can make use of it.

The existing ontologies for science are still under development, but a 
standard such as AnIML should be making use of their potential.
See for example: http://msi-ontology.sourceforge.net/index.htm

Amanda

--
Amanda Clare  http://users.aber.ac.uk/afc/
1851 Research Fellow
Tel: +44 (0)1970 628406  Fax: +44 (0)1970 628536
Dept. of Computer Science, University of Wales, Aberystwyth, SY23 3DB