%META:TOPICINFO{author="GarryJolleyRogers" date="1259118874" format="1.1" version="1.7"}%
%META:TOPICPARENT{name="ClosedTopicSchemaDiscussionSDD09"}%
---+!! %TOPIC%

Bob writes: "I think it is that BDI.SDD_ is more expressive than our current descriptive Schema, which is close to flat. Aside from a monolithic character list, it has a few things that are typed as taxon names, things typed as image names, lists of both such things, and a label attribute on such lists (the "type" of the list). Some of the list types are from an enumerated set, like "similar species", "host plants", and "nectar species" but the list can have any label, because in any case we leave the semantics of the label to the consuming application."

Gregor: This brings up the issue of "list characters", which is in fact an open problem. It has been proposed as an extension to DELTA in the New Delta proposals by Dallwitz et al. It may be indeed be convenient to have a free-form text list character as well as a simple text-block character. So far I have not brought this up, given our struggle in other areas. I am personally reserved (and undecided) about this. Most cases for lists are much better handles as categories, esp. given the option to go beyond the categories, add extra information ad-hoc as unconstrained text, and then rework the terminology by evaluating these bits.

However, especially the cases of geography, host/predator/prey/pollinator etc. you mention are not well handled this way, the number of potential items is huge (all plants on world for a phytopathogenic fungi description...). This could be conveniently placed in a list character. Yet, all cases I can think of are not very well handled as free-form text. Handling them such is convenient and quick, but not satisfying, since the names all refer to other, authoritative lists of names. So:

<ol class="compact" compact="compact">
<li>Can anybody provide an case for a list of text character, that does not depend on geographic, taxonomic, or other authority files?</li>
<li>Should instead of introducing list of text items, we introduce lists of resource items?</li>
</ol>

Issues:

<ul class="compact" compact="compact">
<li>Given our current structure, we unfortunately have to add another resource section, because the taxonomic names have in Lisbon been redefined as "Class names of things being described", and moved from resources to the new Entities section. Names of host/predator/prey etc. (in general organism interactions) are, however, not entities in this definition.</li>
<li>for each citation of a name, do we need an additional data item "typing" this name?</li>
<li>should this be a general character type, or a special category "organism interaction", that lists organism names, PLUS for each a category which interaction: predator, prey, obligate biothrophic pathogen, necrotrophic pathogen, pollinator, nectar visitor, ...?</li>
</ul>

-- [[Main.GregorHagedorn][Gregor Hagedorn]] - 10 Feb 2004

(Just a reminder: this topic is still unresolved after the Berlin meeting!)

-- [[Main.GregorHagedorn][Gregor Hagedorn]] - 28 May 2004