[tcs-lc] Poa acroleuca var. ryukyuensis H.Koba & T.Tateoka

Sally Hinchcliffe S.Hinchcliffe at rbgkew.org.uk
Thu Mar 10 06:47:03 PST 2005


Bob wrote:

> In the case A) could someone please give me an example of "let the 
> questioner work out which was the record they want" when the questioner 
> is a software agent and not a person? If there is such an example, does 
> making A) meaningful require additional specification so that two agents 
> would always make the same decision? Is it problematic if they don't?

This is what worries me about A) - I can't see how it can be done. 
Although there might now be a way in TCS to indicate which are 
the 'answer' records and which are the 'related supporting' records 
because I know I raised the same issue with TCS to do with 
supplying data from IPNI. Jessie?

> Case D), or anything else with transient discovery mechanisms, makes 
> caching problematic. The resolver of the discovery mechanism has to 
> offer the same time-to-live guarantees as the data provider, which 
> probably only works if the resolver and provider have the same 
> administration or at least if the resolver and all providers agree on 
> expiration date policies---an unlikely scenario.

With D, the three records would be included as three LC records 
bundled up into one TCS record using internal links which are 
transient - e.g. running numbers generated during the query - and 
used only to make the links. So the data comes in a complete 
package and there's no need to go back to the provider for any 
more information. Does that answer your problem with D? The real 
problem with D as I see it is that it's the same as A - the receiving 
end has to work out which is the 'real' record, and which are just 
there to make the data complete.

> So, blissfully ignorant of any of the scientific impact of A-E, and 
> assuming that the English meaning of each of them is as it seems, I 
> would say C) seems to be the most congenial to software systems.
> 
Well there seems to be a consensus on the mailing list (!!!) that C 
and B are the best (they are not mutually exclusive)... which is 
good because that's the structure (as long as we update the id 
references to use resolvable ids) the current LC draft version uses. 

I think I'm going to quit while I'm ahead on this one :-)

Sally

*** Sally Hinchcliffe
*** Computer section, Royal Botanic Gardens, Kew
*** tel: +44 (0)20 8332 5708
*** S.Hinchcliffe at rbgkew.org.uk



More information about the Tcs-lc mailing list