[MOBY] Re: [MOBY-l] MobyRequest or CentralImpl ?

[Paul Gordon]

Then perhaps we should have some examples of that in the API (i.e. data 
*instances* with articleName attributes)?  Under the MOBY Data Class 
Ontology section, none of the examples have articleNames, even though 
the GO objects contain several sibling String elements... or am I 
missing something?


Mark Wilkinson wrote:

>Martin is 100% correct.  In fact, the articleName attributes of
>sub-objects really defines an ontology of their own, with the
>articleName being the "relationship predicate" linking the sub-object to
>its parent object.
>
>M
>
>
>On Fri, 2004-05-21 at 07:24, Martin Senger wrote:
>  
>
>>>This brings up a new issue though that requires a policy formulation on 
>>>the part of the community:  Suppose VirtualSequence had two Integer 
>>>members, how would these be distinguished in the XML (i..e the Length 
>>>field vs. the foo field)?
>>>
>>>      
>>>
>>   Articles can be named. The API says: "Articles are, optionally, named 
>>using the articleName attribute. This might be used if, for example, the 
>>service requires named inputs. The order of non-named articles in a single 
>>Input or Output set MUST not be meaningful.".
>>   Mark, is this correct answer? I am still not sure about calling 
>>services :-(
>>
>>   Martin
>>    
>>