[MOBY] [MOBY-l] New Object Registration
Mark Wilkinson
markw at illuminae.com
Wed Aug 20 21:36:33 UTC 2003
Aha... the API is ambiguously worded... it should say "with the
exception of the Classes representing primitives, **and classes derived
from those by ISA relationships**.
In my mind, PubMed-MEDLINE ISA String, and therefore by inheritence has
all the qualities of a String, including the ability to have textual
content. But you are right, this is not necessarily obvious from the
documentation.
I will update the API to be more explicit about this.
Cheers!
Mark
On Wed, 2003-08-20 at 15:27, Vijay Narayanasamy wrote:
> >From MOBY API website:
>
> ------------------------------------------
> The value of the Object element is ignored with the exception of the Classes
> representing primitives (e.g. Integer, Float, String, etc.; discussed in the
> Class Ontology section below). For example, in the following object
>
> <Object namespace="GenBank/gi" moby:id="163483" > this has no value
> </Object>
>
> the text "this has no value" is ignored by both client and service.
>
> -----------------------------------------------
>
>
> That is the reason I have a "String' sub object for the document content to
> pass it to a "future service" (say a text mining tool).
>
>
> Can I pass the content of the document without having a String sub object?
>
>
> --Vijay
>
> -----Original Message-----
> From: Mark Wilkinson [mailto:markw at illuminae.com]
> Sent: Wednesday, August 20, 2003 3:00 PM
> To: Vijay Narayanasamy
> Cc: mobyl
> Subject: Re: [MOBY] [MOBY-l] New Object Registration
>
> Excellent! Very MOBYesque!!!
>
> Only one **small** comment :-)
>
> On Wed, 2003-08-20 at 14:16, Vijay Narayanasamy wrote:
>
> > Output <moby:PubMed-MEDLINE namespace="PMID" id="12925988">
> > <String namespace="primitive" id="">
> > ...CONTENTS OF PUBMED DOCUMENT IN MEDLINE DISPLAY FORMAT...
> > </String>
> > </moby:PubMed-MEDLINE>
>
> > Let me know if can register object class "PubMed-MEDLINE" (ISA text-plain,
> > HASA String) at the main MOBY.
>
> This is *almost* the correct structure, but not quite. Since you are
> planning for PubMed-MEDLINE objects to inherit from text-plain, there is
> no need for the String sub-object; text-plain ISA String, therefore the
> content of your PubMed-MEDLINE object can be a string, verbatim, without
> any sub-XML-elements. In addition, **if** you were to create an object
> such as you describe above you would need to give the String element an
> 'articleName' attribute, which you are missing at the moment (but since
> PubMed-MEDLINE ISA String, there is no need to do so).
>
> Congratulations on your services! They are **exactly** the kind of
> thing we are hoping MOBY will do!!
>
> > p.s.: I still do not have "collection" input
>
> Is there a reason for that? Can I help? I know it is a pain in the butt
> because you can't test these using the dumb CGI client program, but
> if that is the "correct" way to build your service, you should probably
> do so anyway - eventually we will build a better client program.
>
> M
--
Mark Wilkinson <markw at illuminae.com>
Illuminae
More information about the moby-l
mailing list