Thanks Bill. A lot to digest. I ll run this by a documentum engineer. It sounds like a performance concern to pull the hub in... the unique id qualifier soundsOct 13, 2006 1 of 11View Source
Thanks Bill. A lot to digest. I'll run this by a documentum engineer. It sounds like a performance concern to pull the hub in... the unique id qualifier sounds intriguing yet due to its complexity in terms of input factors making up the id it could potentially be subject to frequent change/update...
Gesendet: 13.10.06 12:33:52
Betreff: [TaxoCoP] Re: Question regarding Taxonomy use and Metadata
Thanks for the compliment. I am no Documentum expert, so you are way
ahead of me on that front. But I can elaborate a bit more on the
Let's assume that you are using a taxonomy server or service in which
you can model these multiple vocabularies. Let's then assume that
there are cross-mappings between them, so that the taxonomy records
the fact that a given element in Meddra is mapped to an element in
MeSH, Snomed, home grown list etc.
Further, you have chosen to treat one of these vocabularies as a "hub"
and the others as spokes. Therefore, if element A in the hub is linked
to 3 elements in other vocabularies, when a fourth vocabulary is
added, it will inherit relationships to the others.
At tagging time you can call the hub and retrieve any or all of this
information - i.e the hub element, its attributes, the elements in the
other vocabularies to which it is linked, the vocabulary sources etc.
If however, there is a problem around parsing this data into the
document repository and storing it, (step forward a Documentum
expert!) you could choose simply to use a unique concept identifier
from the hub taxonomy, rather than the XML above. The attribute would
therefore be simple, and should leave your data definition unchanged.
But at the next runtime - information access, for example - you
retrieve all the mappings and attributes from the taxonomy, in real time.
Hope this helps!
--- In TaxoCoP@yahoogroups .com, Karin Schneider <kschnei1@.. .> wrote:
> Hi Bill,
> I can always depend on some good input from you! For my
clarification: You say: you get some xml to apply to the document in
Documentum.. . what do you mean? I assume you are not talking about
parsing xml into attributes because that would lead to the original
problem, right? If you are talking about storing it as part of the
document, how would that work? Obviously, I cannot have alterations of
the content triggered by meta data. That would cause a whole range of
other issues (business approval process, versioning, content integrity
etc.). I cannot store it as a rendition and have a unified information
access/search experience with other metadata sitting in the RDB. I may
be looking at this problem with traditional DMS knowledge and hence
overlooking important clues.... Would you mind giving me a bit more
> *Von:* TaxoCoP@yahoogroups .com
> *Gesendet:* 12.10.06 14:24:29
> *An:* TaxoCoP@yahoogroups .com
> *Betreff:* [TaxoCoP] Re: Question regarding Taxonomy use and Metadata
> If you use a separate taxonomy database, you should be able to solve
> this problem fairly easily. Here are some of the approaches we use -
> 1 Ensure that each taxonomy element has an attribute giving its
> vocabulary of origin. Then call the element name + attibute
> combination via an API. You will get some XML to apply to the document
> in Documentum. By the way, you should be able to create these
> attributes as a batch process on upload.
> 2 Use other attributes, where appropriate, to show equivalences. For
> example, MeSH provides UMLS identifiers. Each taxonomy element could
> store a set of these equivalences as attributes.
> 3 Use relationships between taxonomies in the taxonomy database to
> show equivalents, when the mappings are a constant.
> 4 Consider whether all of this information can be condensed into an
> identifier. This may not be possible (for example if you are exporting
> the metadata to external systems or 3rd parties) but if it is, you
> will certainly minimise the physical impact.
> Re RDF, all of the above data can be modelled in RDF triples, in XML,
> but I am not sure whether Documentum accepts RDF. You will have to
> provide Documentum with the XML in a supported format.
> --- In TaxoCoP@yahoogroups <WBR>.com
[mailto:TaxoCoP% 40yahoogroups. com], Karin Schneider <kschnei1@<WBR> .>
> > Hi all,
> > I'm working in the pharmaceutical industry and wrestling with the
> following problem:
> > For certain regulated documents we have to index an indication (e.g.
> Hypertension, Schizophrenia. <WBR>..) as metadata value. Now,
> where the indication metadata will be used for different values have
> to be used, e.g. Meddra for European Clinical Trial Reporting, SNOMED
> for Product Labeling etc. In addition there is a searchability
> dimension that may call for another value list - sometimes even 'home
> > What is the best way to ensure that you can associate multiple terms
> from multiple dictionaries with your document while maintaining the
> knowledge of the originating dictionary (name and value). In order to
> minimize impact on the physical data layer definition, is there a
> smart way to associate these values in one repeating attribute in the
> database instead of adding physical data fields? Has anyone resolved
> the problem and is willing to share how you did it?
> > Looking forward to your great ideas!
> > Thanks,
> > Karin
> > Viren-Scan für Ihren PC! Jetzt für jeden. Sofort, online und
> > Gleich testen!
*http://www.pc-<WBR>sicherheit. <WBR>web. de/freescan/ <WBR>?mc= 022222*
[http://www.pc- sicherheit. web.de/freescan/ ?mc=022222*]
> [http://www.pc-<WBR>sicherheit. <WBR>web. de/freescan/ <WBR>?mc= 022222
[http://www.pc- sicherheit. web.de/freescan/ ?mc=022222]]
> Viren-Scan für Ihren PC! Jetzt für jeden. Sofort, online und kostenlos.
> Gleich testen! *http://www.pc- sicherheit. web.de/freescan/ ?mc=022222*
[http://www.pc- sicherheit. web.de/freescan/ ?mc=022222]
Der WEB.DE SmartSurfer hilft bis zu 70% Ihrer Onlinekosten zu sparen!