[MOBY-dev] Changing only the description of an ontolgy entry

[Pieter Neerincx]

On 27 Jan 2009, at 16:07, Paul Gordon wrote:

> But in Moby, the description is what tells you the meaning.  The  
> user will infer the appropriateness of using a term based on that  
> free text.  Whether that inference is reasonable or not depends on  
> the user and the person who created the term.  I didn't say that the  
> description is necessarily *correct* or precise, but that it *is*  
> the usage contract in Moby.

Yep, I totally agree. The description is what gives you the meaning,  
especially for the primitives: String, DateTime, Float, Integer and  
Boolean. The fact that something is a string is not extremely  
informative. So we add an articleName attribute. Let's say I use a  
String primitive to store the name of a species. So now we have  
<String articleName='species'> ... </String>. That still doesn't tell  
me if I should use the systematical Latin name like for example Homo  
sapiens or the common name like human and in the case of a common name  
it doesn't have to be in English anyway, there could be case  
sensitivity issues, etc.... So there are a million ways to use this  
object. If this object was designed properly the description will take  
any confusion away and make it 100% clear what grammar to use for the  
species names. If the description doesn't specify it you'll end up  
trying to find out how the service works by painstaking trial and  
error. Note that the namespaces are not helping here as they cover  
only the context of the id attributes, but not the CDATA content of  
primitives nor the context of the element names.

It's virtually impossible for BioMoby Central to figure out if you are  
only trying to correct a typo / add a more meaningful description or  
whether you are changing the meaning. The only way to make a shortcut  
as compared to taking "the long way" to update an object with many  
dependencies is if it's changed/hacked by someone with "raw" access to  
the SQL database, which basically means we got back to whether or not  
to curate Moby Central...


> Andreas Groscurth wrote:
>> Of course the description is important for the understanding and  
>> the use of the datatype - but as its free text where everybody can  
>> add any information (and therefore also less information) i wouldnt  
>> count this as the *essence* of the datatype. If we would talk about  
>> a description ontology or other controlled vocabulary... I would  
>> agree. You could also say that random characters do not give an  
>> implied meaning - thats basically what the description is.
>>
>> Just as an example what i mean (actually this shows it for the  
>> description of a service) - if you have a service called  
>> "runBlast"... the description about is "Runs a Blast".
>> Does this give you any implied meaning ? what version of blast,  
>> against what database, what blast program etc etc...
>>
>> Sorry - this is a general issue ... a random user based text should  
>> not be seen as the essence of any element in the ontology.....
>>
>> And my suggestion in the way i want to change them is basically add  
>> new information, being more specific or erazing typos... I know  
>> that allow this would also enable people to completly change the  
>> description of a datatype... but as we allow people already being  
>> that much I wouldnt dare to give this opportunity.
>>
>> Cheers
>> Andreas
>>
>> Paul Gordon wrote:
>>>
>>>> I agree here with Martin ... i dont want to change anything which  
>>>> will change the LSID and therefore which might change any  
>>>> dependencies - in my case its ONLY the description.
>>> I respectfully disagree: the description is the *essence* of the  
>>> data type.  Otherwise the data type is just a random assortment of  
>>> fields with no implied meaning in the real world, and therefore  
>>> useless.  If you change the description, you change the version,  
>>> even if you think its minor, others may not.
>>>
>>> I agree that you should be able to change it without deregistering  
>>> though, that's a major flaw...
>>> _______________________________________________
>>> MOBY-dev mailing list
>>> MOBY-dev at lists.open-bio.org
>>> http://lists.open-bio.org/mailman/listinfo/moby-dev
>>
>>
> _______________________________________________
> MOBY-dev mailing list
> MOBY-dev at lists.open-bio.org
> http://lists.open-bio.org/mailman/listinfo/moby-dev

-------------------------------------------------------------
Wageningen University and Research centre (WUR)
Laboratory of Bioinformatics
Transitorium (building 312) room 1034

Dreijenlaan 3
6703 HA Wageningen
The Netherlands

phone:  +31 (0)317-483 060
mobile: +31 (0)6-143 66 783
e-mail: pieter.neerincx at gmail.com
skype:  pieter.online
-------------------------------------------------------------