[go] Fwd: [Go-database] Synonym length

[Michael Ashburner]

I agree with Amelia, the database should not dictate content.

Michael
On 26 Oct 2007, at 17:01, Jennifer Deegan (nee Clark) wrote:

> Hi,
>
> It seems to me that if we have the option of being comprehensive in  
> something that can be done systematically like this then it is well  
> worth doing. I'd hate to be standing up in front of an audience in  
> the coming year and telling them that we have added 9,991 out of  
> 10,000 chemical synonyms, and that we left 9 out because they  
> wouldn't fit. It seems to go against what we're doing somehow. If  
> Chris is happy to stretch a point, I'd be keen to keep them in (and  
> encourage the RSC to keep on sending them).
>
> Jen
>
>
>
> Midori Harris wrote:
>
>> In fact, it's for the RSC's text mining system that the synonyms  
>> were requested and added in the first place.
>>
>> (There were 9 over the limit; I took them out yesterday but could  
>> put them back, as Chris says he's increased the limit in the  
>> meantime.)
>>
>> m
>>
>> On Fri, 26 Oct 2007, Jennifer Deegan (nee Clark) wrote:
>>
>>> Hi,
>>>
>>> They might also be useful for reasoning and natural language  
>>> processing applications where a human would not be involved to  
>>> spot that the text is very long.
>>>
>>> Jen
>>>
>>>
>>> Amelia Ireland wrote:
>>>
>>>> Back in Gotham City, Harold Drabkin wrote:
>>>>
>>>>
>>>>> I would doubt someone would search on such a synonym.  Our  
>>>>> synonyms should aid in searching for the right term.
>>>>>
>>>>
>>>> As a general rule, though, we should not let the database structure
>>>> dictate what information the GO holds; if more cases come up  
>>>> where there
>>>> are lengthy synonyms, we should consider allowing longer entries  
>>>> in the
>>>> synonyms field to accommodate them. With some of these chemical  
>>>> names, it
>>>> might be that people see them in a paper or online somewhere,  
>>>> and they
>>>> just cut and paste them into the search box in AmiGO (or  
>>>> whatever GO tool
>>>> they are using). They probably wouldn't stop to consider how  
>>>> long the name
>>>> of the chemical is.
>>>>
>>>>
>>>
>>>
>>>
>
>
> -- 
> Jennifer Deegan nee Clark
> EMBL-European Bioinformatics Institute
> Gene Ontology Consortium
>