[Mulgara-dev] 100k models?

[David Wood]

On 12 Apr2007, at 17:22, thomas wrote:
> --On 12. April 2007 08:51:34 -0500 Paul Gearon <gearon at ieee.org>  
> wrote:
>>> Like Paul, I too prefer just TQL these days since the "i" has
>>> become an unnecessary adjective.
>>
>> Maybe we adopt this as an official position?  It would mean  
>> cleaning up
>> the docs.
>
> tucana, kowari, northrop-grumman, mulgara ... why not make a real  
> move and name it "MQL"? or "mQL" (looks better)? or "ssQL" (for  
> "semantic store query language which resembles SQL")? or "§QL" (the  
> same, but cooler)?
> sorry, couldn't resist ...

Yeah, that's not a bad idea.  It would deviate from the TQL used by  
Kowari and Tucana users, but Mulgara's query language is evolving  
away from there anyway.

> btw: when the decision was made to make TQL look like SQL, what  
> were the alternatives? were there any close contenders? felt an  
> SQLish design just as natural (or nearly so) as those or was it  
> chosen mainly for "political" reasons (acceptance, easy adoption etc)?

The decision was very much political, to ease acceptance, as I  
recall.  We were creating a database, and so should develop a query  
language that was easy to learn and absorb.  I still wish we could  
develop a JDBC/ODBC driver, but it just won't really work.

The syntax borrowed from Squish, RDQL, SQL and a bit of Bash.  It has  
evolved, of course and still is evolving.

I think that we have shown good reasons for Mulgara to continue  
having its own language.  SPARQL is hardly mature and lacks key  
features (e.g. adding data to a store!).  Although I fully support a  
SPARQL implementation for Mulgara, I think we will be committed to  
having at least two query languages for some years.  That is sure to  
impact the architecture.

Regards,
Dave