Re: S.O.D.A. database Query API - call for comments
Date: Sat, 21 Jul 2001 18:01:46 GMT
Message-ID: <9cusnr$lot$01$1_at_news.t-online.com>
Vadim Tropashko wrote:
> >> >CERN needs to store megabytes of data in milliseconds.
> >>
> >> True. I bet those are just simple flat records, however. Why bother
about
> >> objects then?
> >
> >What do you want to bet?
> >The generated data is all but simple and flat.
>
> Do you imply that you have real time applications with objects and it
ships
> objects to sensors via Corba? :-) Anyhow if you have a set of sensors,
they
> measure just scalar values, right? I understand that you might want to
nest a
> flat vector of data into objects, as the vector components map into
containment
> hierarchy of the device. Or am I missing anything else?
Sorry, I am not informed about the technical internas but I am absolutely positive that object-orientation was essential for this project. You can probably investigate a lot over the web: www.cern.ch
Sensors (or whatever hardware) would of course deliver flat data.
Wild guesses:
Data might just be too much to store in real time. Pre-processing could be
used to create objects from possibly analogous data. I would assume that it
makes sense to link the output of multiple sensors. Sensor types might have
inheritance hierarchies. The 3D positioning of the respective sensor would
probably also be an interesting and variable factor.
Kind regards,
Carl
--- Carl Rosenberger db4o - database for objects - http://www.db4o.comReceived on Sat Jul 21 2001 - 20:01:46 CEST
