Re: [boost] Re: Possible Boost addition; Ditto framework

From: Joel de Guzman (djowel_at_[hidden])
Date: 2002-11-08 20:28:13

Martijn,

I think your library can be a layer on top of Dmitriy Arapov 
and Arkadiy Vertleyb's RTL. It would save you lots of
time and you can concentrate on the features that
you are focusing on: Views. 

--Joel

----- Original Message ----- 
From: "Martijn W. van der Lee" <gmane_at_[hidden]>

> Hi,
> 
> Currently Ditto is working though there are some areas in which it can be
> improved and being a one-person-design there's bound to be some gaps, for
> this I would very much like the input of some more experienced C++ coders so
> I am indeed looking for team members or atleast some comments.
> As for comparing it to the Relational Tables these seem to be somewhat
> different reading from Arkadiy Vertleyb's mails;
> 
> - RTL uses expressions for sorting/filtering whereas Ditto uses functors.
> - RTL does not seem to automatically update it's views (or whatever it is
> known as in RTL terminology); when a record is added to a table (or view),
> it is not immediately available in all other views based on that table.
> - Ditto can stack views as to provide multiple levels of refinement of the
> underlying table.
> - RTL can mix multiple tables into one singular view, Ditto currently cannot
> (though the architecture is capable of this).
> - Ditto can assign events (using callback functors) whenever a record is
> added/changed/deleted and even when other events are triggered.
> - Ditto can keep a pointer to a single record by means of a tracker class
> which also plugs into the event mechanism, I don't know if/how RTL handles
> this.
> 
> I may be wrong here with regards to RTL and perhaps the concepts of both
> libraries can be mixed into one library, which would be ideal.
> I'm also under the impression that the goals of the libraries are different,
> RTL seems to want to be an in-memory relational database, Ditto tries to
> provide a means of easily sharing data without a database in a GUI
> environment.
> Performance differences are also unknown to me and it may be impossible to
> do a decent comparison if both libraries are indeed too different.
> 
> regards,
> Martijn van der Lee
> 
> > You very well know that I am scouting for an in memory
> > data base. I wish to use one in a future project. My
> > question now is when do you think Ditto will be completed?
> > Are you looking for other team members/developers?
> > Also, in terms of features, how does it compare to
> > other libraries such as the recently proposed Relational
> > Tables by Dmitriy Arapov and Arkadiy Vertleyb?
> >
> > Regards,
> > --Joel
> >
> > ----- Original Message -----
> > From: "Martijn W. van der Lee" <gmane_at_[hidden]>
> > Subject: [boost] Possible Boost addition; Ditto framework
> >
> >
> > > Hi,
> > >
> > > I've been quietly working on a framework which I think could make a nice
> > > addition to Boost. Currently it compiles on Borland C++ 5.5.1 and GCC
> 2.95.3
> > > (Dev-C++/MinGW) and seems to be stable (though the testcases are not
> 100%
> > > implemented yet).
> > >
> > > In short it is an in-memory generic table and view framework where a
> "table"
> > > class contains records similar to a database table and "view" classes
> > > contain sorted and filtered (by means of functors) version of that
> table.
> > > These views are updated dynamically so adding a record to the table will
> > > automatically make it available in all views.
> > > Furthermore, views can be stacked to allow multiple levels of
> refinement,
> > > single records can be selected by using a "tracker" class and events can
> be
> > > triggered on any change in the frameworks' classes by means of a
> "monitor"
> > > class.
> > >
> > > The original purpose of this library was to replace Borlands' VCL
> mechanism
> > > which requires data to be located inside the class handling the visual
> > > layout, this made it extremely difficult to have multiple separate views
> of
> > > the same set of data, the Ditto framework succesfully eliminated this
> > > problem at the cost of recreating the visual components but with the
> benefit
> > > of performance, memory requirements and especially code quality.
> > >
> > > The project is published at sourceforge; http://ditto.sourceforge.net
> but
> > > the site is a bit outdated and documentation is especially lacking, the
> > > current version in CVS is stable and can be downloaded from the project
> page
> > > at http://sourceforge.net/projects/ditto
> > >
> > > There are currently still some minor issues with the library (mostly the
> > > need to specify a dummy filter functor to views when no filter is
> required)
> > > but with some help (hint! hint!) I believe the framework could be in a
> > > finished state within weeks.
> > >
> > > regards,
> > > Martijn van der Lee
> > >
> > >
> > >
> > >
> > >
> > >
> > > _______________________________________________
> > > Unsubscribe & other changes:
> http://lists.boost.org/mailman/listinfo.cgi/boost
> > >
> >
> > _______________________________________________
> > Unsubscribe & other changes:
> http://lists.boost.org/mailman/listinfo.cgi/boost
> >
> 
> 
> 
> 
> _______________________________________________
> Unsubscribe & other changes: http://lists.boost.org/mailman/listinfo.cgi/boost
>