Re: [guppi-list] Re: categorical data in Goose

["Asger K. Alstrup Nielsen" <alstrup diku dk>]

> > 1) Is there any good reason why the categorical type should not be
> > dynamic? 

[Problems with consistency of statistics that rely on a specific
categorical type.]

> I really think that the best and safest policy might be to only allow
> you to create auxilliary data structures from "locked" categories.
> And if this is the policy, you might as well lock them at cache-time,
> or everyone will always be forgetting to lock their categories.  It
> will be bad enough remembering to call complete_category()...

Should we not solve this problem in a different way?

The problem is that this auxilliary data structure should not *reference*
a categorical type, but the categorical type should become *part* of the
auxilliary data structure.

The auxilliary data structure should keep a Categorial for reference,
and interpretation of the data it contains.  And in order to make sure
that things are consistent, this copy of a Categorial should be declared
const.

I.e. instead of letting the categorical type lock itself, the auxilliary
data structures that use a categorical should lock them.

The responsibility of keeping the data up-to-date should not be distributed
out to the Categorical type, IMO.  The responsibility of the validity of
the data from a given class belongs to that class.
For this contingency table to be able to certify the validity of the results
it presents, it should take the appropriate steps it self.

Maybe there is something wrong with this view-point?

> > 2) Would there be any foreseeable problem with deriving it from the DataType
> > abstract base?
> 
> Probably not.  I'd have to look at the DataType again and think about
> if there would be any problems.

I don't think there will be any, but maybe it was a design decision.

> > Regarding the GooseSet:  Last week, I did some work on this myself, but I
> > haven't committed it to the cvs yet.  I'll try to integrate my setup with yours
> > when I find the time.
> 
> Shouldn't be too hard, given how minimalistic mine is...

I'll hopefully have time tonight.
I won't touch the Categorical though [while the discussion is still ongoing].

Greets,

Asger