Re: Singular Value Decomposition (SVD) Ada Algorithm

[jpwoodruff@irisinternet.net (John Woodruff)]

aschwarz@acm.org (skidmarks) wrote in message news:<35f054ea.0410302259.7510a56e@posting.google.com>...
> > 
> > You can acquire the Drexel U Matrix Math items from  
> > http://dflwww.ece.drexel.edu/research/ada/
> > They are marked:
> > -- Copyright (c) Drexel University, 1996           
            
> > Their package  Generic_Real_Arrays.Operations defines a subprogram
> > "Singular_Value_Decomposition".
> > 
> > This might be what you want!
> 
> Thank you. I haven't looked closely but have looked enough (to say) it
> seems to be what I need.
> 
> By-the-by, any reason that you made it a generic package? 

Well, it wasn't I who did the making.  But I'll offer the
"conventional" explanation:  application writers are wise to select
specific types to represent the numeric quantities, and the generic
formal in such packages as "generic_real_arrays" is instantiated to
reflect those application-specific types.


<...> I developed a
> doubly-linked list. List management is a normal package body. The
> generic package provides space, in the generic body, and interfaces to
> the non-generic list manager. The effect is to have a no-cost generic
> package interfacing with a shared non-generic package. 

I'm not sure I followed that claim;  you might consider posting an
example (folks here are pretty constructive, so you won't get
pilloried :)

However when I read "generic package provides space", I interpret that
to mean maybe you're coercing lots of different "components" of some
list, using 'address' to manage memory (?).  If that's your
suggestion, I don't think I'll agree.

In any event, I don't think there is a "cost" argument in favor of
avoiding generics -- but I'm interested in the discussion.

John