DData revision

[JP Bernardy]

--- Wolfgang Jeltsch <wolfgang at jeltsch.net> wrote:
> Am Sonntag, 14. März 2004 18:31 schrieben Sie:
> > [...]
> > > And why does Seq.concat have the type
> > >     Seq (Seq a) -> Seq a
> > > instead of
> > >     [Seq a] -> Seq a
> > > which would be more consistent with the other
> modules?
> >
> > I gave it some thought, and concluded that it is
> more
> > important for Seq to be consitent with itself than
> with
> > other modules. This really should be solved with
> type
> > classes, yet we chose not to introduce new classes
> with
> > DData.
> >
> > Moreover, Seq is an instance of Monoid, so
> 'mconcat'
> > can be used.
> 
> But why does LambdaSeq.concat have the type
>     [LambdaSeq a] -> LambdaSeq a
> instead of
>     LambdaSeq (LambdaSeq a) -> LambdaSeq a
> then?

I kept the old sequences (implemented as lambdas) as
LambdaSeq, but did not work on it any more, so it is
rather not up to date. I thought to remove it
eventually. Do you think it is any worth keeping?

Cheers,
JP.




__________________________________
Do you Yahoo!?
Yahoo! Mail - More reliable, more storage, less spam
http://mail.yahoo.com