Quantcast

Hat-Observe N+M Counter?

classic Classic list List threaded Threaded
3 messages Options
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Hat-Observe N+M Counter?

Barzille
Hi,
I am wondering if there's no more M+N Counter. In the documentation is an explanation that it shows the number of evaluated and nonevaluated functions when using :info. Is there an option to set this option? I am using hat 2.04

Maik

--
Maik Barz

Wilhelmplatz 4 - 24116 Kiel

www.Barzille.de
_______________________________________________
Hat mailing list
[hidden email]
http://www.haskell.org/mailman/listinfo/hat
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Hat-Observe N+M Counter?

Colin Runciman
Maik,
> I am wondering if there's no more M+N Counter. In the documentation is
> an explanation that it shows the number of evaluated and nonevaluated
> functions when using :info. Is there an option to set this option? I
> am using hat 2.04
The M+N counts in the :info tables of hat-observe are still shown just
as before.

You mention 'evaluated and nonevaluated' applications, but that is not
what M+N counts represent: counts of this form are used to indicate M
applications that were demanded but not successfully reduced; and N
reduced applications.  Applications whose results are never demanded are
not counted at all.

Regards
Colin R

_______________________________________________
Hat mailing list
[hidden email]
http://www.haskell.org/mailman/listinfo/hat
Reply | Threaded
Open this post in threaded view
|  
Report Content as Inappropriate

Re: Hat-Observe N+M Counter?

Malcolm Wallace
In reply to this post by Barzille
"Maik Barz" <[hidden email]> wrote:

> I am wondering if there's no more M+N Counter. In the documentation is
> an explanation that it shows the number of evaluated and nonevaluated
> functions when using :info. Is there an option to set this option?

You are right, the (m+n) count of (reduced applications + unevaluated apps)
has been removed.  I could not find anything in the commit log messages
to say why.  Would this be a useful feature to re-instate?

There are actually three counts available:
  * reduced applications
  * applications entered but never completed (blackholes)
  * applications that were never demanded
The original (m+n) count included the first two of these.  We reasoned
that the third count was uninteresting, because if a value was never
demanded, then it had absolutely no effect on the computation.

Regards,
    Malcolm
_______________________________________________
Hat mailing list
[hidden email]
http://www.haskell.org/mailman/listinfo/hat
Loading...