Re: voigt error... followup

On Fri, Nov 16, 2007 at 13:50 -0500, rgibson_at_email.domain.hidden> I think I found the problem.  The voigt.vtherm parameter goes
> negative during the fits (because I allowed for a minimum value
> < 0 in set_par).  I think it wasn't obvious because fit_counts
> would crash before updating the vtherm parameter when it went
> negative.

In this case, "crash" is the wrong word isn't it?

For example, if a car's "check engine" light came on, it
wouldn't normally be called a "crash".

> 
> Suggest robustifying voigt.vtherm and voigt.fwhm to not halt
> ISIS if evaluated when negative?
> 
> Thanks,
> Rob

I don't think this would be an improvement.

When an attempt is made to evaluate a function outside its
valid domain, I think the function should generate a fairly
severe error.

For example, when computing a real-valued expression involving
sqrt(x), I think a severe error (an "exception") should be
generated when x<0. The exception serves as a warning that
something has gone very wrong.

For what it's worth, you can use slang's try/catch/finally
construct to handle exceptions of this nature.  For example,
see:

  http://www.s-lang.org/doc/html/slang-15.html#ss15.2


Thanks,
-John

----
You received this message because you are
subscribed to the isis-users list.
To unsubscribe, send a message to
isis-users-request_at_email.domain.hiddenwith the first line of the message as:
unsubscribe