[Users] Remove default variable guess from Multipole

[Ian Hinder]

On 10 Sep 2010, at 21:49, Erik Schnetter wrote:

> On Fri, Sep 10, 2010 at 2:00 PM, Ian Hinder <ian.hinder at aei.mpg.de>  
> wrote:
>>
>> On 10 Sep 2010, at 20:49, Roland Haas wrote:
>>
>>> Hello all,
>>>
>>>>> When the Multipole thorn was first written at Penn State for
>>>>> debugging, we
>>>>> were using it with WeylScal4 and to avoid incompatibilities with
>>>>> early
>>>>> parameter files, when the "variables" parameter was added we
>>>>> decided that if
>>>>> it was left blank, it would default to the WeylScal4 Psi4
>>>>> variable.  This is
>>>>> not a good logical default, as the thorn provides generic mode
>>>>> decomposition.  I would like to remove this, and the patch is
>>>>> attached.  OK
>>>>> to commit?
>>> Would it be possible to make the default for the parameter
>>> "variables" to be "WeylScal4::Psi4r{sw=-2 cmplx='WeylScal4::Psi4i'
>>> name='psi4'}"?
>>> I would expect that for most (vacuum) runs done with the code in the
>>> Einstein Toolkit one would decompose Psi4 so this default would make
>>> sense. Also it would keep the thorn backwards compatible :-)
>>
>> Sorry - I only saw this message after committing the patch.  What you
>> are asking for, as far as I can tell, is the behaviour that was
>> implemented.  It has been printing the following warning for 3 years:
>>
>> WARNING: No Multipole::variables string has been specified.  For
>> compatibility, the current default is to use \"WeylScal4::Psi4r{sw=-2
>> cmplx=\'WeylScal4::Psi4i\' name=\'psi4\'}\" but in future, this
>> default will change to \"\".  Please update your parameter files to
>> include this new parameter.
>>
>> so I think you have had enough time to update parameter files :)
>
> I don't think Roland suggested to have this "silent default" where the
> thorn checks for an empty parameter value. I think Roland suggested to
> set this as the default value for the parameter, using the standard
> Cactus mechanism for this.
>
> Roland, this is a good idea.


Ah sorry I misunderstood.  I can see why this would be convenient.

I think it might cause a little confusion and breaks the orthogonality  
of the design of the various thorns.  I would prefer to leave the  
default as it is now, i.e. an empty list of variables, rather than  
couple it to WeylScal4, which is an unrelated (except by common use)  
thorn.  Of course, this isn't a hugely important point, so If there is  
strong support for Roland's suggestion, I would accept the view of the  
majority :)

-- 
Ian Hinder
ian.hinder at aei.mpg.de