Re: [Miase-discuss] MIASE paper

[Jonathan C. <jon...@co...>]

Dagmar Köhn wrote:
> Hej Jonathan,
>
>> Rules:
>>
>>    1. Parts A and C are very similar. I would be inclined to include
>>       boundary conditions within A. Peter Hunter's comment refers to
>>       PDE-based models, where the same model (in terms of PDEs) can be
>>       discretised in many ways (even using different weak forms). How
>>       much of this should be considered part of the model, and how
>>       much part of the simulation algorithm, is debatable. This is
>>       also covered to some extent by 2.E. So it might be best to
>>       remove 1.C entirely, and just have an explanatory comment about
>>       PDE models.
>>
>
> Hmm, not sure I agree or not. In any case it would make sense to swap 
> the order of rules. Like 1B-1A-1C.
That sounds reasonable.
> Whether or not to merge 1A and 1C I cannot decide.
The mention of 'initial conditions' in 1A does imply time course 
simulations, which might be too restrictive (this goes back to the scope 
of MIASE).  What about boundary value problems, for instance?
>
> I'd propose:
> D. If a model is referenced as a piece of implementation code, then 
> all information needed to simulate it correctly must be provided.
>
>   1. Use of open code is encouraged as much as possible, as it is the
>      best way to evaluate the quality of a simulation.
>   2. If  using closed code (black box), then all information needed to
>      simulate it correctly must be provided. Several independent codes
>      must provide the same result.
The phrase "all information needed to simulate it correctly must be 
provided" is still repeated.  But then, it's important enough that it 
could  be worth emphasising.  I don't have a strong opinion on this.
>
>> Information on the Models:
>>
>>     * I'd be inclined to agree with MC's comment that little needs to
>>       be said about changing parameters. Perhaps just "Model changes
>>       do not only include atomic changes such as the assignment of new
>>       values (such as covered under rule 1.A.)."
>>
>
> I think mentioning some concrete examples makes it easier for the 
> reader. But if I am the only one we can well shorten the paragraph.
I don't think it's a big issue.
>
>>     * Information on the Simulations:
>>
>>     * Regarding parameter scans, perhaps "such as the range of
>>       parameters considered in a parameter scan"
>
> Could not find where this related to? 
It was related to the comment "find a better expression for step size". 
So change "That includes but is not limited to the simulation algorithm, 
or information such as the step size in the case of parameter scans, for 
example." to "That includes but is not limited to the simulation 
algorithm, or information such as the range of parameters considered in 
a parameter scan, for example."

Jonathan