[Miase-discuss] [Fwd: Re: MIASE paper]

[Dagmar K. <dag...@un...>]

Hej Frank,

David disagreed on this paragraph from the motivation (one of your 
favorites in the paper as well):
"Imagine that a modeler wants to use the model, as it is, as a working 
module to build a larger system. In an ideal world he would want to 
retrieve the model from a model repository and explore it using the same 
setting the model's creator used, in order to determine whether this 
model is suitable for his needs. Currently this is not possible, and the 
user would have to put effort into understanding the deep mathematics of 
the model, verify with the paper to discover the simulation algorithm 
and implementation that were used, and how the parameters would have to 
be changed in order to arrive at the desired behavior of the model."

He proposed to rewrite it to something like (draft):
"In an ideal world the model development workflow would follow these 
steps: (1) query model repositories for existing work that can be 
utilized; (2) add in modifications and new model components; (3) apply 
the complete model(s) in one or many simulation experiments 
[Nickerson&Buist, Phil Trans 2009]. While some initial work has been 
completed in early investigations of the technologies which can be 
applied to achieve this [Nickerson et al Bioinformatics 2008; Nickerson 
& Buist Prog Biophys & Molec Biol 2008], the focus of the models used in 
these developments has been limited to cellular physiology. Effort is 
now required to extend the concepts developed in this earlier work to 
encompass the larger range of models addressed by MIASE. Achieving this 
goal will enable the above workflow to be utilized by model authors, 
model users, and  tool developers removing the barriers imposed by 
current workflows (such as the effort required for understanding the 
deep mathematics of the model, verification with the paper to discover 
the simulation algorithm and implementation that were used, and how the 
parameters would have to be changed in order to arrive at the desired 
behavior of the model.)"

My personal feeling is that the workflow description is quite specific 
and I am not sure I would like to narrow down the papers' motivation 
section so much to that workflow.
I also do not know if this workflow would be a commonly accepted view on 
the work in model development in general as I have seen quite different 
ones. And I'd suggest not to get into any discussions about model 
development workflows with any reviewers.
But I agree that we should mention part of this work.


What do you think?
Dagmar