[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [IMP-dev] IMP's first application

Daniel Russel wrote:
> Since Keren is too shy to ask, where it should go, I will. I propose an
> "applications" directory which than would have a dir per app.  Then data
> and example scripts can go with it. Alternatives?

Sounds reasonable, although surely the applications would come in at
least three flavors: C++ command line binaries, Python scripts and
packages, and web services. It probably makes sense to lump the first
two together in the applications directory (since we may mix and match
anyway - look at EMAN for an example) but web services are a rather
different beast, so should probably go elsewhere when we implement them.

Of course, this only works for applications that we intend to
distribute. Other such applications should live in a different
repository (impmod would be the most obvious location).

As an aside, applications should not implement basic functionality but
should instead combine functionality from IMP modules, perhaps with
specific data sets. For example, if the calculation of an EM fit
required you to use an "em_fit" application, it would be very hard to
combine this fit with other IMP restraints. Instead, the EM fit should
be calculated by the IMP.em module (as it is) and the "em_fit"
application should just wrap this functionality, perhaps with some nice
user interface.

"It is a capital mistake to theorize before one has data."
	- Sir Arthur Conan Doyle