[Reprozip-users] Reprozip vs CDE?

Rémi Rampin remirampin at gmail.com
Sun Oct 30 23:58:52 EDT 2016


2916-10-29, Jonas Wagner <jonas.wagner at epfl.ch>:

> One thing I'd be curious about is if you know a way to work around
> licenses. Say I have an experiment written with Matlab (or some other
> proprietary software). I'd assume it would be illegal for me to pack it up
> with Reprozip, wouldn't it?
>

It would definitely be illegal. ReproZip is primarily meant for people's
own software or opensource packages.

You might be allowed to distribute some proprietary software if you leave
out your license file or equivalent, you should check what the conditions
are. In any case you might run into said software's copy protection
mechanisms, which ReproZip doesn't try to avoid.

In the case of MATLAB, I think there is a way to "export" scripts as
stand-alone programs that don't require MATLAB to run, and you can
redistribute these.

The other thing I've wondered is about differences in the environment. For
> example, I have a few experiments that are tuned for the 40-core servers
> where we run them. I guess I could pack those up with Reprozip, but it's
> not clear to me currently if I could do something to help users with
> different hardware configurations reproduce the results.
>

You will probably see differences in performance/run time, but the software
should still run on a difference machine. The biggest issue we have in term
of hardware incompatibility is GPU drivers (where basically the new machine
has to have the same hardware, and in case of Docker, the same driver
version...)

[...] And again: thanks for doing the cool work. I wish we'll see more
> reproducible benchmarks in science in the future!
>

It is our hope as well. Thanks for your interest and feedback!

Best
-- 
Rémi Rampin
ReproZip Developer
NYU Tandon School of Engineering
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://vgc.poly.edu/pipermail/reprozip-users/attachments/20161030/7c2fbe86/attachment.html>


More information about the Reprozip-users mailing list