IRC channel logs
2021-07-08.log
back to list of logs
<civodul>TL;DR: a service for reproducibility based on Docker <PurpleSym>The main problem with their approach is that it’s an extra step before publication, not something you can use during development. <civodul>good point, i hadn't thought about it in these terms <PurpleSym>Yesterday I gave a very short (10 min) talk about reproducibility using Guix. One of the questions was: Can we run/convert existing research? And the answer was, unfortunately, “maybe”. <dstolfa>civodul: have you hit any other issues with guix-jupyter, or does it make sense to close the bug now? <dstolfa>i haven't hit any other issues so far <civodul>PurpleSym: yes, i started prototyping around that last year, but unfortunately haven't gone further <civodul>dstolfa: great that it works for you! in that case you can close it <civodul>i should make a release one of these days <PurpleSym>civodul: You mean like rekado’s guix.install? <dstolfa>civodul: alright, will do! thanks again for taking a look at it on such short notice <civodul>one problem is that we don't have a peer-reviewed paper on this topic <civodul>so they didn't bother mentioning it, unlike the Maneage work <civodul>(well, this one *is* peer-reviewed, but it's also not quite the right topic) <PurpleSym>Oh, I see. I’m very curious whether that one will be fully reproducible in 10 years. <maxwell_TGAP>rekardo_: exporting IIRC=~/Desktop/jup-data still produces the same error <civodul>PurpleSym: yeah, we can never know for sure