IRC channel logs
2023-01-20.log
back to list of logs
<rekado>zimoun: the x-debbugs-cc must not have worked. I didn’t receive any notification about #60885. <zimoun>arf, maybe I screwed up something. :-) <rekado>I’ll make a change to the description of r-roptim <rekado>I find pyproject-build-system to be doing a better job at actually running tests <rekado>zimoun: guix lint found something: r-scistreer@1.1.0: description contains typo 'This packages', should be 'This package' <zimoun>rekado: indeed, a better idea to have pyproject-build-system. <civodul>incidentally, i'm disappointed Guix-Jupyter wasn't met with more enthusiasm in Jupyter/RR circles <rekado>perhaps these circles are just too big? <rekado>I think it’s a common problem with very large, diverse communities <rekado>sometimes you don’t make the right amount of splash in the right part of the pond and the waves don’t manage to reach those who would be interested <zimoun>civodul: in Grenoble and then later, I discussed with Jupyter aficionados and mentioned Guix. Bah, the setup with Guix appears to them too complex, and it is also “a bit too fragile” (sometimes, it crashes). <civodul>zimoun: you mean Guix-Jupyter crashes? it's possible there are bugs indeed <civodul>rekado: heh, i like that pond analogy, that's probably what happened <zimoun>well, I have no investigated more with them. :-) My impression about Jupyter aficionados was they are not ready for the Guix shift in the first place. Maybe, these 2 difficulties are not really ones. <zimoun>well, the trend is to move all in the browse via WASM. <zimoun>In Grenoble, a person from QuantStack who is in ome Jupyer Board explained that the future is to have all in the browser. <zimoun>Arf, the slides are not available. <civodul>ah sure, but i don't subscribe to that point of view <civodul>it's not the future i want to live in anyway <civodul>in reality, "the future" will have less computing anyhow <civodul>we should think about removing layers, not adding more <zimoun>yeah, I guess I imagine the same “future” as you. <civodul>but hey it's Friday so let's be lighthearted :-) <zimoun>just, you asked why Jupyter folks are not so much about Jupyter-Guix. Mayne, because not the same future. <civodul>and for Jupyter, i can see the appeal of doing everything in the browser <civodul>no more fiddling with kernels: just compile compilers to wasm, have them produce wasm as well, and do all that entirely client-side <zimoun>yeah, maybe Guile could compile for the browser. ;-) <civodul>it's been discussed a number of times <civodul>there was even a promising GSoC project (for JS at the time, not wasm)