IRC channel logs


back to list of logs

<civodul>hey ho!
<zimoun>is it possible to check if some entry are in without many lines of code?
<civodul>"guix lint -c archival" does that
<civodul>i guess one could reuse that code
<civodul>Timothy's preservation-of-guix must have something similar
<zimoun>for instance, disarchive-db is expose and it is possible to check the entry
<civodul>(BTW, i'll start an rsync of the database on berlin later today)
<zimoun>I already did it on my machine here, in case. :-)
<civodul>ah yes, the HTTP endpoint is this:
<zimoun>thanks for the endpoint.
<civodul>this one is 404 but there are examples at
<zimoun>yeah, it was an example. :-) I found the one I wanted. And 404 means not yet in. :-)
<zimoun>Arf, I subscribed for Julia webminar about 1.9 blabla. The link is… Zoom! And one needs to accept a lot of “terms of service”. Now, not sure I am still interested. :-)
<civodul>heh, let'em know
<zimoun>Ahah, Cascad is « the first certification agency for scientific code & data » and « A cascad certification allows researchers to signal the reproducibility nature of their research to their peers »
<zimoun>Ok, it sounds cool, isn’t it?
<zimoun>Let check the first item of “table of certifications”
<zimoun>This one:
<zimoun>The beautiful “Reproducibility Certificate“
<zimoun>Ok, let see the “Code & Data”. It points to Zenodo
<zimoun>Then the .zip file is empty.
<zimoun>Ahah! That’s a yet another certificate I trust! :-)
<civodul>certification for the win! :-)
<civodul>there are also cases where they don't even claim to publish code & data:
<zimoun>Ahah! Even better! :-)
<zimoun>I am sure they claim it falls under
<rekado>the execution report is also great
<rekado>they keep qualifying things
<rekado>right off the bat: “The results have been compared to those displayed in the original notebooks, which vary very slightly
<rekado>from those in the manuscript. This is due to a later change in the code made to simplify it.”
<rekado>“There are a few differences with the original manuscript: […]”
<rekado>“The results are the same for couples with 1 or 2 children. For those with 3 or 4 children, they’re a
<rekado>bit different (10^-2).”
<rekado>“Almost identical results.”
<rekado>“the results are similar.”
<rekado>and so on
<rekado>this deserves the highest reproducibility rating “RRR” aka “Perfectly reproducible”
<zimoun>Yeah, RRR… it’s a comedy Fench movie!!! ;-)
<zimoun>Ah, last the certification is costly
<civodul>those examples are impressive, rekado
<civodul>they call it not-for-profit and say it has support from CNRS
<zimoun>it seems a spinoff of
<civodul>there's a UK-based one too, forgot the name
<civodul>berlin doesn't let rsync traffic out, so i'm rsync'ing over Tor...
<rekado>can’t you use ssh for that?
<civodul>i could, but it just felt easier, and it's reasonably not-slow :-)
<civodul>(i always have to think twice when i use "ssh -L"...)
<zimoun>civodul: Sabrina from SWH pointed to me:
<zimoun>« I was shocked at how easy it would be to reject or hide criticism of the editorial process. There should be greater transparency and other measures of accountability over editors, senior authors and reviewers. »