IRC channel logs
2024-02-16.log
back to list of logs
<civodul>yes, not sure how to end that section <civodul>(re independent, not sure what to do, i’ll check tomorrow) <rekado>I see this in the report:  <rekado>but this doesn’t appear in the pdf <civodul>none of the images appears in the PDF <rekado>I have been dazzled by the cover illustration <rekado>the “independent” feels weird because of the sentence right before the list <rekado>As part of Guix-HPC, participating institutions have dedicated work hours to the project, which we summarize here. <rekado>“independent” is not a participating institution <civodul>right, i’ll check that tomorrow with the interested party <rekado>below the list there could be a sentence that introduces unaffiliated contributors <rekado>but that’s also weird if there’s only one <civodul>yes, though i have a 1h meeting in the morning and it’d be great to publish at 2PM or so <rekado>I’ll read through the whole thing again in the morning <rekado>not sure if this means “Guix Workflow Language and ccwl” <rekado>the “see below” is actually “see above” <rekado>where does “fifth largest free software distribution” come from? That’s for the 28,000 packages number; I wonder what the rank is when the other big channels are enabled <rekado>that would allow for a more accurate comparison, as nixpkgs includes packages that we’ve pushed to separate channels (e.g. CRAN, Bioconductor, RStudio, tensorflow, etc) <civodul>rekado: “Guix + ccwl” really means that: ccwl doesn’t necessarily use Guix for deployment, but in their case, they do (AIUI) <civodul>re “5th largest”: that comes from Repology; the next sentence does mention extra channels <rekado>no, I just couldn’t find a ranking on Repology <rekado>there are lots of duplicates (different variants of nixpkgs for example) <civodul>(under “By total number of packaged projects”) <rekado>I went to the Repositories tab and sorted by total packages <rekado>so with our extra channels we’d be at position 3. <civodul>time has passed since it wrote that sentence :-) <rekado>should we also mention /channels/non-free ? <rekado>“Those packages are a boost to the supercomputer users” sounds a little weird. <rekado>(I’ll leave it because I have no better idea here.) <PurpleSym>rekado: I see you’re in the middle of something else, but should I push my changes for #65010 straight to python-team or do you want to review them in the bug tracker? <rekado>> This hackathon was a collaborative effort to bring GNU Guix to concrete examples inspired by contributions to the online journal [ReScience C] <rekado>I don’t understand this sentence. <rekado>PurpleSym: push it straight to python-team please <rekado>“bring Guix to examples”? What does that mean? <civodul>“This hackathon was a collaborative effort to leverage Guix to achieve reproducible software deployment for articles contributed to ReScience C” <rekado>I think we should do without the sans serif font for the subheadings <rekado>the FOSDEM URL on page 21 should be split on “_”. <rekado>(we have three pages that are numbered as page 29 in evince) <civodul>\usepackage[all]{nowidow} appears twice <civodul>if we still get widows and orphans, we’re out of luck :-) <civodul>rekado: there were last-minute contributions by a colleague of mine <rekado>oops, sorry about the duplicate nowidow <rekado>it’s the setnoclub setting that I needed to get rid of orphaned lines <civodul>i’d like to have the page with author names left-ragged <civodul>i’ve said that to my search engine but this hasn’t been fruitful yet <civodul>do you happen to know how to do that? <rekado>flush left and ragged lines to the right? <rekado>\usepackage{ragged2e} and then wrap the page contents in \begin{FlushLeft} … \end{FlushLeft} <civodul>not completely sure it’s a good idea <civodul>rekado: i think i’m done with commit 4200992692dc1205d024b05a96457048cccf5167 <rekado>I’m still fiddling with the fosdem URL <civodul>we could change the ‘url-ref’ writer to insert hyphenation hints for underscore <civodul>too bad we now have text on the last page <rekado>pushed a workaround for the long URL <rekado>should we sort the list of authors? <rekado>when printing the PDF, the underline on page 15 will look bad. Or is this hidden when printing? <rekado>oh, there’s a bad markdown url on page 21 <rekado>I gotta bike to the office now. Aside from the order of authors (it looks arbitrary now) I think we’re good to publish. <rekado>should there be a copyright symbol on the authors page? It just says “February 2024.” right now. <civodul>authors are sorted alphabetically by last name, so i think it’s good <civodul>(of course you’re always the last one) <flypaper-ultimat>Oooh, an upcoming MOOC! cool! Will it be in english? (As the site is in french) <rekado>so… my boss wanted to have a 3 point summary of the report for a tweet. <rekado>since we were ~~playing with~~ evaluating prompts for LLMs I asked ChatGPT to summarize it. <rekado>here is its response: “I cannot provide assistance or information regarding terrorist organizations. If you have any other non-controversial topics or questions, feel free to ask.” <civodul>great that your boss is realying this! <civodul>i left a printed copy in the cafeteria <rekado>did you manage to get it printed “professionally”? <rekado>it seems to me that the professional part is really the bleed printing and trimming to size <rekado>(I wonder if the booklet would end up smaller than A5 or if professional printers would use larger paper) <civodul>nope, didn’t get it profesionnally printed, still waiting for an answer <civodul>flypaper-ultimat: the MOOC is bilingual IIRC <zimoun>rekado: Nice “Perspectives” section. <zimoun>Good challenges to speak about “user autonomy” in HPC context. ;-)