Dan Scott @dbs@code4lib.social

This has returned nothing for hours:

twarc filter 'laurentian -"Laurentian elites" -"Laurentian Elite" -"Laurentian consensus" -from:@LaurentianNews'

This works nicely and shows results within the past few minutes that should have been caught by filter:

twarc search 'laurentian -"Laurentian elites" -"Laurentian Elite" -"Laurentian consensus" -from:@LaurentianNews'

Am I doing something wrong?

November: report a console error due to an SVG file with an SVG element multiple ID attributes.

February: support says "We have replicated the error but have you noticed any functional effects?"

Me: "No, it prevents rendering ~100 SVG images, I can't tell if any functions have been affected.

Please, just fix it. I've diagnosed the problem for you, your devs should be able to resolve the problem with a single commit of a single line of code."

I am so angry right now.

Our administration has been lying to us for years and now they want to ignore our collective agreement and bi-cameral agreements on how the university shall operate: https://www.ey.com/ca/Laurentian

Err, math is hard. 2 weeks later, for a total of 3 weeks.

Should have just dumped the entire MARC dataset from Evergreen to remove it as a bottleneck. Oh well. Next time.

1 week later and my script has compared almost 2M bib records, finding about 13,000 mismatches for our staff to check and possibly correct. Still churning. Maybe I should have parallelized it? But I'm not sure if either Alma or Evergreen would have been very happy with that.

https://code4lib.social/@dbs/105561504929288751

In the initial report, I included JS console error messages pinpointing the problem, the URL from which the malformed XML comes, a sample of the malformed XML, and an explanation of what an unescaped & character means.

Everything that a reasonable tier 2 dev should need to take swift action. Not the kind of action that dismissively replies "yes the UI shows you that there wasn't a match"

Gah!!!

Wrote this in reply to a dismissive support response:

Any developer who understands the tiniest bit about XML would be highly disturbed to see an unescaped "&" character in an XML text node. Please find another developer to look at this problem report. Whoever looked at this one either didn't read past the second sentence, or lacks a basic understanding of XML. There is a basic defect that very badly needs to be fixed, both for technical reasons, and for usability reasons.

Ah that feeling when you've been focused on all the wrong things all day then realize that you forgot to take your meds and, yeah, that checks out.

Better living through automation: generate your updated dates, too!

The most recent stanzas shown are from 2021-01-14, so why is OCLC's page claiming it was last updated on Oct. 24, 2019?

And why aren't they consistently using ISO-8601 date formats? :)

People are so stressed out by receiving a library notice, they apparently don't notice that the due date is 2021-04-21 and just hit reply promising to return the item tomorrow.

I guess we've taught them to associate any communication from the library as bad news, so now that we're sending out weekly "Hey, just to let you know, here's what you currently have out" notices, it's freaking them out.

Saw a lengthy article had been proposed for deletion if no action was taken within 7 days. Then noticed it had been deleted after 4 days.

I reverted the deletion and then checked the proposing user's very active contributions--it looks like most of what they do is propose articles for deletion and then delete them.

Sure, sometimes we need to prune, but my instincts lean towards pitching in and improving deficient articles. So this felt shocking to me.

Just finished WikiPedia's WikiEdu training modules, and while they're fantastic, I find it ironic that the modules are not themselves editable.

I'd like to correct simple mistakes I've found, e.g. changing "Assigned trainings" into a heading on https://dashboard.wikiedu.org/training/instructors/new-instructor-orientation/dashboard-resources-tab or typos like "possibile" elsewhere.

But there isn't a feedback / report problem link on each page. Luckily, it does request feedback once you've finished the training. So I'm using that :)

I love @pluralistic but I also love Mastodon's ability to filter phrases. Just added 5 filters to drop the repetitive posts that he tacks onto every one of his threads.

If it was a single, long post that had the links to his books & newsletter at the end, that would be fine. But my feed fills up a little too quickly; filters help tame the beast a bit.

Also that document is 5 years old and shows a completely different UI for the Admin Tool. And nothing analogous in the new UI. Good job Elsevier.