A tentative use case for machine learning in academic libraries

By Mark Eaton | Published: April 16, 2023

Being a subject selector in an academic library is pretty repetitive. I’m basically applying the same selection criteria to different materials over and over again. In my specific role, I’m almost always looking for books (and ebooks) that are for lower-division undergraduates/general readers; that are from reputable academic presses; and that fall within the subject in question. I have some other more subtle requirements that I won’t attempt to explain here, but the point is that I’m applying these same criteria every single time.

Perhaps we could train a machine learning model to do this work. If it watched me select books long enough, I bet such a model could even pick up on the finer nuances by which I’m choosing titles. It could then probably do a great job of applying those criteria. The relatively constrained nature of the problem might make this a feasible problem to solve.

Librarians would probably want to train the model with their own preferences for their subject areas. A human-in-the-loop approach might be sensible. Maybe the team at ProQuest’s Oasis (for example) would be willing to build us such a feature?

Posted in acquisitions, ai | Comments closed

Dawdling

By Mark Eaton | Published: April 15, 2023

I’m really taking my time writing some JavaScript for our library homepage. The current code (that I want to replace) is working fine, but that’s not the reason I’m dawdling.

I’m going slowly because I want JavaScript to move more slowly. In my opinion, the whole language can just slow down. I know that’s an absurdly Quixotic idea, considering how fast the JS ecosystem churns through ideas and code. But if open source has taught me one thing, it’s that you make your code how you want it, and others can take it or leave it. A very important part of open source is having code that meets our needs. If I want slow JavaScript, I’ll make it slow, damnit. (I know there are very real practical problems with this, such as ending up with codebases that are no longer supported, etc., but I’ll do my best!)

I guarantee that I’m not the only one who wants this. If enough people want a slower moving JavaScript, it can happen.

Posted in javascript, open source | Comments closed

Bilingual

By Mark Eaton | Published: April 5, 2023

I’m currently reading a book by Kyran Dale called Data Visualization with Python and JavaScript, and it’s super interesting. In my experience, it’s very unusual to find a fully bilingual programming book. This is a great example. Most technical books, for better or worse, focus exclusively on one programming language.

I’m still in the early pages, but I find the comparisons of Python and JavaScript really interesting. This is stuff I mostly know, but for some reason it’s fascinating to see it spelled out systematically and comparatively.

The tools described in this book are also very interesting to me. Long-time readers of this blog probably know that I have an ongoing interest in data visualization in both Python and JavaScript. I think this book hits the sweet spot between “stuff I already know” and “stuff I still need to learn”; it’s at the right place in the learning curve to be very constructive for me.

Anyhow, the second edition of the book is very new, so it’s (pretty close to) the latest info! Highly recommend.

Posted in books, javascript, learning, python | Comments closed

Code4Lib 2023

By Mark Eaton | Published: March 17, 2023

Code4Lib 2023 wrapped up today. I attended in person this year, as it was nearby, in Princeton, New Jersey. As always, it was really great. One of my favorite things about this conference is that I always go home feeling a lot more optimistic about my profession.

I was happy about the large number of newcomers at the conference this year. Having new faces in the room is really important for maintaining the vibrancy this community.

Also, I’m (finally) starting to feel like less of an outsider at Code4Lib. I’ve attended 4 conferences; presented at one of them; participated in the Slack; wrote an article for the journal; and even recently became an editor of the journal. You’d think I’d feel at home by now, but it has taken a long time. Maybe what has happened is that the importance of the social connections I’ve made are finally starting to overcome the impostor syndrome I started with.

Posted in code4lib, conference, imposter syndrome | Comments closed

Modern javascript on campus

By Mark Eaton | Published: February 24, 2023

JavaScript has changed a lot over the years. Old browsers are not equipped to handle new syntax. ES6 (in 2015) was a major change. That was a long time ago. But unfortunately, on our campus, we still have browsers deployed in labs that predate ES6.

Why is this? My guess is that old system images still get installed, without much consideration of the browser versions. And for some reason, the browsers are not permitted to auto-update. Why this is the case is mystifying to me, but I assume there are reasons.

This makes it hard on people building websites for our campus community. You can’t use modern JavaScript and expect it to work everywhere on campus. You can write old-timey JavaScript if you like, but this can be annoying to manage and maintain. There are very good reasons why we moved on from those old JavaScript ways.

Maybe you could write conditional code that accommodates old and new JS. This is possible, but becomes complicated quickly. Or you can lobby for newer browsers — this is probably the best solution — but you may face intractable campus politics, which can be very quickly dispiriting. At this juncture, on our campus, it seems like a hard problem to solve.

Posted in javascript | Comments closed

Vue 3

By Mark Eaton | Published: February 18, 2023

I had been meaning to upgrade our library website to Vue 3 for a while now. We had been on Vue 2 for some time. But there was a hard deadline for switching over, because Vue 2 reaches end-of-life at the end of 2023. Out of prudence I didn’t want to push up too close to that deadline, in the event that the switch might require more than a trivial amount of effort.

Anyhow, this weekend is when I finally made an attempt. It was a particularly auspicious time, because we are between semesters right now, so website traffic is way down. Of course, during the switch-over I did manage to break the production site for about 5 minutes, so I was glad to have picked a quiet weekend.

Aside from that 5 minute hiccup it went well! Because our Vue code is rather simple, the switch was pretty easy. Basically, I had to change how the Vue app is initialized, but that’s about it. Everything else worked perfectly fine when ported over to Vue 3. It feels good to be up to date with the latest version of Vue, and I am happy to no longer have that end-of-year deadline hanging over me.

Posted in vue | Comments closed

The last days(?) of the Kingsborough Twitter archive

By Mark Eaton | Published: February 17, 2023

It looks like it might be the end of the road for our bot that archives tweets about Kingsborough Community College. Recently, the official @twitterdev account announced the impending end to the free tier of the Twitter API. On top of that, Mr. Musk himself suggested that the new replacement basic paid tier should cost $100/month, which is wildly prohibitive for our library. There’s still a lot of confusion and uncertainty about what these announcements mean for bot makers, but the gist of it is that it is not looking good for our archive.

The archive has run more or less continuously since February 2015, which is quite a good run, in my opinion. I’ve documented some of the technical bits here, here and here. We now have a dauntingly huge collection of tweets, which will hopefully serve as grist for someone’s analyses down the road. I am tempted to have a go at extracting some insights from the data set myself over the coming weeks. For now, while I’m sad that the archive is probably ending imminently, I’m happy that it ran successfully for so long.

Posted in api, archives, bots, twitter | Comments closed

Guest blogging

By Mark Eaton | Published: February 1, 2023

For the next little while, I’m going to be guest blogging over at Choice’s new LibTech Insights blog. My first post is up already! It’s about Mastodon.

My intention is to keep Library Technology at Kingsborough going as well. Wish me luck!

Posted in meta | Comments closed

The ghost of Code4Lib future

By Mark Eaton | Published: February 1, 2023

I was complaining about the Computers in Libraries magazine website recently, for pretty obvious reasons. I wanted to link to an article I wrote for the magazine from my faculty bio page, but I can’t in good faith send people to that website. I mean, it’s fine that the site is old; it’s actually kind of wonderful in a retro way. But I can’t present anything there as my professional work … I just can’t.

This has got me thinking about how online and professional communities age with the rest of us. Computers in Libraries has its own demographic, for better or worse. I suppose that my generation has Code4Lib, but we’re all getting older at Code4Lib too, and I wonder how long it will be before Code4Lib starts resembling the Computers in Libraries magazine webpage.

To be more positive and forward-looking about all of this: what’s next after Code4Lib? I want to see what that community is up to.

Posted in code4lib, old | Comments closed

Become illegible

By Mark Eaton | Published: January 19, 2023

One thing I appreciate about Mastodon (and the Fediverse) is the desire among some users to not be legible. Some users don’t want to be seen, understood, or to have any “reach”. There’s a desire to not be visible, to not be widely understood, and to basically be baffling to outsiders.

I’m taking this idea of legibility from James C. Scott’s book Seeing Like a State. I’ve seen some Mastodonians explicitly reference Scott’s book. Others don’t refer to Scott directly, but are often speaking about the same thing.

There are numerous strategies for being less legible: having a private account, not posting to the public timelines, not threading posts that would make much more sense threaded, not using a real name, deliberately not explaining references, not providing context, sub-tooting, and so on. This is all behavior that is common to inward-looking groups elsewhere on the internet, but only on Mastodon have I seen illegibility as being the explicit, stated goal. It is a decidedly interesting and compelling goal, and I humbly submit it here for your consideration.

Posted in mastodon | Comments closed

  • Subscribe to this blog


     

    Rest assured that your email won’t be used for any purpose other than sending you posts from this blog.
Skip to toolbar