I run a software book club
notes.eatonphil.com
notes.eatonphil.com
Now, the first challenge was finding fellow enthusiasts among my software engineering colleagues, who seemed to communicate exclusively in binary. During our daily stand-up, I announced, "Who here likes books?" I was met with a silence so profound that I could hear the hum of the office air conditioner. Undeterred, I added, "There will be snacks."
This got their attention. After all, the way to a programmer's heart is through their stomach. By the end of the day, I had three tentative members: Jake, who only read sci-fi, Priya, who was into self-help books, and Alex, who claimed to enjoy "anything as long as it isn't too long."
For our first meeting, I chose a conference room and sent out an email titled "Book Club Kickoff - Bring Your Appetite!" I brought in a selection of snacks: chips, cookies, and, because I was feeling fancy, a cheese platter. The turnout was surprisingly good. Even people who hadn't signed up showed up, lured by the smell of free food.
We decided to read "The Hitchhiker's Guide to the Galaxy" by Douglas Adams, which seemed to be a safe middle ground between sci-fi and humor.
Wise move. More people will come if they think you have punch and pie.
This has been my experience. I led a book club at one of my jobs and it was poorly attended and those who did attend hadn't read the book.
My guess is that there is just a low percentage of people who meet all the criteria: 1) interested, 2) have the time to read the book, 3) have the commitment to read the book, and 4) social enough to attend the book club meeting.
I understand that many people are busy and may not have time to read the book or have a schedule conflict the day the book club meeting happens.
So unless your company is huge, you'll not have enough to keep a book club going. There also may be pressure to avoid using company communication tools to promote the club.
1. Reading a whole book every one or two months works well when everyone has a regular reading habit. Without that people realize that they need to read a whole book in a few days before the meeting. That either led to poor quality discussions or people showing up without reading. 2. Reading book chapters every month was less demanding but the frequency of meetings was still too low for people to build reading into their routines. 3. At two weeks the frequency was high enough that it became a routine for the engineers. 4. We also reminded everyone one day before to read the article. Even if they had forgotten to read till then it is easy to find and hour in the day to read up. 5. If the reminder was too far in advance (say 3 days before) then people ignored the reminder.
I guess that wouldn't be unreasonable if the content is work related and certainly helps with attendance, but I'm not sure if that would fly within all corporations.
I've approved or started (or both) this sort of thing on company time before, at more than one place.
For context, this was typically early stage R&D, and many/most of the employees had some academic background, though that ranged from "decades ago" to "we just hired you after a masters degree".
In this setting, it's a pretty natural continuation of the common "journal club" approach in academic research groups. It spreads the effort around, helps the team stay on top of new work, and generally improves professional development - if done well. It does take a bit of effort to keep fresh.
The trick to make it sustainable is to make at 3-month calendar of covered papers from multiple sub-areas (e.g. data management to machine learning, new programming languages to compiler topics) and share that widely. Not every topic is relevant to everyone, so naturally each topics will only attract a subset of the crowd beyond a small number of open-minded champions, but that's okay.
While in theory, everybody thought it's a good idea, not everyone had time to read it, and when it came down to it, everybody just preferred to work on their regular tasks. In the end, nobody wants to risk delaying a promised deadline because of a book club.
So it all fizzled out within a month...
Maybe because I've been a reader my whole life, but I can't imagine getting value out of reading a chapter a week. It's rare for me to read less than 1 book per week. Even if it were a technical or educational book, one chapter at a time is a miserable way to read something.
(In a previous startup, we had cross-department tech talks, partly because we had some really specialized problems, so it was useful for, say, releng to hear about what the ML team thought was interesting, and vice versa, just for generally improving communication and reducing friction. We even threw in an occasional "lightning talk" session, which was popular, though it was as much for getting people comfortable giving talks so they could level up to bigger ones.)
One of the key (unexpexted) benefits of it was how it facilitated discussions about SW development practices between the different teams at the company. The content was usually good for giving us new ideas to try etc, but the usefullness of the discussions surprised me.
I've written more about it here:
Not sure what it is, but in general software books just don't do much for me.
Only one I truly liked and got a lot out of was Physically Based Rendering[1]. It had a great blend of explaining the concepts but also showing how to implement them, and for me the literal programming style worked very well in this case.
But apart from that I struggle with getting something out of software books. Not fiction though, I read tons of sci-fi (especially hard) in my 20s.
My guess is that a lot of those people who voted on Despair and proposed it were trying to show off how smart and educated they were but those type of people don't actually follow through because it's work.
So we switched to making selections in person at a session; if you can't come, you can submit suggestions, but the people who actually bother to show up make the final choices. We choose a monthly book for a whole year at a time, so we just do an hour-long discussion in December and pick 11 titles. In a monthly group you could probably do it in 10 minutes after the regular discussion.
Also, as the organizer feel free to put your thumb on the scale: pick 3 books you'd personally like to read, and ask them to choose one. You may want to be completely democratic about it, but people often appreciate someone else reading reviews and limiting the choices (avoiding the paradox of choice).
Here are the meeting logs and some statistics: https://susam.net/cc/log.html
Here are some posts and content pertaining to these book clubs: https://susam.net/tag/meetup.html
Quite similar to the original post (OP), we also had about 5-8 regular participants in our meetups. However, ours were virtual rather than in-person. Given that the author of the OP mentions 5-8 consistent attendees at in-person meetups, it's quite impressive!
Do you feel like the pace is ever an issue with technical book clubs, or is it actually just fine?
The pace was not an issue for me because we used to go really slow. We read every line of the book together, ensuring we understood each argument before moving to the next one. As the meeting host, I came prepared for the meetings and spent a lot of time illustrating proofs with additional details, perspectives, and examples (see https://offbeat.cc/iant/boards/ for notes).
This approach suited me well, as I need time to digest new mathematical concepts and form intuition around them. I hope it worked for others, but we didn't survey participants, so I can't say for sure.
Sometimes we picked some exercise problems, solved them, and walked through the solutions during the meetings. But most exercises were done offline by the participants in their own time.
I have written a lot more about the learning experience here: https://susam.net/one-hundred-meetings.html
Related, I have had mixed experiences with work book clubs. I generally like them as an outlet for my socialization needs.
Sometimes you get people seeking "brownie points" and they don't really participate.
Sometimes you get the contrarian who thinks they know better than the author and does not debate the topic(s) in good faith.
Sometimes you get eager people and have great discussion.
All that said, I think book clubs are very useful and valuable and generally have zero issue getting the time allotted at work for folks across org boundaries to participate.
However, the company's support is very important. Fortunately, my employer pays for the books and they even have a person charged with supporting book clubs, including ordering books for you and create a Confluence template for notetaking and attendance tracking.
(rather than actually getting an audio book version, I'll typically get the epub and use a third-party reader with a read-aloud feature. I mention this because audio books are often not available, and because first-party readers typically won't allow you to read books aloud specifically because they to upsell you on the audiobook)
I've switched a few times, and anything on an Apple platform will typically just use the speech synthesis APIs and have the same set of voices available: https://developer.apple.com/documentation/avfoundation/speec...
That's fair. This is definitely the main reason why my comprehension is lower and I have no solution for it.
Yes and: our study group use(d) approval voting to choose the next topic(s). With both suggestions and voting open to any one.
After closing nominations and before voting, we briefly chatted about each suggestion. No more than 15min, usually a lot less.
Topics (books, papers, libraries, whatever) were added to the queue, ordered by interest (votes).
The organizer calls another "election" as needed. To refill the queue, or if interest wanes, or if a book proves to be a stinker, or...
Please let me know if you find any. Happy to help start one too.
https://eatonphil.com/2024-understanding-software-dynamics.h...
Just curious what the frequency of meetings you found to be most effective, weekly, bi-weekly etc?
Thanks
2 chapters weekly worked well because the chapters were short, and were mostly brief overviews that discussed the pros and cons. (Also we had a ton of experience in the room, so that helped guide the discussion.)
If the chapters were dense I would have longer gaps between meetings to allow people more time to read and reflect.
Thanks! Really good read.
Basically a condensed version of this blog post. :)
There's also Eric Zhang's NYC paper reading group [2] that seems cool but I've never been to.
It's a good book though.
Completely worth it for the first part already tho.
So, luckily, no I don't think so.
I often find that the material in the classics (programmatic programmer, mythical man-month, etc.) are so ubiquitous that I don't get much from actually reading them.
I agree with you that most recommended books are so-so.
More so a traditional mailbox.
I wish read receipts never existed. It doesn’t make sense to know when I check my mailbox.