Origin story of bh and Jens?

@bh and @jens seem to be fairly good friends and from my understanding, they work together on Snap!, but how did they meet? Was there a time when Snap! (or perhaps BYOB) was being built by one but not the other?

(@bh If you happen to have one of your really fun 20th-century computer stories in stock that would explain this, please share.)

P.S. Sorry for posting this under Computer Science, but I wasn’t sure where to post it and this seemed like the best fit since both of you can be reasonably described as computer scientists.

@bh Off-topic, but just as an FYI, your website’s broken:

There’s this interview with Jens, he explains how he got into making Snap! and how Brian met him.

That’s actually quite informative! Thanks, @d016!

Maybe this should be a FAQ.

Long ago, the group at MIT that had developed Logo, the first programming language designed for kids, made Scratch. When I came to visit, they showed me Scratch and the first thing I said was “how do you make your own blocks?” And they said “You don’t.” And then I said “well, how do you do recursion then?” And they said “You don’t.” So I said “what kind of programming language is that?” I was unimpressed.

Nevertheless, they made me come to the first Scratch Conference. So I organized a rump Birds of a Feather session called “Things you can do in Logo but you can’t do in Scratch.” One of the things I demoed at that session was Vee, the project (in Open>Examples in Snap!) that my friend Paul Goldenberg invented to introduce students to recursion.

Unbeknownst to me, one of the people who attended that session was Jens, who was at the time on the Scratch Team. I’m a complainer; Jens is a doer. He went off and added to Scratch the ability to write procedures. He called this Scratch mod “Build Your Own Blocks.” It was a proof of concept meant to convince the rest of the Scratch Team that this was something that could usefully be added to Scratch itself.

Meanwhile, Jan Cuny, a program manager at the (US govt) National Science Foundation, decided to fix the underrepresentation of women and minorities in computer science. When I was a kid, that was a problem in all areas of science and engineering, but by 2005 or so there had been a lot of improvement in most STEM areas, except for CS, as measured by the number of different flavors of high school kids taking AP courses in each area.

Although we lose women and minorities at all levels of education starting in middle school and continuing through grad school and faculty hiring, Jan decided to focus her efforts on high schools, for several reasons, one of which was that she had an important tool available: the College Board exams that high school students take before applying to colleges. This is important because in the US, unlike most technically advanced countries, there is no national curriculum. If you want to change something about schooling in the US, you have to do it by convincing one school district at a time, or one college at a time, except when it comes to college preparation, in which the College Board exams form a de facto national curriculum.

So Jan talked the College Board into creating a new Advanced Placement exam, to be called AP Computer Science Principles. It took some talking-into, because AP CSP was going to be different from all the other AP courses. Usually, the College Board develops the AP exams by looking at what all the colleges teach in their first course for students who intend to major in that field (or sometimes in adjacent fields, e.g., engineering majors take the same first physics course as physics majors do) and whatever the colleges mostly teach, that’s what goes in the AP exam. There was already a course like that in computer science, called AP Computer Science A (because there used to be an AP CS AB exam covering the first two college CS courses, but by this time they no longer offered AP CS AB because hardly any schools taught it and so hardly any students took it). But not all colleges offered the course Jan had in mind, which was a CS “breadth” course for kids in non-technical majors. And among the colleges that did offer such a course, there was no general agreement about what topics it should include.

Jan’s plan was that CS Principles wouldn’t just teach the skill of programming, but would also include an understanding of how computer technology affects everyone, including non-STEM kids, from Internet commerce to cell phones and secret profiling of individual people. The hope was that this would make the course more appealing to women and minorities than an all-programming course would be.

My Berkeley colleague Dan Garcia was very active in the early planning for this campaign, and Berkeley was chosen as one of five colleges to get NSF funding to develop pilot versions of a CS Principles course. (In the end, we were effectively two of the five, because one of the others, Tiffany Barnes, initially at the University of North Carolina and later at North Carolina State University, decided she liked our curriculum better than what she had proposed.) A later round of funding supported, iirc, 20 colleges developing CS Principles courses.

In the first round of funding, nobody knew what a CS Principles course should contain, so there was a lot of variation. Berkeley already had a CS breadth course, which was a programming course taught in Scheme, so that it would prepare students who decided they wanted more CS and would take our first course for majors, which used SICP, the best CS book ever written. But Scheme was famously offputting to non-techies because of all those parentheses. (That’s a silly reason to dislike a programming language; all the other languages had ways to bracket part of a program, whether it was {braces} or BEGIN…END, or, worst of all, counting the number of spaces at the beginnings of lines. But whatever.)

So our idea was to use Scratch. We figured, if it’s easy and welcoming enough for eight-year-olds, it’s easy and welcoming enough for English majors. Several of the other people developing CS Principles courses had the same idea, but we all ran into the same problem: You really can’t get very far into computer science if you can’t write procedures. Some other groups decided on what we rejected as a bait-and-switch tactic: teaching the first week or two in Scratch and then switching to some text language.

At this point I discovered a blog post by Jens about BYOB. I asked him if we could use it in our course, and he was glad to agree. That got us started, but we still weren’t really happy because we couldn’t write higher order functions (MAP, KEEP, COMBINE) in BYOB, which still lacked first class lists and first class procedures.

As I said, I’m a complainer, and Jens is a doer. I still maintain that my main contribution to Snap! was teaching Jens about lambda. That sounds kind of esoteric, but if you think about MAP instead of thinking about lambda you can see the power of the idea and why it’d be important to our curriculum.

So anyway, that started a year of intense collaboration between Jens and me, in which every day I’d complain about some missing feature in what was then still BYOB, and Jens would say “oh, I couldn’t do that, it’d be way too hard” and then I’d wake up the next morning and there would be the new version in my email. Finally we released BYOB 3.0 to universal acclaim, with ticker tape parades in every major city in the US and in Germany. Just kidding.

And by then we had actual teachers teaching our course in actual high schools, but we got a few loud complaints from a few teachers who said they couldn’t use a program called “BYOB” in their schools, because it would be taken as an endorsement of teenage alcohol drinking. So with great reluctance I had to ask Jens if we could change the name. We chose “Snap” because the subdomain “snap.berkeley.edu” was available, which wasn’t the case for several better names we considered. And I added the exclamation point because “Snap” by itself seemed so boring to me. And I guess I was influenced by the name of “Oliver!”, which is my favorite musical.

Thanks for sharing, I love these stories.

Oh yeah forgot to answer this. Basically all UC Berkeley Internet services were off the air for a while following a security breach, but it should be back to normal now.

These stories are enthralling for me to read! Keep pushing out more of these, I love to know more about the development.

Hehe, Brian’s right about him teaching me about lambda. I even gave a TEDx talk about this a while ago: https://youtu.be/MZhl7Ix8_xQ?si=D8U161ZDk7JOwLct

Security breach?! Nothing from Snap! was harmed, right?

Excuse my lack of knowledge in the subject, but what are MAP, KEEP, and COMBINE out of curiosity? I think I might remember MAP, I believe I used it to apply a mathematical operation over a list, but a little refresher would be nice.

Evil educators…


Thanks for sharing this story, bh! You know, at some point you should write a book, maybe with an HTML version, where it’s just a compilation of your stories. You’re a really great storyteller and it’s a shame that I can’t have your tales for bedtime reading.

They exist as snap blocks:

Oh yeah, I remember watching that. Very inspring stuff!!

@pppdud wants to know what are MAP, KEEP and COMBINE generally, not only as Snap!
blocks. Sort of, in Wikipedia articles, they are known as Map, Filter and Fold respectively…
[Map (higher-order function), Filter (higher-order function), Fold (higher-order function)] :eyes:

time to bedtime read stories about Snap! to my grandsons when I’m 80…

Lol.

“Back in my day, well I didn’t know much about this ancient 1960s computer, but some guy named Brian Harvey did.”

@bh Your website’s up again, but I think you accidentally left some of the HTML code as text:

Edit: I just realized that you wrote this page in HTML 2.0. Just so you know, there’s this really cool thing called HTML 4.0 that exists now and is supported by virtually all browsers from the new millenium onwards. /hj

Did you know there’a rhis cool thing called HTML5, which is the current version of HTML that is used by almost every single website out there, including this forum?

If you’re using HTML 2.0, then clearly you’re trying to acheive compatibility with something, so HTML4 is a better fit. HTML5 is a very recent language version that deprecates lots of useful tags and attributes, and not everything supports it.

Wow, that is quite interesting! Here I thought I’d already known most of Snap’s history. If I read correct, there is direct a line from you, Snap!, and the course I took last year?

It was released in 2008, which in the world of computing I can agree it’s recent, especially when it comes to compatibility, but it’s not “very” recent. Also, I’m fairly certain that everything that doesn’t support HTML5 is old deprecated programs that no one seriously uses. I have a bigger feeling that bh just made the website when HTML2 was the standard, and just hasn’t bothered to update it because frankly there’s not much reason to.