> Working on Bel was hard but satisfying. I worked on it so intensively that at any given time I had a decent chunk of the code in my head and could write more there.
I'll have to say reading this makes me feel a bit sad for pg. It seems that he worked on Bel extensively for four years, and the end result was something which appeared on the front page of Hacker News for one day and then disappeared. I haven't seen it mentioned in any community which are actively working on programming language design (e.g. Rust, Zig, TypeScript). Maybe he's happy with the result regardless of how useful people have found it, but surely it must be somewhat disappointing to see it go unnoticed by?
> I'll have to say reading this makes me feel a bit sad for pg. It seems that he worked on Bel extensively for four years, and the end result was something which appeared on the front page of Hacker News for one day and then disappeared.
This is not unlike a doctoral dissertation. Invest years of your life and effort, produce nice results, have people cheer for you after you successfully defend it and then nobody cares for the results anymore.
Well a more forceful version of your counter-argument would be that Claude Shannon laid the foundations for the information age in his thesis, rather than linking to some obscure Lisp work. But really, that wasn't their point -- most theses go nowhere, the few that do are obvious exceptions to that rule.
If writing Bel did not make pg better at Lisp and programming generally I will be very much surprised. Producing something that anyone cares about for any length of time is more than almost any school project achieves.
The Bel part felt deeply in tune with the rest of Paul's life - it's something he did because it was interesting to him and he wanted to.
If anything, Viaweb is the odd thing out, since it was pursued with the explicit intent of making money (though I expect the whole affair was fascinating from the inside anyway).
Not everything has to be evaluated in terms of its popularity, "success" or "impact".
Another PG (Phil Greenspun) had this comment on Bel:
If you weren’t persuaded by the existing 100+ dialects of Lisp that have been created over the years, Bel from Paul Graham should change your mind and lure you aware from the dark and tedious arts of C and Java.
After you’ve saved bigly in development time on your next project, you can thank me!
As a Lisper, I had not been able to understand what Bel was when it was released, and still do not understand it after reading the shorter explanation in this post.
Does anyone else feel the same ? Perhaps this project just needs to be "sold" a bit better.
As somebody who liked reading about Bel but gave up after the first paragraph of the source code, something to note is that (a) most new ideas are failures, (b) I’m happy for Mr pg for trying new things, I like when people try to do something new, without regard for success, (c) great ideas take time to be appreciated in full - it may very well be tha t Bel will be a great success, but once understood by people far smarter than me and (d) (importantly for me), I appreciate the reminder that the path least travelled can be the most rewarding and it reminds me not to lose heart in my own crazy projects. I really liked the quote that the intentions on why we do things is important - the going will get rough, but we will persevere if we do it for the right intentions (our own happiness).
Bel will most likely not gain traction; but the fact that Mr pg spent four years and enjoyed his time developing it, makes it to me a highly successful outcome for himself:) outward Success is not a requirement for a successful project, the only thing matters is whether we achieved our internal goals on it.
Exactly. I think that's well captured if you extend the quote above:
> Working on Bel was hard but satisfying. I worked on it so intensively that at any given time I had a decent chunk of the code in my head and could write more there. I remember taking the boys to the coast on a sunny day in 2015 and figuring out how to deal with some problem involving continuations while I watched them play in the tide pools. It felt like I was doing life right. I remember that because I was slightly dismayed at how novel it felt. The good news is that I had more moments like this over the next few years.
That certainly sounds like a successful endeavor to me.
Give that pg is someone who can walk away from running YCombinator to work on other stuff for fun, he probably doesn't care about maximizing utility or impact at this point.
There's no onomatopoeia in the English language for the kind of laughter that I'm currently emitting. There have been plenty of technical examinations of Bel [0]; it's just not novel.
There's no techincal examination of Bel in that essay. It's a dismissive aside at the end, based on threads the author perceives in Graham's intellectual history.
There's a couple good comments in the discussion thread. In particular, there's a comment halfway down the page which points out that Bel's original introduction [0] has only a few specific characteristics (metacircularity, long contemplative period before implementation, formal methods) and that those have been core concepts in the Lisp community for decades.
If you mean https://lobste.rs/s/jec21l/thought_leaders_chicken_sexers#c_..., that's not serious criticism. (And like the other article, is obviously motivated by extrinsic animosity. Articles like that get reactions based on how people already feel. Those who share the animosity like the hit, those who don't don't.)
pg's idea for Bel was to express existing programming language constructs in the style of McCarthy's original Lisp, so to complain that it doesn't introduce new ones misses the point. You can't make serious criticism without knowing what the project was trying to achieve. (IIRC, Bel does actually contain a couple of unusual constructs for a Lisp, but not because pg was trying to invent any. They came up as side-effects of making the program clearer and smaller.)
Similarly, it makes no sense to complain that Bel isn't being used as a programming language. That's not its intent. Its intent is to be the minimal executable explanation of what a programming language is, in the way that McCarthy's Lisp explained what computation is. Obviously that would be unusably slow as a real-world platform. The important thing is that it runs at all.
I'll have to say reading this makes me feel a bit sad for pg. It seems that he worked on Bel extensively for four years, and the end result was something which appeared on the front page of Hacker News for one day and then disappeared. I haven't seen it mentioned in any community which are actively working on programming language design (e.g. Rust, Zig, TypeScript). Maybe he's happy with the result regardless of how useful people have found it, but surely it must be somewhat disappointing to see it go unnoticed by?