Hacker News Reader

Comment by beaconstudios | Hacker News Reader

I hope darklang succeeds too, but I personally believe that visual programming is at least partly the way to go. This is because I firmly believe that the "most-correct" view of regular programming is that of a directed graph of computations, and directed graphs don't compress very well into text.

I'm building something visual to prove (or disprove) this concept, but I'm also thinking that power users would want a textual language for faster input, as GUIs are better for output where TUIs are better for input.

View on HN

I agree that visual programming can be great - my argument would be that starting from a model that we know works and making incremental steps towards something more visual is more likely to succeed, than attempting to jump all the way to something fully visual based

I'm a fan of what Observable is doing. Combine bubbles of classic code, with intermediate audio/visual output as you build up to full interactive infographics/demos.

Some types of code, hopefully less than more is just not very flow based and doesn't have a concrete visual representation.

in general I agree that incrementalism is the best way to improve a model. I'm just not certain how to do that with such a drastic change as going from textual to visual - but darklang seem to be doing a great job at that, and I do hope they succeed at their mission.

Thanks! We're very much of the opinion that we got into this mess we're in (gestures broadly) because everything has been incremental to date. That's how you get layers upon layers.

Yes 100% - my founding hypothesis was that because language progress so far has all been improvements on opcodes, that the idioms of programming are still based on the human catering to the computer rather than the other way around. It gets subtler over time, but we still have to, for example, emulate execution in our heads while coding, which is really really hard to learn.

Text is much more searchable. Any binary format is ultimately text. We use it for non-input, non-human-readable formats.

Graphs are great for showing relationships and high level information at a glance. Human can recognize visual shapes (this can apply to text as well but text as well. Many editors now show text minimaps).

But after a certain scale, the best we have for relationships is probably hypertext links and other uri references.

Dark actually started with a directed graph concept. It worked pretty nicely, but users (developers) hated it. We tried to get them to write a simple fizzbuzz, and it was very challenging, and ultimately we abandoned it.

Hey Paul, I'm a big fan of what you guys have been doing. Yeah I did see that you went the visual route early on and it didn't work out, but I think there's quite a large landscape yet to explore. I'm also aiming to work with people who aren't currently developers (ie the usual nocode market) - I think developers have already gotten over the hardest hurdles to learn textual programming and thus aren't interested in the velocity for simplicity tradeoff that visual programming usually involves. This is all hypothetical for now though as I'm still building the platform.

Reply on news.ycombinator.com