Recently did a digital ad entirely in processing and then ported it to p5js. In my experience with it…
* processing is still much easier to work, even though p5js seems to catch-up. I did work with it like 10 years ago and it was already very easy
* it was very easy to spot bugs or missing implementation in p5js
* While an almost identical port was possible, it’s not 1:1 and takes learning some APIs
* p5js is not even close to
d3 or three.js in terms of performance
* tbh it seems to me lot easier to animate in web with modern css3 and some helper libs, than use p5js. In fact I would see a very straight pipeline from free vector apps into css with very little shading without p5js. Which makes me wonder how is p5js relevant in 2024, apart from educational tool.
* For things which involve pixel level compute there is no easy way to use the GPU efficiently
For me this whole 450k funding is quite bizarre, and in particular the fact it got directed to a not so popular framework.
Regarding performance, I once had an experience that I never bothered to look into. I made a cool little interactive simulation in the p5js sandbox; ran flawlessly. I then copy pasted it into a codesandbox, and it was choppy and abysmal. I checked the version numbers and how it was being included and couldn’t see anything obvious. Maybe codesandbox has some isolation that is taxing? Like I said, never looked into it.
I think this is missing them point of p5js. Khan Academy has brilliant kids courses on p5. A bright first grader can learn it.
P5 is not designed to solve the same problem as D3. It's for creative coding. It's what kids did on Logo, BASIC, or an a graphing calculator. For that, is brilliant.
It's for a weekend playing making digital art.
It's for teaching math and physics.
It's not something anyone could, would, or should use for any production code. You can't do a11y, i18n, responsiveness, or a million other annoyances that come with real-world coding. It's not a work framework.
It's there for FUN.
(And if you have kids and the to help them out, it's absolutely better than Scratch; marginally slower to get going, but no ceiling)
These are some great insights! Thanks for summarizing them. It will save me a lot of time in the future. I would love to see this digital ad! Is it public?
the p5js part was more challenging, as I had to only draw every 5th pixel every frame, in order to get some decent performance. but since this is a transition, you can't mention it.
the whole loop is not much, but took a good 6 hours to get together. then another 4-5 go rewrite it for web.
but if you mean beginners, why not stick with processing?
BESIDES - Processing is much more mature and more usable with electronics. Really the whole point of p5js and this grant is beyond comprehension for me.
* processing is still much easier to work, even though p5js seems to catch-up. I did work with it like 10 years ago and it was already very easy
* it was very easy to spot bugs or missing implementation in p5js
* While an almost identical port was possible, it’s not 1:1 and takes learning some APIs
* p5js is not even close to d3 or three.js in terms of performance
* tbh it seems to me lot easier to animate in web with modern css3 and some helper libs, than use p5js. In fact I would see a very straight pipeline from free vector apps into css with very little shading without p5js. Which makes me wonder how is p5js relevant in 2024, apart from educational tool.
* For things which involve pixel level compute there is no easy way to use the GPU efficiently
For me this whole 450k funding is quite bizarre, and in particular the fact it got directed to a not so popular framework.