Lots of great ideas here. However I think this makes dependency management really difficult. For example, if you have a large function which is found to have an important bug in some old version, it seems like it would be really hard to identify the users of that code and patch the bug just for the affected cases.
I came up with this idea of storing programs in structural format aroun 2013 - originally thought its future just like you think. I later however found value in the text representation - just more structured languages, but still text. I think now that is future not this. Also learned that big-idea languages usually are not the future - they are useful experiments as full-size prototypes more often than solutions. -- But its really interesting to see something similar to what I imagined back in uni (and also talked to people about) come to exists ;-) Its also so interesting to see its "same core idea made differently" - also the distributed part is interesting and all where you went with this idea is actually pretty awsome..
As long as there is a text editing interface the code at some level its still just a 'more structured language'. The bad stuff happens when you try to prevent users from being able to express invalid code by requiring that they only express code as a self-validating data structure. You need to be able to put square pegs in round holes when programming. If for no other reason than to demonstrate why you otherwise shouldn't.
How often do names change? Really? A straw man leading away from the many upsides of Merkel trees for “versioning”. In actuality, I would even include the name of definitions into that Merkel tree itself.
![YoungBoy Never Broke Again - Missing Everything [Official Video]](http://i.ytimg.com/vi/dTYeMZGzOzw/mqdefault.jpg)
You guys are crazy, love it
I utterly love this. Btw, one thing to consider re runtime: WASM & WASI.
Seems to be very promising language with distributed features enabled inherently. Jumping in with nix setup. Thanks for the presentation
Lots of great ideas here.
However I think this makes dependency management really difficult. For example, if you have a large function which is found to have an important bug in some old version, it seems like it would be really hard to identify the users of that code and patch the bug just for the affected cases.
I came up with this idea of storing programs in structural format aroun 2013 - originally thought its future just like you think. I later however found value in the text representation - just more structured languages, but still text. I think now that is future not this. Also learned that big-idea languages usually are not the future - they are useful experiments as full-size prototypes more often than solutions. -- But its really interesting to see something similar to what I imagined back in uni (and also talked to people about) come to exists ;-)
Its also so interesting to see its "same core idea made differently" - also the distributed part is interesting and all where you went with this idea is actually pretty awsome..
As long as there is a text editing interface the code at some level its still just a 'more structured language'. The bad stuff happens when you try to prevent users from being able to express invalid code by requiring that they only express code as a self-validating data structure. You need to be able to put square pegs in round holes when programming. If for no other reason than to demonstrate why you otherwise shouldn't.
If it had array programming where the old arrays were dropped after writes, I would use it
pretty cool
How often do names change? Really?
A straw man leading away from the many upsides of Merkel trees for “versioning”.
In actuality, I would even include the name of definitions into that Merkel tree itself.