Apologies for rambling on a bit - but I also have some questions about VPRI. As far as I can gather, it was never the intention to publish the entire system (The whole stack needed to get "Frank" running)? If so, I'd like to know why not? Where you afraid that the prototypes would be taken "too seriously" and draw focus away from the ideas you wanted to explore?
The VPRI reports, and before that some of the papers on Croquet (especially the idea of "teatime" which might be described as event-driven, log-based, relative time with eventual data/world-consistency) are fascinating, and I'm grateful for them being published. Also the Ometa-stuff[o] is fascinating (if anything, I think it's gotten too little mind-share).
It seems to me, that we've evolved a bit, in the sense that some things that used to be considered programming (display a text string on screen), no longer is (type it into notepad.exe) -- it's considered "using a computer". At the same time some things that were considered somewhat esoteric is becoming mainstream: perhaps most importantly the growing (resurging?) trend that programming really is meta-programming and language creation.
ReactJS is a mainstream programming model, that fuses html, css, javascript and a at least one templating language - and in a similar vein we see a great adoption in "transpiled" languages, such as coffee script, typescript, clojurescript and more. HN runs on top of Ark, which is a lisp that's been bent hard in the direction of http/html. I see this as a bit of an evolution from when the most common DSLs people were writing for themselves were ORMs - mapping some host language to SQL.
In your time with VPRI - did you find other new patterns or principles for meta-programming and (micro) language design that you think could/should be put to use right now?
Other than the web-developers tendency to reinvent m4 at every turn, in order to program html, css and js at a "higher" level, and the before-mentioned ORM-trends -- the only somewhat mainstream system I am aware of that has a good toolkit for building "real" DSLs, is Racket Scheme (Which shows if one contrasts something like Sphinx, which is a fine system, with Racket's scribble[2]).
Do you think we'll continue to see a rise of meta-programming and language design as more and more tools become available, and it becomes more and more natural to do "real" parsing rather than ad-hoc munging of plain text?
The VPRI reports, and before that some of the papers on Croquet (especially the idea of "teatime" which might be described as event-driven, log-based, relative time with eventual data/world-consistency) are fascinating, and I'm grateful for them being published. Also the Ometa-stuff[o] is fascinating (if anything, I think it's gotten too little mind-share).
It seems to me, that we've evolved a bit, in the sense that some things that used to be considered programming (display a text string on screen), no longer is (type it into notepad.exe) -- it's considered "using a computer". At the same time some things that were considered somewhat esoteric is becoming mainstream: perhaps most importantly the growing (resurging?) trend that programming really is meta-programming and language creation.
ReactJS is a mainstream programming model, that fuses html, css, javascript and a at least one templating language - and in a similar vein we see a great adoption in "transpiled" languages, such as coffee script, typescript, clojurescript and more. HN runs on top of Ark, which is a lisp that's been bent hard in the direction of http/html. I see this as a bit of an evolution from when the most common DSLs people were writing for themselves were ORMs - mapping some host language to SQL.
In your time with VPRI - did you find other new patterns or principles for meta-programming and (micro) language design that you think could/should be put to use right now?
Other than the web-developers tendency to reinvent m4 at every turn, in order to program html, css and js at a "higher" level, and the before-mentioned ORM-trends -- the only somewhat mainstream system I am aware of that has a good toolkit for building "real" DSLs, is Racket Scheme (Which shows if one contrasts something like Sphinx, which is a fine system, with Racket's scribble[2]).
Do you think we'll continue to see a rise of meta-programming and language design as more and more tools become available, and it becomes more and more natural to do "real" parsing rather than ad-hoc munging of plain text?
[o] https://github.com/alexwarth/ometa-js
[s] https://docs.racket-lang.org/scribble/getting-started.html
http://lambda-the-ultimate.org/node/4017