To clarify:

- unevaluated expressions are called blocks, and with them the line between code and data blurs, since you can treat block as either one at any time (homoiconity);

- there're no scopes, only namespaces (contexts);

- yes, Parse is one of the Red's crown jewels, many people (and author of the article himself) treat it as some kind of "plain English regex", while in fact it's not limited only to string parsing, but can be applied to any other series-like value and code (i.e. blocks) itself, which allows for some cool metaprogramming tricks;

- R/S (without any optimization layers whatsoever) is 2-3 times slower than C.

Thanks for the clarifications.

> it's not limited only to string parsing, but can be applied to any other series-like value and code (i.e. blocks) itself

Sounds somewhat similar to OMeta (http://www.tinlizzie.org/ometa/) which can pattern match streams of objects.

With contexts it sounds like almost every word can be late bound? Also it seems blocks and contexts are decoupled which means you could evaluate multiple blocks from different parts of your program in the same context, or give each a unique context, etc?

Every word (the thing what author calls a "variable") has a binding. In short, it's a pointer to the namespace where value associated with that word resides.

Blocks can nest any other value, including words. At any point in runtime you can change this pointers (bound word to some value) for every word in a block or for separate words.

You can pattern match streams of values, and those values can be any lexical form that Red supports.

I'm not sure how easy it is to do DSLs. You don't see this popup in the Red or Rebol docs for a reason. It is homoiconic like Lisp, so it should be feasible. I'd love to hear some of the core devs like doc respond.

How hard it is all depends on the DSL. I just found another article on HN and whipped one up as an example. See: https://news.ycombinator.com/item?id=16010971

It's not black magic, by any means, but it's also not yet point and click. We'll build more tools, higher level pieces, and more in the future.

It might get hairy if DSL is non-trivial, IIRC Doc has plans to implement meta-DSL wrapper (a la Rascal) sometime after 1.0 launch to ease the process.

I hope Red puts up a couple of really simple examples on how to build DSLs. It's one of the great things Red/Rebol has to offer, but it isn't obvious how to do it.

I guess you can start with [1] and ask people in community, there may be a couple of simple examples to show.

[1]: https://github.com/red/red/wiki/Red---Rebol-Dialects:--Selec...

I've tried asking the community before and the response wasn't good. With that being said, the links you sent were pretty nice, so thank you!

You're welcome! I actually rather interested in DSL creation myself and want to dabble my feet in this topic sometime. And I also feel that creation of dialects is rather obscure art for old-timers, and that there's a need for learning material.

With that said, I would appreciate if you'd listed what would you like to see in "How to bake DSL with Parse" guide of some sort, that would be really helpful when the time will come for writing tutorials and documentation.

Sorry, I guess what I'm trying to say is I'm aware what a DSL is (like View dialect), but no clue how to start creating one other than it has something to do with blocks and Red's homoiconicity. There needs to be a specific spot on the documentation page that gives a good overview with simple examples and finally a more complicated one. In fact, I hope they do that with all the planned additions of erlang like concurrency...etc. You know what I mean?