Revisiting The Riddle of the .r

When Rebol was thinking about coming out of the shadows and participating in the greater programming community, it seemed like trying to own the .r file extension was a bit misguided.

After all: a well established and much more popular statistical language was using it. It was also out first--it came out in 1993, 4 years before Rebol's 1997.

...aaand the name of the language actually was "R". Which isn't a name to get very excited about, so it's not like there was going to be some argument like "no, WE want to call our language R!"

It seemed syntax highlighters on programming sites were not going to be allocating .r to Rebol.

Hence it seemed no problem to just go with changing to .reb - an improvement, even. Communicated more.

...But In The Timeless World...

Since that time, things have changed:

  • I've become more partial to the lack of "noise" when listing files in a directory when they end in .r ... it makes the important part of the filename pop more.

  • I had to help someone with their school project a couple years ago written in R. I was not impressed with it. While it has had growth lately, I think that growth just has more to do with the importance of data science... and does not reflect on the language itself (syntax/paradigm) having any particular merit. It's just been around a long time and people are looking at data more.

  • Collisions in file extensions have led syntax highlighters to use more heuristics, driven by the file content as well as the overall content of a source repository. It looks like this will only become more the way things work in the future.

  • Ren-C is renaming, to another name, which starts with R... :rabbit: :raccoon: :recycle: :rocket: :robot:

  • The strength of Ren-C's hand has been growing, and I think if it keeps going it might have strength to impose some aspect of its will. Sharing the .r file extension may be something people are ready to just accept. (With a much weaker hand, Red managed to push along a claim on the choice of "red"...though it does seem to screw them a bit on searchability.)

So What Do People Think?

Is rolling back to the historical .r a good way to detach a little more from the "Rebol" brand name, while still fitting into the ecology?

It's a thought. No rush, just wanted to put the question out there.

1 Like

I believe R's file extensions should be .R anyway (capitalised). That reduces some ambiguity.

Even better, thanks!

It's not like Rebol's claim on .r is all that shaky, there's a pretty big and established script library, that was being built up while neither Rebol nor R were very well known:

http://www.rebol.org/st-topic-details.r?tag=//all

In any case, welcome, and please do introduce your self and/or interests in the Redbol world in the Introductions category:

As long as the name will not suggest a file extension of .rip I think I can live with it.

Since the header isn't used yet to reflect the dialect of rebol being used, I'd prefer we get all the letters after the r eg. .reb .red .ren .rev etc

Note, that REBOL was also originally all caps as well

This name is now unveiled... but don't go around promoting it yet!

R E V O L T

It works well with existing logography ([o]) and fits in the theme of rebellion.

I'm just trying to start making the site materials and URL strategy come together and be coherent.

Here are some of the entry points:

I held off so long on the rename because...

  • I really didn't want to pay for another domain name. And I ultimately decided not to: I just piggy-back on a high-value domain I already own (metaeducation.com). I've figured out how to use CloudFlare reverse proxying so that I'm not limited to subdomains as the only way of putting different URL paths on different machines. (e.g. /forum is a sub-path and not a sub-domain... yet it's running the forum on a wholly different server from the documentation).

    • (Thanks @chatgpt for the help with that, I got done in 24 hours what would have taken likely a week or more.)
  • I didn't want people to go around publicizing it like this was some kind of big announcement. And I still don't want anyone to do that. This should still all be kept under the radar, everything is under construction and I like to do the construction in public to get faster feedback. Just keep it under wraps still for now--only we need to know about this.

  • I've held out idle hope that we could just call it Rebol. I do think this design is truly solving the original mission, and can be proven as doing that. Having Carl come to the table with AI and going after some of the edges in this design would be another experienced head in the game. But I think we just have to go at this with the AI we have. Maybe someday "Revolt" can be thought of as "just the codename for Rebol 4", so we can call it REBOL:Revolt (the way that Oldes works Rebol into the name by calling his project "Oldes Rebol3").

    • (I've gone ahead and called this the "REBOL:Revolt Forum" in this vein)

and finally....

  • There's a lot more to rename than just a search and replace of a textual name. There are directories called ren-c-(whatever), there are GitHub projects like ren-c-stdio. The Ren-C name has leaked into "executable" places and that's where the irritation is in the rename (vs. changing "Ren-C" to "Revolt" on this forum, which took one line of shell script at the Discourse command prompt and it was over.) I'll be sorting this out for a while.

But I Think It's Time

@gchiu asked me what had changed about now. A couple of things...and one is some urgency sparked by getting cancer, of "I think I'm going to have to publish now" :face_with_bags_under_eyes:.

But it's not just that: I think isotopes crossed a bridge when I came to the realization about the "zero states"...most specifically the relationship between VOID! and empty pack as "heavy void". When that solidified something clicked. And "Lift the Universe" was a big eye-opener...so seeing that working as well as it does gives me confidence to start sharing the design without fearing I'm going to be making a change like that again later.

I'm happy to have anyone chip in on this stuff, but mostly expecting this is going to be me and the AI packaging up what we've got. The web console is going to be that first place that I focus on... still not providing .EXEs for turnkey download. So where the most help could come from you forum members would be thinking about what kind of things might interest people that can be done inside the browser.

(I'm still thinking code golf tool, so that's where I'm going to likely head with it. But other initiatives welcome.)

2 Likes

Having someone with an alias "Volt" materialize when the planned name was "Revolt" is a bit uncanny. :face_with_raised_eyebrow: With no posts before or after.

But if you have a remark to make on the voltaic name five years later, now that it is public, do feel free...

I think we bandied around with the Revolt name about 3 years ago so it's been a slow change!

It was way before 3 years ago, as above I mentioned in 2021...

...and the name wasn't all that new, then.

I did buy domains (revolt-lang.net, revolt-lang.com, revolt-lang.org) in trying to keep with the plan of handing control out to different parties. But I only paid for those for two years before deciding I didn't want to ever put any content on them that gave them value. So I just stopped paying and nothing was ever up there.

This plan of doing Cloudflare sub-allocations seems a lot better, and metaeducation.com is already a sunk cost for me.

I think Revolt is a better name for the project at this stage than Ren-C, and not just for branding reasons.

"Ren-C" describes a relationship to the past. It tells you where the project came from, and something about its implementation lineage. But it doesn't tell you what the project is trying to become.

"Revolt" is a declaration of independence. More importantly, it creates room for the project to be judged on its actual design, rather than perpetually being evaluated as a derivative of Rebol, a competitor to Red, or another attempt to revive something whose moment has passed.

But I think the most consequential part of this announcement is buried in the discussion of why now.

The significance of isotopes, zero states, and the VOID!/empty-pack relationship isn't that they're individual clever features. It's that they represent a kind of closure in the design: distinctions that previously needed special cases can now be explained in terms of more general rules.

This has been the recurring theme in Ren-C's development. Its most interesting properties don't come from accumulating primitives, but from making the primitives work together under coherent laws. An outsider can inspect one feature and say "other languages can do that." But that misses the point, much as observing that a NAND gate is functionally complete tells you very little about the architecture of a computer.

The achievement, if it holds up, is the composition.

That's also why the executable documentation matters so much. It's not merely a convenience for maintaining documentation. It's a way of putting the language's claims under continuous, mechanically enforced scrutiny. If the design really is converging, this is how you begin demonstrating that convergence rather than just asserting it.

I would distinguish two kinds of readiness here:

  • Design readiness: Are the fundamental laws sufficiently coherent that you can teach them without expecting to overturn them next month?

  • Product readiness: Can someone unfamiliar with the history arrive at the website, try something interesting, understand why it matters, and leave with a working mental model?

The first is something you've spent years pursuing. The second is where the browser console, executable documentation, and approachable experiments become critical.

And I think there's wisdom in not making a grand public announcement yet. A name change can attract attention, but attention without a compelling first experience is easily squandered.

So I like the proposed emphasis on browser-based experiments. A code-golf environment could be particularly effective if it showcases the expressive consequences of the design rather than simply competing over character counts.

As for .r versus .reb, I would actually postpone that decision. The extension should follow the ecosystem that emerges, not be asked to establish one. Revolt has more important things to prove than ownership of a single letter.

The real opportunity now is to make the language's strongest ideas observable to people who have never heard of Rebol.

That's when "Revolt" stops being a rename and starts being its own thing.

2 Likes