I agree, the vertical-bar symbol-syntax is a very smelly onion. And as far as I could tell the last time I checked, the only place arc really depends on it is in webserver code, to define a handler for the application root uri ( "/" ) - this path is defined by the empty symbol, which is represented as ||
You could get away with totally ignoring vertical-bar symbol-syntax, and add this somewhere before you define your webapp:
(assign || (sym ""))
Hindsight is great. I wish I had thought of this before writing all those vertical-bar tests :)
"Do people consider srv.arc to be 'part of arc' for compatibility purposes?"
I would consider any Arc program to be "part of Arc" for compatibility purposes because any Arc program is supposed to use only the core and things defined in arc.arc (along with its own libraries, of course).
Obviously all bets are off if it uses the $ macro to drop into Racket, but then you're not really writing Arc code anymore, so I'm ignoring that case.
"(assign || (sym ""))"
That's a pretty good way of solving that one particular use-case, but it won't help in general because there might be some Arc program that actually does use the |...| notation more extensively.
However, I suspect such programs are quite rare, and with good reason. So I don't think it's a major deal if an Arc implementation chooses to not implement the |...| notation. However, such incompatibilities should be documented, so people using your implementation know what works and what doesn't (and why it doesn't work).
And in that case, it would probably also be a good idea to have your implementation throw an error if a program ever uses the |...| notation. That makes it easy to figure out which programs are broken, and makes it easy to pinpoint where the problem is so it can be easily fixed, rather than failing silently or at a later step in execution.
"I don't understand. If I write a hello world arc program, are you saying that's part of arc?"
For the sake of compatibility with Arc, yes. An implementation claiming compatibility with Arc should be capable of running programs designed for Arc 3.1.
"what do you consider to be the core of arc?"
We were discussing what constitutes "part of Arc" for compatibility purposes. "srv.arc" is not a part of the core of Arc, nor are any libraries written in Arc. But it is still "part of Arc" for the sake of compatibility, in the sense that a compatible implementation of Arc must be capable of running them, warts and all.
Unfortunately, that means that certain programs may end up relying on things inherited from Racket, such as the |...| syntax. I think it must be decided on a case-by-case basis whether compatibility with Arc 3.1 is worth it or not. In the particular case of |...| I don't think it's worth it. But I'm not the one implementing Arcueid: that's up to dido to decide.
Hmm, core-evaluation-test.arc seems to hang Anarki, as well as Arc 3.1.
Well, I've tried to run the tests that do work fine with 3.1 and Anarki under Arcueid and find a lot of issues. For starters, I had no idea that Arc treats symbols with |'s specially. Looks like more accidental behavior inherited from MzScheme/Racket. Scheme48 says that '|abc| contains illegal characters. Guile creates a symbol |abc|. I don't see anything in R^6RS that mandates any of this behavior. Heh, looks like I've got a lot of work to do!
Apparently the bars in symbols is a sort of convention when it comes to case sensitivity of symbols in Scheme. It seems that Arc, in its current implementation anyway, is unintentionally inheriting a lot of onions from MzScheme...
I first encountered this idea in a theorem-proving class - that if you have a hard time proving a theorem, often a really good approach is to find a more general 'lemma' to try to prove. Stronger theorems are often easier because they let you assume more in the induction step. Recursion has a similar 1-1 correspondence with inductive proofs.
Two problems with (o arg default) - you need to remember not to use 'o at the start of a destructuring list (and not get confused when you see (def handle-request ((i o ip)) ...) ), and as akkartik says it's paren-inefficient, a single keyword to delimit required/optional args would mean fewer tokens.
The first problem is easy to fix though - use a symbol that's less likely to be an arg name to identify optional args. How about '= ?
(def myfun (a b (= c (something)) ...)
it has the advantage of similarity with ruby:
def myfun a, b, c=something
disadvantage: looks silly when you don't supply a default value:
What is the issue with  in rainbow? ... Do you have an example?
I think you'll find it's a very simple example. :-p
rainbow.parser.ParseException: Encountered "]" at line 1, column 2.
As for 'macex1, there's definitely one inconsistency, which is that it doesn't work unless car.expr works. I also briefly worried about ssyntax, but I'm not sure it's a problem; official Arc lets (assign a.b nil) cause (bound 'a.b) to be true, but I don't blame Rainbow for not emulating that.