I'm not sure what this means for Arc, but other than pulling the plug on this site, the only thing I can see that would be bad for the community would be if he publicly said that Arc's dead, or something like that.
I agree -- while it would be sad if pg stopped using arc, that itself wouldn't be the end of the world.
The problem I see is that if pg publicly stops using arc, the already-slow trickle of newbies would slow even further.
Honestly, I'm starting to tire of this community being so small. I've been avoiding Clojure -- mostly because I really like Arc, partly because I've had bad experiences with some of the people in Clojure's community -- but Arc isn't getting to be much easier to get things done. And I'm not sure I'm going to be able to put it off much longer.
The problem with Arc as opposed to Clojure is that it is incomplete. It is missing some of the most basic features such as a a module system, a good debugger, and useful error messages. The reason none of us has come up with a solution to any of these deficiencies is that they are hard problems solve. We are all just waiting for pg to try to fix these problems instead of going out and solving them ourselves. If someone can solve these issues, pg or not, Arc should gain traction. I hate to say it, but if no one is going to do anything, Arc might as well already be dead.
I'm certainly not waiting for arc to become perfect. My side project may well end up needing me to reinvent a computing stack from scratch: http://akkartik.name/about. But I fully expect it to take 20 years, so don't hold your breath :)
Are you fully sure there is no other way to do what you want to do without it taking 20 years?
That's a question worth pondering, not just because big things tend to fail or good design is simple, but also because life has a way of changing what people choose to do over time.
Yeah, I'm actually constantly looking for shortcuts, ways to not need to reinvent the stack, ways to build things that will help existing projects, or allow people to use existing languages and tools.
When I said 20 years I wasn't saying I have a plan that's going to take 20 years. I meant that I'm building small projects that -- when I step back -- seem to be headed in that general direction over time. Projects like wart[1], tracing tests[2] and layers[3]. I'm building these little projects out in the open, so hopefully I won't fall victim to the "hacker comes out of basement after 20 years with a compiler and OS that nobody cares about" failure mode. No, my goal is to fail in some more creative way :)
I'm less worried about my goals changing. On the one hand, if they do they do, one can't be anything but fatalistic about it. On the other hand the goals I outlined above seem to have hooked into my motivation at a deep level. I may well end up switching solutions I explore, but it seems like the goal is unlikely to change.
I haven't really started building anything for the past couple of months, but I've been thinking lately about three potential little projects: a) a simple multitasking OS without virtual memory or processes, maybe even without preemptive scheduling, b) a better language for teaching programming by separating scope from the notion of functions, and c) an assembly language with structured programming and lisp macros somehow. Something 'easy' to build without needing an entire C compiler. Rephrased: "I'm willing to do my own register allocation. What can you do for me, lisp?" :)
Mostly I just get up every morning and work on whatever I want. And these are the things I've been ending up working on.
I've had bad experiences with some of the people in Clojure's community
When arc advances enough to appease you, it will also be appealing to others; then when a larger community develops around arc you will start to have bad experiences. That's the nature of mass appeal.
Even pg noticed the quality of HN's content and community degraded with popularity, which is why he's adding the pending change discussed (amongst others).
And let's all be honest, if arc solved half the problems Clojure handles, I doubt you would notice much of a difference between the languages. After all they are already very similar.
If you were to draw a line with liberal being at one end and conservative on the other, then plot various languages on that line you'll find arc and clojure would be close together as opposed to far apart.
Even Yegge's post would seem[1] to confirm this with his own data:
1. I assume arc to be like lisp. Many of the things that need a new interpreter atop scheme (lisp-1, quasiquote, unhygienic macros) are borrowings from lisp. The only major scheme-ism arc retains is call/cc. Am I missing anything?
2. I actually think arc is more batshit liberal than traditional lisp. Perhaps I'm reading too much into the absence of a module system, but I consider it a deliberate omission.
3. You're assuming that there's only one way to build out the missing infrastructure, and that is to be more like clojure. I'm skeptical of that.
Yeah, I can see how one could draw a little larger of a distinction between Arc and Clojure using the conservative/liberal categories, but even so I'm not convinced it even stands up as a good overall measure for similarity/differences (even if it's a good one). So setting that one criteria aside for a moment, I look at Arc and Clojure and here's what I also see:
- They are both Lisps with dynamic type handling, prefix notation, strings as sequences, heavy use of hash-maps, a large number of functions and macros with similar names doing the same things.
- The syntax's are so similar you could pretty much swap out color scheme templates (in fact I initially used the arc textmate bundle for Clojure).
- They both depend upon pre-existing language for compilation. Clojure drops into java, Arc drops into scheme.
- The online comparisons between them are more frequent than arc and any other language.
I could go on, but I think you get the point.
And while I know you could create a list to show differences I'm simply saying that overall they are very similar which was my initial claim.
As for "having to be more like Clojure" to advance, well no, but I'm pretty sure that it's more likely than any other unstated options. Here's what I do see to support this:
Ok, I'm convinced :) By pg's comments in particular.
(There's a lot of red herrings in your comment, though. That they look the most similar syntactically isn't too surprising, and it's irrelevant that erlang can't compete, or that they build on an existing language, etc. You're measuring position to compare trajectories.)
In any case, it feels quite delusional to compare arc to clojure at the moment :)
Let me ask you this: if you have clojure why do you care about arc? Genuinely curious.
> “There's a lot of red herrings in your comment, though. That they look the most similar syntactically isn't too surprising, and it's irrelevant that erlang can't compete, or that they build on an existing language…”
I realize many of my arguments have holes in them when viewed independently, but when put together those “soft arguments” contribute to painting an overall picture that supports my claim better than had I not given them. Albeit the "they build on an existing language" is pretty weak overall.
> “You're measuring position to compare trajectories”
Position is not something you can measure it's something you can map/use. That aside, I don’t believe what it is I think you intended to say is true. I used position to substantiate the claim of similarity, I used trends to substantiate the trajectory (i.e. the last half of my reply, what people are saying and what they are doing are trends).
> "if you have Clojure why do you care about arc?"
I like both languages. I don’t use Arc because its unfinished and unusable for the things I need to do. I often get the sense that when I promote Clojure on the Arc forum that people get defensive and wonder why I'm even around, but remember I am only saying what pg himself is:
> Would you recommend Arc to modern startups in general?
PG: No, I don't think so, not in its current state.
> Why not?
PG: It's still missing some things that most people take for granted.
> Are there any Lisps you would recommend.. ?
PG: Clojure is probably the best bet...
I like the arc language and I like the community. I view arc as unfinished and look forward to giving it a shot when it is, but until then I am going to continue to be realistic about the current state and promote Clojure as an option.
"I used trends to substantiate the trajectory, i.e. the last half of my reply, what people are saying and what they are doing are trends."
If you think arc is whatever pg says it is, then what we are saying and doing here is irrelevant :)
If you think arc is whatever we're saying and doing here, the things we take from clojure seem a tiny minority compared to the sum total of conversations. I'd say arc starts out kinda like clojure, but is looking to steal ideas from all sorts of languages including erlang.
---
I tried to phrase my question very carefully to avoid seeming defensive about something I don't care about. So let me just come out and say it: I have no problem with you talking about clojure all you want here. If you did, maybe I'd get to talk to you more! :) Arc is indeed absolutely unfinished, no disagreement there either.
Maybe what I'm actually defensive about is the prospect of any two languages becoming more and more similar. That just seems bad, nothing good can come of it. I'd rather see arc and clojure evolve in different directions and give me more ideas and more data about how good those ideas are.
Copying ideas and creating hybrids is totally fine, that's what we are good at. But then the hybrid starts at the intersection of its influences and sets off on a whole different trajectory.
So let me rephrase my question: is there something you miss from arc in clojure that has you wishing for a superset?
(The answer may take time to emerge from the subconscious. At least it has for me in similar questions.)
My comments are a targeted response to statements made in a specific comment (http://arclanguage.org/item?id=18421). I think you're treating the conversation in this thread as you treat this forum; as some place to mingle languages and abstract all ideas. That's fine if you want to do that, but I'll suggest you'll be more effective if you follow the thread and consider context when responding.
And I'm suggesting more care be taken in this regard, because it leads to you being offended (i.e. "If you think arc is whatever pg says it is, then what we are saying and doing here is irrelevant...") on statements that should be considered only in relation to the thread. It also leads to unintentional attacks such as "In any case, it feels quite delusional to compare arc to clojure ...". I don't think you realize you suggested I'm being delusional by adding that comment... Arc and Clojure are both modern lisps that have eliminated the overuse of brackets, so really is it delusional to suggest these two languages are more similar than not? Am I creating some injustice by telling someone, that's already leaning towards Clojure, that these languages are very similar anyway?
I'm going to end this thread here as I think it has already gone off the rails.
Sorry I'm causing offense. It was indeed utterly unintentional, and fwiw, I actually was never in the slightest conscious of feeling offended by anything in this thread. I'm entirely the offender and not an offended party.
I didn't mean to trigger associations like "injustice". When I said "delusional to compare arc to clojure" I was trying to be self-deprecating. Arc is just a toy, clojure is real. You're absolutely right in your defense of clojure.
Not to worry, I've known you long enough now to know you're not mean spirited or intending offence. Hopefully I didn't get too grumpy, but I needed to put and end the thread because I could see we had different ideas on what the thread was even about.
I can relate to the frustration. It's hard to predict when the next great stack will come along. I try not to make irreversible choices so I can jump to it as long as possible.
No danger of it happening here. Arc forum is running some ancient version from what I can tell. They probably learned to stop messing with it after the voting ring detector went haywire. I wonder how many of us remember that :)
