Mocl – a Common Lisp implementation for mobile platforms | Hacker News Reader

Mocl – a Common Lisp implementation for mobile platforms

wukix.com

54 pointspavelludiq61 comments

At 200 bucks a license there isn't going to be more than a couple people who are ready to answer stack overflow questions about this, there isn't going to be a good plugin to most IDEs, the tooling support is going to be poor in general and there is going to be libraries for approximately nothing (yes I know, it takes a lot less time to write them in Lisp, but it still means you need to dive into the Oauth spec to work with Twitter, whereas I can just download a library for Java that works).

I say this as a huge fan of Lisp and there was a time where it made sense to buy a commercial compiler - that time has passed, because when you buy a compiler and a new language you are also buying into an ecosystem, and the more people that are in that ecosystem the better for you.

I can see maybe using Lisp to do some special parts of the program, but even then there is Chicken Scheme which is free and already have a bunch of eggs (extensions), oh and it compiles to C too.

"there is going to be libraries for approximately nothing"

You do realize this runs Common Lisp, right? There are plenty of Common Lisp libraries.

Maybe. CL has a bunch of different implementations that are slightly incompatible and some of the libraries (or their dependencies) have to be chased down on obscure FTP sites that haven't been updated in a long time.

Don't get me wrong, as a language is it the best that I know of. As an environment, it is not.

the eclim talk mentions that many CL libraries work with it so maybe https://github.com/fons/cl-twitter or https://github.com/skypher/cl-oauth will work?

Anyway, lispworks seems to be doing fine selling their lisp, so maybe it isn't hopeless for wukix :)

Competitive acquisitions often result from this kind of development.

It's funny to me that people will happily spend $40 for a t-shirt, $100+ for a pair of shoes, $300 for an ounce of weed -- but ask them for $200 for a professional development tool and they won't budge.

I for one, think the price tag is reasonable.

I'm totally willing to pay for a great professional development tool. However, I think it's way too risky to use a closed source niche product in this day and age. I would have preferred something like the Ravenbrook licensing model: a viral open source license that would necessitate a commercial license to create derivative products or develop commercial applications.

We'll sell you the source code. That's not enough?

No, because it would still have extremely limited network effects.

I don't blame you or what you are trying to do, it is the nature of the beast. Introducing a new programming environment is hard enough when it is free, it is nearly impossible when it isn't.

The thing is that it’s easier to predict the personal value of a real-world item. You don’t know in advance whether you’re going to get $200 of value out of a particular development product, because there’s no reasonable standard of comparison.

You would pay $40 for a T-shirt if it were some sort of collector’s item you could resell. You would pay $100 for running shoes because that’s less than $15000 for knee surgery, and you like having comfy feet. You would pay $150–300 for an ounce of weed because you know what weed is like, and you like it, and the price is driven up by legal issues. But what justification can you come up with for spending $200 on what is essentially a productivity tool, if it won’t necessarily make you more productive?

The same justification I use for buying books on programming: I am trying to make myself a better programmer and I'm willing to take a gamble on myself.

But book stores, unlike the company that produces mocl, won't give my money back if I find that the book didn't help me within the first 30 days.

So in this case, it's even a relatively safe gamble (assuming of course, that the company does in fact, honor the 30-day money-back thing).

Well it depends on the book store really. For example Barnes and Noble lets you return books up to 2 weeks after purchase. Amazon allows you to return kindle ebooks, but I'm not sure of their policy on physical books. And of course you could always go to your local library and see if the book is available there.

For me, it isn't the price tag that is a problem. It is the lack of information about something that has a price tag. I would personally be interested in a product like this, and I have paid $200+ for development software before, but there isn't enough info here to let me know if it is worth it.

Fully agree, I still remember the days I had to pay for every single piece of software I was using.

Nowadays there seems to exist a trend developers should work for free.

Sounds like a cool project. But really, no documentation, no sample code, not even a whitepaper? Just a Buy Now button?

It probably isn't, but it sounds like a scam. Better put out some actual code to show that this is for real.

Agreed. I'd be really interested in a good common lisp for mobile development but I need more to go on than a page with a buy button.

It has been released 30 minutes ago, so web page is a work in progress.

Hey, mocl creator here. I apologize for the minimal website. We are building it out as quickly as we can. Bit hectic with a fixed release date we were trying to meet (and did meet).

It is real, though. I gave a talk about it at ECLM 2013 http://weitz.de/eclm2013/.

Then why even release it? You're releasing a new "coding method" for mobile platforms with no documentation, no sample code, no example applications, no benchmark backups.

Nobody is going to hand over $200 without at least some of those items listed above. So it just seems a bit strange to even want to release it.

I'm just saying you should slow down a little bit, take a look at http://www.rubymotion.com/ for example. It's the same sort of thing, Ruby for iOS. They provide both examples, publicly available documentation, they even have a way to try it out yourself and see what it looks like.

For what it's worth I just put up some code so you can see better how it works.

So, you write the UI code in Java for Android and Objective-C for iOS, then there is some build process for both platforms?

This seems a little awkward. A sample project setup to build for both platforms would be useful to see.

For some types of applications, the native UI boilerplate would probably be minimal, but for apps with complex UIs, this might add a lot of dev overhead.

Don't get me wrong though, this looks cool!!

congrats on reaching your deadline, but now you've shot yourself in the foot in another way by tarnishing the..."brand", for want of a better term. you can pat yourself on the back for hitting your target date without having to prematurely publicly release the whole thing.

A month ago mocl was considered vaporware by many. Not releasing it was damaging their "brand".

This is exactly what I am looking for and I would probably buy it if I had some reassurance that it would actually work. Where is the docs? The sample code? Proof of concept apps that can be downloaded from app stores? Fine print for what is allowed in the different licenses?

Documentation, and a little bit of sample code is included in the download. The license agreement is viewable at http://wukix.com/mocl-license-agreement.

With fine print I meant answer to questions like "what do I get in a commercial license that I do not get in a personal license?" as it is written it could mean that with a personal license you are not allowed to earn money from apps, or it could mean that it would only cover a single developer (then how many developers do I get for the commercial license?)

Also by actually reading that license it seems to contradict what you promise elsewhere, it says: "All fees shall be non-refundable".

Personal just means you are buying it with your own funds and not being reimbursed by a company. You are free to make paid apps with it. I'll work on language to clarify.

Thanks for pointing out the contradiction there. We may be going above and beyond what is required of us by the license, in the case of refunds. But we'll still do so.

Can you guys provide something to look at before I pay? A free but limited version like Lispworks will be ideal and may increase your numbers.

Seems to me like LISP might defend itself as a mobile language through relative ease of GPU integration, given the parallelization possible through OpenCL/CUDA, the difficulties of using them with traditional languages, the mobile platforms' challenge of electrical power consumption, and the limits of CPU-based processing.

A little bit expensive and with no trial (no, I don't want to buy something to return it later).

It's actually cheap for a production ready(if it is) commercial cl implementation, but I agree, a trial version is essential.

The same applies to a lot of software. If you just speed up a few hours then the ROI is very high.

BUT from the commercial perspective you must follow some "commercial software standards". You can even lose customers because the commercial proposal is far from the standard and "against" the customer.

This comment was downvoted without a reply

However, note that most commercial lisp implementations, such as LispWorks, Allegro CL each have a free (gratis, not libre), albeit limited, personal/noncommercial edition as well as evaluation and trial licenses for their other versions.

mocl creator here. Trial version will be coming in the weeks ahead. Didn't have time to put it together yet. Sorry about that. We do accept returns though.

I am not sure that a trial version is really necessary if you provide several solid complete examples (source, build setup for both platforms) and pre-built apps for the demos that we could try on our Android and iOS devices.

Nice, and congrats to the wukix Team for their effort. However, will Apple accept those apps in it's store? There was this guy who wrote a game in Gambit Scheme (iirc) and was rejected back then. This seems to be the same process here (generate C code from Lisp and glue it toether with GUI code), just with CL.

Apple no longer cares what language you use, as long as new code isn't downloaded to distributed apps after installation. So it's okay to use CL as long as there is no facility for interpretation of CL downloaded from the net.

I think Apple will accept anything compiled with mocl. A similar project is RubyMotion (for Ruby obviously) and it seems to have no problem with Apple.

I'm very happy about mocl's release. Great work, and congratulations!

Clojure on Android: https://www.youtube.com/watch?v=1NptqU3bqZE

You do realize that Clojure is not Common Lisp ?

That being said, the linked video is from 2011 (published in 2013, but from 2011).

Clojure isn't Common Lisp, but it's A Lisp, and a pretty decent one...

Clojure is incompatible with any other Lisp...

What do you mean exactly?

I mean, Common Lisp is incompatible to Scheme for example.

Common Lisp was designed to be backwards compatible with Maclisp. Maclisp was developed from McCarthy's LISP.

Clozure OTOH is compatible with nothing.

It's compatible with all devices which support a JVM, which these days outnumber the ones supporting Maclisp.

Yeah, it is compatible with the JVM. But that does not make it a Lisp. Not a bit.

How isn't Clojure a Lisp?

What is a Lisp according to you?

I'm not asking this in an agressive way, btw, I'm just really curious.

    Clojure 1.5.1
    user=> (car '(1 2 3))
    CompilerException java.lang.RuntimeException: Unable to resolve symbol: car in this context, compiling:(NO_SOURCE_PATH:1:1)

There are several things wrong with above.

I would expect:

    ECL (Embeddable Common-Lisp) 12.5.1 (git:6f2fd5413066103f46a7d9f70148c2f0541698f7)
    Copyright (C) 1984 Taiichi Yuasa and Masami Hagiya
    Copyright (C) 1993 Giuseppe Attardi
    Copyright (C) 2000 Juan J. Garcia-Ripoll
    ECL is free software, and you are welcome to redistribute it
    under certain conditions; see file 'Copyright' for details.
    Type :h for Help.  
    Top level in: #<process TOP-LEVEL>.
    > (car '(1 2 3))      

    1

So simply that a language does not support car/cdr but uses first/rest instead disqualifies itself as a Lisp?

So? Is Scheme compatible with Common Lisp?

Clojure can run on any platform the JVM can (so basically anywhere except iOS). Clojurescript can run anywhere Javascript does (so everywhere). It's a practical, immediately useful language, that's seen significant uptake and activity.

I like Common Lisp, but it's much easier to build useful apps in Clojure.

> I like Common Lisp, but it's much easier to build useful apps in Clojure.

Unlikely.

From language implementer perspective your ECLM slides are lame.

Is this ECL re-packaged you charge > $ 200 for without even giving a trial whatever? :-)

could you only put a gist with the code of a hello world application ?

Like AppDelegate and a label with hello world in it ?

Looks very interesting!

Some blog articles showing development of sample applications would also be nice :-)

So is this based on an existing CL implementation, or a totally greenfield project?

It compiles to C; there is no JIT/bytecode/interpreter. It may be created from scratch.

http://wukix.com/dist/mocl_eclm2013.pdf

Is this same as Gambit-c or Chicken scheme, but for CL?

If what you ask is if mocl compiles to C first (like Chicken and Gambit), than yes. It is the same.

interesting

I'm going to assume a few things here...

- No one has ever used this. - It is filled with bugs and has no full time developers supporting it. - It is extremely limited. - The benchmarks are completely false.

I'd buy something like this if it didn't look like a good rich scheme.

Reply on news.ycombinator.com