Hacker News Reader

Comment by geuis | Hacker News Reader

I agree. I haven't been much of a fan of Cappucino since seeing a review at a Google JS meetup several months ago. I have just never been sure what the point of a language written on top of javascript is. The whole point of really knowing not only the core language but also the inconsistencies between IE and the rest of the browsers is that I want to build web apps that perform fast and robustly. Abstracting that away into yet another layer that runs moderately slowly on the 70% of common browsers just means that it runs that much more slowly. To be fair, I give all the credit in the world to the 280 North guys, but charging $15 for access to a beta program for an environment and language that really has no broad adoption is just not worth it.

View on HN

"I agree. I haven't been much of a fan of the NeXTstep class library since seeing a review in [a prominent publication of the period] several months ago. I have just never been sure what the point of a language written on top of C is...

"To be fair, I give all the credit in the world to the NeXT guys, but charging $3500 for a microcomputer without a vast application library and that really has no broad adoption is just not worth it."

That said, here's how I see things.

(1) 280 North took a broken environment and tried to apply a not-entirely-broken substrate atop it. Is it going to be slow? Sure, but consider also that Javascript is inherently slow to begin with.

(Slow begets slower, blah blah, I don't want to hear it. I don't care.)

(2) It's a start, and when it comes to 280 North's business practices, I'd be willing to bet that there might very well be a standalone compiler for Objective-J at some point if enough people ask for it.

... if I don't beat them to it first. It's getting close to the point where I'm considering just hacking something together in OCaml to scratch my own itch in this vein.

(3) The 280 North guys are smart guys. They listen. Complain. Send them mail.

On point number 2, the compiler ships with the open source project. It's written in Javascript. It runs in the browser, or on the command line.

We are planning on getting a new compiler in 2010, and it will still do all those things. It will also be more robust, introduce some sorely needed new features, and with any luck be faster to boot.

With respect to point one, I'm not sure anyone commenting in this thread really understands the complexity of optimizing application performance in the browser. On an absolute level, message passing is slower than function calling. But taken in context, that's a small percentage of what happens in an application.

Cappuccino, because its an application framework, can make certain assumptions to dramatically improve performance. Assumptions that can not be made here and there by individual app developers. For example, coalescing DOM updates and drawing events on the run loop. Combining timers together to minimize overhead. Delegating events from the document level to prevent unnecessary build up of event handlers. On the tools front, we can strip some dead code from projects, automate spriting of all app resources, target different code paths to different browsers automatically, and a slew of other things. Performance is complicated.

Remember that because Objective-J is a superset of Javascript you can still write native JS code in an app. This is the beauty of Cappuccino: it provides the benefits of a framework (consistency, robustness etc) with the flexibility to pretty much still do whatever you want.

The 280 North guys are the first to admit that CP isn't for every web app, but it serves a specific need exceptionally well. And, personally, I think its' application and adoption will only broaden.

I can't wait to part with my $20! They have to monetise somehow, and if this helps the next year of development - hell, make it $40!

Yup, I couldn't agree more. In fact, I've even got an internal webapp for my company that I have planned to build using Atlas, so as of tomorrow morning, I'll be signing up for the beta program. $20 really is peanuts for a professional development tool.

yep $20 is nothing for a professional development tool. even 200$ is totally acceptable. But they only provide access to a beta version of a unproven tool... not my definition of professional.

i dunno what the tool worth and i dont plan to pay 20$ to find out if i need this tool.

i dont blame the product, just the moneytisation is broken. Ok to sell a tool, but first convinces potential buyers

That's fine. We're not looking to convince anyone. We are recognizing that we have something useful now, and that there are people who are willing to pay for it now. We are providing it to those people, and making it known that we believe the application is still not ready to be called 1.0.

We happily admit that a paid beta program is a bit of an experiment, but it's hardly unprecedented. See Mac OS X Public Beta for an example. We do not want to set up the false expectation that Atlas is free. It isn't.

Will there be a substantial discount on the final version for beta testers? If so, sign me up!

another marketing glitch: dont launch a professionnal tool on sunday. do it during a work day, say around 10am pst to maximize the effect.

I don't really think it matters when it launches.

That said, if you wanted to hit the ground running on Monday morning, launching on Sunday allows for the blogosphere to start talking about Atlas on Monday morning.

HN needs to get private messages.

I can not continue this thread without appearing as a 280n hater and i didnt find a way to reach boucher directly.

ross at 280north dot com

I think the potential buyers are already convinced, it's the people who have already started developing with Cappucino.

Having a paid beta, means that people who haven't used Cappucino probably won't try it and it's a good thing because they would have been harder to support...

It's $20 actually. But you're only paying for a development tool. The language, the framework, the runtime -- everything else is open source.

The extra abstraction layer Cappuccino provides seems worth it to me. Much as I like jQuery et al., an MVC-style framework for writing desktop-like web apps, along with an Interface Builder clone, blows manipulating the DOM out of the water. IMO, of course.

Reply on news.ycombinator.com