I agree that it's a stretch to imagine Shoes on the browser, but there are several main drawbacks to Java / Flash / Silverlight. These are all designed mainly for sandboxed execution and rich graphical applications, but often times for prototyping, you need tighter desktop integration or the ability to drop into native code, both of which J/F/S aren't designed for.
In contrast, the Shoes / Ruby MRI stack at least gives you is (a) an easy way to extend web or web apps with native code (C or Ruby) via Ruby gems and Ruby/C extensions (b) a simpler DSL for UI authoring (in contrast to the XML-heavy, type-rigid directions of modern J/F/S) (c) an emphasis on Ruby and on tinker/bricolage innovation more generally (more Haml than XAML, more jQuery than Dojo) To be specific, my most recent research involves prototyping with the Mapnik mapping library written in C++, and there are many more rich media libraries that go beyond <canvas> -- think more SIGGRAPH 2009 than SIGGRAPH 1974. For example, how do we support rapid prototyping of multi-touch on the web? Ubiquitous computing? Computational photography? While it's great that Processing.js and the like exist, to be really bold and experimental, we (in the advanced prototyping / research / hacking space) need to dip and dive deeper than modern notsofreaky sandboxes allow, and hope that HacketyHackers won't be confined to such small but Flashy spaces. ~L On Wed, Sep 3, 2008 at 5:41 AM, Bluebie, Jenna <[EMAIL PROTECTED]>wrote: > I think you need to understand that it's a bit of a stretch that shoes will > work as a desktop platform on OS-X, Linux, and Windows well, because > currently, it doesn't it kind of sort of does... We are all hopeful, and > there is plenty of progress each and every day towards that fantastic goal. > There are many great places shoes could visit after that, like, perhaps the > iPhone, and Android devices, and perhaps the web browser too. But until > then, Shoes has plenty to be, plenty to live up to, and frankly, we don't > need a new browser plugin for interactive widgets. We already have java > (think jRuby + processing), flash (javascript is a fantastic language on > it's own, available in a high speed bytecode compiled version in flash), and > indeed also Silverlight, which is able to run ruby applications as a widget > on macs and pc's via .NET, and is gaining support from open source platforms > too. > What the web really needs now, are quality web browsers with enough power > to make fantastic javascript applications. Nobody wants to install one more > plugin. They're a pain in the butt, and as google points out in their own > comic, compromise google's ability to create a secure browser. Google is one > of many pushing the forward now, especially with Chrome's V8 javascript > engine, which I expect is even faster than ruby 1.9 would be, given it > compiles javascript in to raw x86 machine code that runs straight on your > cpu. There is no VM... the javascript becomes executable binary. > > So now we have compiled code execution speeds, brilliant css support, > canvas, <audio> and <video> (at least in webkit, anyone tried these in > Chrome yet?), and with John Resig's Processing.js, we have a good drawing > api as well, with many more surely to follow. What exactly about embedded > shoes would one up any of those things I wonder. > > How about this though, maybe this is better? Who wants to implement a java > applet that builds in jRuby and a fake shoes interface that replicates shoes > functionality via java's drawing api's, which thanks to Processing, we now > are all quite aware, do not suck much. Could be nifty, and doesn't require > users to install an obscure plugin with an even stranger name. > > On 03/09/2008, at 9:59 PM, Leslie Wu wrote: > > By now, some of you may have seen Google's new browser announcement, aka > "Chrome" (http://www.google.com/googlebooks/chrome/). > I think what's exciting about this development is that there will finally > be a modern, open source WebKit-based multi-platform browser engine that > supports a plugin architecture and holds the promise of easier (?) > hackability. > > Last year, I experimented with a whole bunch of different ways to better > integrate the browser-as-app-engine and desktop (http://jinsync.com/), and > looking now I see a blog post in May 2007 where I talked about the "future > of the web (app)" -- http://jinsync.com/?q=node/14 -- and alas it took a > while but Gears came out, and a year+ later, a beta version of Chrome is > out. (Interesting also that Chrome generally addresses the notions I > mentioned in March 2007 -- http://jinsync.com/?q=node/12) > > And while a RESTfully Organic FF didn't quite happen ( > http://lwu.vox.com/library/post/on-the-future-of-the-fox.html), I'm at > least hopeful that Shoes and perhaps HacketyHack will find their way mayhap > into Chrome. What do you folks think of that -- who wants to implement Shoes > as a Chrome plugin so that somedaysoon HH will be but a button clicks (3x) > away? > > I've been brainstorming interesting Chrome <-> Shoes interactions, and I > generally like the idea of being able to embed Shoes apps into webpages, > whether statically or dynamically, such that Shoes apps can hit back and > troll the DOM jQuery/Hpricot style and Chrome can send events to Shoes. > Maybe what I'm really saying is, why not Shoes where today we find Flash? > While I'm not proposing that Shoes take over (yet) the SWF world, Shoes does > have nice native platform integration in terms of Ruby / Ruby gems and of > course the ability to touch local resources through Ruby and/or > C-implemented libraries (Hpricot / Mongrel). > > But if that's off in the horizon, have any folks experimented with embedded > Shoes in Mongrel or versa vice? I'd like to be able to send data RESTfully > from my browser to Shoes apps (to do super colorful graphics and the > like)... > > ~L > > >
