GWT generates overbloated JavaScript ? Where do you have that one from ? GWT's Java to JavaScript compiler is one the best out there (The closure compiler beeing imho the best). And GWT 2.5 just added the closure compiler as a compilation option.
2012/8/30 Carlos Rovira <carlos.rov...@codeoscopic.com> > AFAIK Google has made the same with GWT as Adobe with Flex. But GWT > has the problem to generate overbloated JS code... > > 2012/8/30 Nick Tsitlakidis <ni...@perfectedz.com>: > > Hello guys, I'm following all the topics here but I post rarely because > > most of the times someone else has said something that I agree with 100%. > > > > This time though, I was trying to think about similar technologies which > > are either compiled to js or they are converted in js in some other way. > > So I thought about GWT. The appproach google has taken with it is very > > similar to Flex. They even have a skin architecture equivalent. > > What I'm trying to say is, what if we could achieve something similar. > They > > seem to be translating Java to JS without a problem because they exclude > > Java features that are not compatible. > > It's a small Java subset, I'll give you that, but developing in Java and > > creating skins just like in Flex is way more interesting and agile > compared > > to pure HTML and JS. > > > > As far as I can tell, both languages are not that different (Java and > AS3). > > > > Any thoughts on this? > > > > On Wed, Aug 29, 2012 at 10:55 PM, Om <bigosma...@gmail.com> wrote: > > > >> On Wed, Aug 29, 2012 at 12:19 PM, Michael A. Labriola < > >> labri...@digitalprimates.net> wrote: > >> > >> > >Can you please elaborate? > >> > > >> > >The point I was trying to make was that HTML5 language itself is not > >> > designed to be extensible. Using Javascript does not really count (in > >> this > >> > >context) > >> > > >> > >As far as using the DOM, I assume you mean the Microdata format. > This > >> > results in non-standard HTML most of the time and is not supported > across > >> > browsers. And it deals more with extending data semantics and >not > >> > functional extension. > >> > > >> > In flex, IMO, we worried too much about extension and not enough about > >> > composition. > >> > >> > >> I think that is besides the point. There is nothing in MXML that > prevents > >> composition. It is just that the current set of Flex components are > built > >> like that. We can fix that given time and effort. There is no need to > >> structurally modify MXML to achieve this. > >> > >> Whereas with HTML(5) there is nothing in the standard that will let us > do > >> specialization (via inheritance or composition) I cannot dream up new > >> elements and expect a browser to understand it out of the box. > >> > >> > >> > So long as I have a good series of patterns (and the discipline to > follow > >> > them) then I can look at the HTML DOM elements as the Atoms of the > >> universe > >> > and assembly them with some bonds (JavaScript) to make an element. And > >> then > >> > in turn assemble those to make any application. > >> > > >> > >> Right, we need Javascript to do this kind of extension to HTML. To do > this > >> in the Flex world would mean that we either > >> > >> * Bring in JS as a language we support in Flex > >> or > >> * Keep Flex as it is (i.e. Actionscript based) and have a AS to JS > >> translation layer. > >> > >> The latter is a better approach because of various reasons ranging from > JS > >> not being a real OOP language, no package organization possible, etc (we > >> all know why AS is better than JS) > >> > >> I think being able to code in MXML and Actionscript would be a key goal > of > >> this cross-compilation effort, right? Unless we want to fundamentally > >> change what 'Flex' means. > >> > >> > >> > So, the key is not trying to extend the Atom but trying to assemble > it in > >> > useful ways and allow those to be extended or recomposed. So far, I > have > >> > found few limitations of this approach and often times ended up much > >> > happier. > >> > > >> > > >> I definitely agree with you on this. But again, this requires > Javascript > >> to assemble things. My above points still hold good as well. > >> > >> > >> > Mike > >> > > >> > > >> Thanks, > >> Om > >> > > > > > > > > -- > > Nick Tsitlakidis, > > > > CEO and Software Architect at Perfect Edge LTD. > > www.perfectedz.com > > > > -- > Carlos Rovira > Director de Tecnología > M: +34 607 22 60 05 > F: +34 912 35 57 77 > CODEOSCOPIC S.A. > Avd. del General Perón, 32 > Planta 10, Puertas P-Q > 28020 Madrid >
