On Wed, Mar 10, 2010 at 9:59 PM, Bhushan Inamdar <[email protected]> wrote:
> I am very much new to all this. But from what I understand, My 2 cents
> w.r.t. the following suggestions made by Omer in his post to Koen...
>
>> Koen:
>> 1) You could use some kind of tag instead like dojo does
>> <event attach="someWidget" signal="someSignal">
>> javascript here
>> </event>
>> A javascript parser is an overkill but it actually sounds like a nice idea
>> to have one as part of the build chain, it can be useful :)
>> 2) Sure, I will do so ASAP.
>> 3) True, again if there is a need for run time efficiency this should be
>> turned off.
>> 4) If you would have looked into my source code that I had sent you you'd
>> see a wrapper class that integrates the model into the widget, if you choose
>> to not use a model you can just create a regular widget.
>
> To add to what Omer and Koen have been talking in point (1) above, I would
> like to know if using Google V8's model for implementing javascript code
> would help? I am not sure about the kind of licensing restrictions Google
> has put for V8 use for commercial purposes, yet it might be worth looking
> at. The basic concept behind V8 is as folllows. Javascript by nature is a
> dynamic language as against the other object oriented languages like C++,
> java, C#, etc. This implies that a var declared in javascript can hold any
> data assigned to it. This is not the case with OOP languages which are
> strongly typed. Also, since Javascript is an interpreted language, it cannot
> have the speed and performance benefits as that of the compiled languages.
> However, V8 is a javascript engine that provides javascript in exactly what
> it lacks by the way of implementing what they call as "Hidden classes" under
> the hood. It compiles the javascript code the first time it is executed. As
> a result, when the same code is executed any time further, it will perform
> better. However, as of now, it does not provide greater benefits if the code
> is compiled and called only once. Google is already working on this problem
> and has its own algorithm to decide if the code should be compiled or not.
> This would really help plug some of the redundancies that Javascript brings
> along with its numerous advantages. There is more as to how V8 works, but I
> will leave it to the readers to explore.
> What I am hinting is creating or using something like V8 that suits Wt. I am
> not sure what language V8 is written in, but most probably it would be Java
> and/or C++. You might still need a parser I believe. But may be a better
> performance at runtime can be achieved at the expense of increased build
> times. It should, therefore be kept optional to use as mentioned in point
> (3).
> Please let me know if that makes any sense, as my knowledge in OOP languages
> interaction with Javascript is very poor...

Tracemonkey tends to be better with such things (Firefox's javascript
engine) and faster in the general case.

I do not understand why you would want to add such a thing to Wt
though, if you want to make a web page in some other worse language
then C++, why not just use one of the already existing platforms
already, there are a few javascript based server backends.  I like Wt
since it is pure C++, although I am not big on the Qt design of it to
be honest, it is C++.


On Thu, Mar 11, 2010 at 12:37 AM, Pau Garcia i Quiles
<[email protected]> wrote:
>> On Tue, Mar 9, 2010 at 6:10 AM, mobi phil <[email protected]> wrote:
>>> 3. The already proposed idea about using Qt Designer to design Wt
>>> widgets, might be a wrong direction. In my opinion a better solution
>>> would be to create a Wt application that would allow the construction
>>> of "empty" widgets, through a web application. The advantage of this
>>> approach is that contrary to Wt Designer, the widgets would render on
>>> the final web application as they render in the Wt. The product of
>>> this designer wt application would be a html template. Any .ui
>>> interface etc. would mean unnecessary overhead in the development
>>> cycle etc. etc.
>>
>> That is *exactly* how I thought that should be designed, even played
>> with it a bit (just a quick creator as I was designing my app so I
>> could figure out how they 'feel', *very* basic though).  This would be
>> fun to do.
>
> I disagree.
>
> "the widgets would render on the final web application as they render
> in the Wt..." *IN* *ONE* *BROWSER*, the one you are using to run
> WtDesigner. If you are running WtDesigner on Firefox, you won't be
> sure how it will look on Safari, IE7, Chrome or whatnot. Further,
> nothing prevents you from making the QtDesigner-based WtDesigner
> render the widgets on a QtWebKit canvas or replace "Show with
> Plastique style", "Show with CDE style", etc from QtDesigner with
> "Show in Firefox 3.0", "Show in Opera", etc. In fact, ideally
> WtCreator would automagically download and install (a-la
> portable-apps, ie. no system-wide installation) all those browsers so
> that you can easily test your design.

Why would it look different at all?!?  Does not Wt go through great
pains to make sure everything looks the same everywhere already?

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
witty-interest mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/witty-interest

Reply via email to