[wx-discuss] Hoyle reviews wxWidgets

[Kevin Ollivier]

Hi Julian and all,

On Mar 6, 2006, at 12:58 PM, Julian Smart wrote:

> At 20:49 06/03/2006, you wrote:
>> On Mon, 06 Mar 2006 20:06:27 +0000 Julian Smart  
>> <julian at anthemion.co.uk> wrote:
>>
>> JS> Do you think any of the existing samples or demos are adequate,
>> JS> or do we need a new one similar to the wxPython demo?
>>
>>  I think wxPython demo is great. It shows everything wx can do and  
>> does it
>> wonderfully well.
>
> I agree, I was so envious when I saw it first :-)
>
>>  I won't be at all surprized that relatively higher
>> popularity of wx among Python programmers compared to C++ ones is  
>> entirely
>> due to the demo existence. I'd love to have something like this  
>> for C++
>> library but, of course it would be quite an effort to reproduce it  
>> in C++.
>
> Yep. I feel another bounty coming on...
>
>>  These 3 would be a nice start. I'd also add at least one HTML  
>> sample.
>> Unfortunately this still leaves out wxList/TreeCtrl, wxCalendarCtrl,
>> wxGrid, wxMediaCtrl, wxRichEditCtrl, ... which are pretty  
>> important. But,
>> still, having few precompiled samples is much better than having  
>> none at
>> all.
>
> Yes; we can do the basics and them come back and package up
> some more. But really we do need the single wxPython-like demo
> since with separate samples it takes a lot of bandwidth for just
> a few seconds of "click, click, OK, I see".

What is wrong with the wxPython demo? We can easily make it into  
an .exe, .app, etc. and from an angle of seeing what wx looks like  
and what it's capable of, I think it's fine. And for those not  
adverse to Python, the ability to alter the demo in real-time and see  
how changes affect it is simply something that couldn't be matched at  
the C++ level. We could get it out the door really quickly, too.

At the C++ end, what IMHO is really needed are complete sample  
projects, with Bakefiles that are not 'hooked into' wx's build  
system, showing how to create an app, an XRC-based app, an MDI-app, a  
wx add-on library, etc. These apps would have skeleton About dialogs,  
Preference dialogs, etc. that matched with HIGs, etc. Projects that  
users could actually copy and paste to use as the start of their own  
app. We could also extract the standard dialog templates to their own  
folder so that existing wx apps could drop them in. (i.e. About  
dialog, Preferences dialog, etc.) Maybe have a templates subdir, with  
further subdirs of projects and dialogs. You'd be surprised at how  
often templates are used, if they're available. :-)

So what I see is a 'new user' process like follows:

1) D/l the wxPython demo or browse some screenshots (and maybe d/l)  
some popular apps, say wow, I'd like some of that.

2) D/l a wx package, containing release binaries where available (Win/ 
MS, Mac, RPMs?) in ansi/unicode packages, a few pre-compiled samples,  
Bakefile, and the templates dir (w/ already generated Bakefiles).

3) Run the installer. When it finishes, the default browser pops open  
a wx getting started page (don't know if we can do this on Mac...),  
pointing users to the compiled samples, explaining the templates dir,  
and offering a quick "here's where to find the Bakefile quickstart"  
pointer and a "here's the command to regen the makefiles". Include a  
brief mention of the most popular IDEs, and lastly also let them know  
where they can find the getting started page again. :-)

I think this would go a long way to improving the initial user  
experience for wx. What's more, we're pretty close to this. An HTML  
page would take an hour to write. As for the binaries, I'm finishing  
up the automated release tarball generator (it is running on Mac/ 
Linux, I just have to test some more), and the next step is building  
test binaries. If the test binaries work, then we could just run  
InnoSetup on the them and include them. We could set WXWIN  
automatically on Windows too using the installer (with the user's  
permission, of course) making this whole experience quick and painless.

Thanks,

Kevin

>>  OTOH, to be completely honest I don't think proposing complete
>> applications (be it DialogBlocks, wxDesigner or BitWise) is a good  
>> idea: it
>> distracts from the real purpose of the demo which is, well, to  
>> demonstrate
>> the library. And the objectives of a real app and a demo one are  
>> rarely
>> aligned so this would create unnecessary conflicts.
>
> I can see that angle but I can also see that people will want to
> see what wxWidgets is _really_ capable of in a real app. But we'll
> offer demos and pointers to real apps so people can choose to
> try what they want.
>
> Regards,
>
> Julian
>
>
>
>
> ---------------------------------------------------------------------
> To unsubscribe, e-mail: wx-discuss-unsubscribe at lists.wxwidgets.org
> For additional commands, e-mail: wx-discuss-help at lists.wxwidgets.org
>


---------------------------------------------------------------------
To unsubscribe, e-mail: wx-discuss-unsubscribe at lists.wxwidgets.org
For additional commands, e-mail: wx-discuss-help at lists.wxwidgets.org