develooper Front page | perl.perl5.porters | Postings from June 2006

Re: Its time we set the score straight on Perl 5 and Perl 6 and debunk our own self generated FUD.

From:
Dean Arnold
Date:
June 17, 2006 14:16
Subject:
Re: Its time we set the score straight on Perl 5 and Perl 6 and debunk our own self generated FUD.
Message ID:
4494711E.90806@presicient.com
> 
>>   Now we have wxPerl at what looks like a
>> mature stage, and can we learn from history and ensure it survives long
>> enough to build up a solid core of users?
> 

<rant>

FWIW: I do appreciate the efforts expended to provide wxPerl. My rant is
offered not as disparagement, but out of frustration. Hopefully,
it will be received as a friendly "wake up" call to its advocates,
because I and others really would like to use wxPerl.

But personally, I don't consider wxPerl "mature". Very capable ?
Yes. Very nice native look/feel ? Of course. Do I want to use
it ? Very much.

I've tried to use it, and after 2 days of hair-tearing and
cussing like a Russian sailor, I *finally* managed to get a
very simple demo with a cascading menu working. At which point I purged
my system of all traces of wxPerl.

Contrary to the frequent assertions that "its just like the C++ API",
it isn't. And even if it were, I code in *Perl*, not C++.
So, in order to figure out how to implement a cascading
menu, I eventually had to pour over the wxPerl source - which (at least
at the time) pretty much lacked any useable comments. Which seems to
be standard practice, since most of the demos didn't have any comments
either.

Meanwhile, Perl/Tk took me all of 15 minutes to figure out
how to implement the same feature. For that matter, I've managed
to randomly pull Javascript widgets from the web and figure out
how to implement cascading popup menus in a browser in
30 minutes or so.

Documentation matters. If the wxPerl community expects others
to jump on their bandwagon (and, really, I *want* to jump on),
they need to invest some time in POD generation (e.g., go to CPAN
and count the number of widgets that have any hyperlink...and
then read the docs for those that do. Trust me, it won't take long.
Then hop over to http://www.wxpython.org/ to see whats possible.).
If, as claimed, the API is nearly identical
to the wxWidgets base, I'd think the task would be a simple
matter of writing a little Perl script to touch up the wxWidget
docs into POD; hell, leave them as HTML, but show me how wxPerl is
used in *Perl*. Then supply some *useful* tutorials (no, showing
how to rubberstamp images onto a canvas isn't useful. Instead, think "Eclipse";
or take the vast collection of demos in the pTk bundle and convert
them to wxPerl).

After the mentions here, I bounced over to sf.net to see the
"documentation" promised on the CPAN page. Yeah, that was worth the trip...

App::GUI:Notepad is swell...but its hardly representative
of a large scale desktop GUI app. Until there are usable docs that don't require
pawing thru the wxPerl source, and then require me to translate
C++ into Perl, its likelihood of adoption by *developers* is pretty limited.
And based on dialogues I've had with others looking for a pTk
alternative, my opinion is not an isolated one.

</rant>




nntp.perl.org: Perl Programming lists via nntp and http.
Comments to Ask Bjørn Hansen at ask@perl.org | Group listing | About