pyFlash

[Nigel Head]

rnd at onego.ru (Roman Suzi) wrote in
<mailman.997339095.27883.python-list at python.org>: 

>On Wed, 8 Aug 2001, MNX tech services wrote:
>
>>Would it make sense to use static Flash 5.0 pages as a browser-based
>>GUI against a Python backend?  Wouldn't it have the advantage of
>>displaying absolutely uniformly across all platforms, along with
>>providing a vast array of widgets?
>
>If you code for the WWW, then it's probably not good idea. Personally I
>avoid flashed sites, .... [snip] ...

And there you have the problem: this comes down (partly) to a marketing 
decision -- the community appears to be split between the purists (who have 
a good point, given the regular security problems with most forms of client 
side scripting) and the rest, who are prepared to accept some of these 
client side technologies.

The other part to the equation though is Flash itself: it seems to have 
some unfortunate bugs, they only get fixed VERY slowly as well. The list is 
longish, some samples -- XML sent with the wrong content type (fixed in 
latest player, but how many have upgraded?), GET instead of PUT under some 
circumstances (not consistently fixed, AFAIK), Content-type header value 
with trailing garbage attached (my latest discovery). And so on... 

Given this I'm seriously considering doing some of my work with Jython 
generated applets instead -- specifically any form based dialogs needing 
client side validation or other forms of smarts. As well, I like to use 
XML-RPC and doing that in Flash, although possible, is more painful than it 
ought to be. I haven't tried with Jython yet but I figure it ought to be 
OK.

Where Flash really can't be beaten is the presentation/graphical/design 
type sites ......

Nigel.

BTW (off topic): Anyone played with CURL yet?