"Leon Atkinson" <[EMAIL PROTECTED]> wrote:
[...]
> When coding, it doesn't really bother me when I mix HTML in the code. I can
> see through the layers of abstraction. And having starting doing CGI in C
> back in 94, I'm comfortable using print statements to output all the HTML.
> But I do see value in pushing that HTML into it's own space to make it
> easier to change things, so in the 2.x tree I've built out functions for
> common GUI elements (like selectors, text boxes, etc) like we have with
> StartForm() in the 1.x tree.
I think the problem I had was more that the HTML was all around, and
it wasn't easy for me to figure out what parts of the finished page
came from where in the code. Sometimes it came from a function
somewhere, sometimes it was inlined, sometimes it was in a
layout/navigation, etc. In the end I just did lots of grepping, and
that worked most of the time.
Mostly, it's not that prints are bad or anything, but that there were
too many places where HTML could get on the screen. Maybe it would
have been easier if, for instance, the screen would fetch and arrange
any information necessary for displaying the screen, then there would
be a seperate section for creating the HTML to represent that
information. It would be kind of an expansion of the action/screen
seperation. More like act/process/display.
I think it's some of what a template system tries to get at, only
without creating a mini-template-language, and without trying to
forsee all possibilities by creating a hard boundary.
--
Ian Bicking
4869 N. Talman Ave., Chicago, IL 60625
(773) 275-7241
[EMAIL PROTECTED] http://www.colorstudy.com/ianb
------------------------------------------------------------
To subscribe: [EMAIL PROTECTED]
To unsubscribe: [EMAIL PROTECTED]
Site: http://www.working-dogs.com/freetrade/
Problems?: [EMAIL PROTECTED]
