Posted by Richard Lynch on 05/12/05 04:36

While this *CAN* work, and a lot of people like it, it tends to add a fair
amount of "cruft" for not that much benefit, really...

What do you GAIN having this big old switch statement?
What data/processing is really really shared in all these steps?

On Wed, May 11, 2005 4:57 pm, dan said:
> Hello, all -
>
> I've been researching how to handle forms properly, and I think I
> figured out a way that might be beneficial for me to use. It is as
> follows:
>
> (index.php)
>
> session_start();
> if (isset($_SESSION['step'])) {
> switch $_SESSION['step'] {
> case "1":
> require('step1.php');
> break;
> case "2":
> require('step2.php');
> break;
> case "3":
> require('step3.php');
> break;

Simpler:

case "1":
case "2":
case "3":
require "step$step.php";
break;

Also not that it's pretty unlikely that your default will kick in, since
somebody would have to intentationally hack $step to be, say, 4 or
something to reach that line of code... [more]

> // add more case statements here if I need to
> default:
> require('step1.php');
> break;
> }
> } else {
> $_SESSION['step'] = '1';
> require('step1.php');
> }

So you might want to start off with:

$step = isset($_SESSION['step']) ? $_SESSION['step'] : 1;
Then you could do your switch, and default to step1, thereby having all
the same functionality, but with fewer branches in the actual algorithm.

>
> Each stepX.php file would look something similar to this:
>
> (step1.php)
>
> // if submitted, check data for completeness
> // if complete, set 'step' to 2, to be used as argument to index.php
> $_SESSION['step'] = '2'
> // redirect back to index.php, use new value of 'step' to direct
> header('Location: http://somesite.com/index.php');
> // else display form data
>
>
>
>
> Now, this is, really, one of my first experiences with doing forms. I
> just want to know if I can/should/would anticipate any problems down the
> road while doing this. I think it would work quite well, but I've only
> been doing this for a short while.

I did it this way at first, but quicly found that the amount of shared
code between steps was so minimal, that it was better to just have each
step as a separate form, with filenames that made sense for the
information gathered at that stage.

It's also sometimes good to let the user fill in different steps in
whatever order they prefer -- depending on the data gathered and what your
business goals are.

I've seen a particularly nice implementation of this at CDBaby (which is
not real useful to anybody reading this unless you happen to be a musician
with a CD you want to sell...) where the steps can be done in any order,
but the first five are MUST DO and are flagged as such in RED until you do
them, and the last three are OPTIONAL and are in yellow until you do them.
Completed steps are changed to green.

Since some of the steps could require a fair amount of background work
(writing/editing/fixing-up your Bio, for example, or getting a complete
track listing with titles in order) he lets you do them in the order that
fits into your life, not in the order he happened to program that morning.
Very nice.

--
Like Music?
http://l-i-e.com/artists.htm

[Back to original message]