why not "divided and conquer"

[searain]

I saw many systems use this approach - "master page"

put all the actions and the display of all these actions (add/delete/edit ...) in one page, and to tell the difference, the action is passed as a parameter. and they use switch statement to tell what action this page should take and what display results it should give.

why not separate these pages into different individual pages, like "divided and conquer", add action will have an add page, and delete action has a delete page, such as

<?php
$id=$_POST['id']
//delete sql here

// redirect to the right page for the results display

?>

this kind page will not be see by the end-user, it will just act in the background/middle step.

but it is much clear than bury all these actions and different action results display pages in one "master page".

[vaaaska]

Actually, I'm currently doing both at the same time. 😉

[pohopo]

IMHO,

if you had a page for each function you would have a hell of a lot of pages and that is much more confusing. what most do is create classes where they group similar functions of objects together. when data is passed to a page you usually have some conditional operator that determines what function in your class to use.

as long as you built your classes correctly it is much easier to not only find your functions, but you can also reuse on multiple pages.

[bubblenut]

I agree wholeheartedly with pohopo, switch statements are great if you've got your class/function structure well built. However, I would like to stress that if you don't have that kind of structure switch statements can be a nightmare. I've seen code where each case of a switch statement was about 100 lines of code and it was horrible. Even knowing what the script was doing behind the scenes I still found it hard to follow. Moral of the story is to spend more time designing the layout of your structure than the code itself if you're doing anything more than just a little script.

[Weedpacket]

Indeed, why not? Then again, why? The idea that there has to be some sort of trivial mapping (such as identity) between files and URLs is a fallacy of tiresome persistence. The structure of the web site and that of the filesystem need only have enough to do with each other that the process of identifying which files are required to satisfy which URL requests is comprehensible to the poor sod who has to maintain the site.

[vaaaska]

Directly and indirectly I've seen this topic pop up in alot of places lately. In another thread here somebody was complaining about Smarty - I believe this discussion is similar.

A few weeks ago I labored (with the help of Weedpacket) to create a function for calling up functions from a library. It basically uses file_exist to see if the page is there and then passes the arguments to the function.

fin('theFunction',$arg1,$arg2,$xnumargs);

Before I did this the speed of things was around 0.xxx seconds and now it's around 0.0xxx seconds so something worked.

I also created a switchboard like layout which grabs URI information (yes, it's validated big time) and then if the function isn't in the registry it will go to a default scenario instead. So, if a user wanted to enter illegal characters they will simply see the same page over and over. I don't know if it's good or what...I'll have to wait for others to tell me later...