It's about time that I start talking to you about WebDev! And how better than with an article I wrote about the differences between WinDev and WebdDev development. The WebDev develoment philosophy is VERY different than the 'normal' WinDev one. If you are not aware of the problem, you'll get in trouble.

Going from WinDev to WebDev should be painless, and on most levels, it is. One big problem remains: the development philosophy. This can sound like a big name, but you'll see very fast that looking at things this way can make you progress fast.

During a classical software development (with WinDev by example), the developer has everything on hands: the code and the UI are merged and everything is always accessible…

As soon as you open the WebDev code editor, you can see that something is amiss: browser code on one side, server code on the other, functions that are available only in one or the other, code execution order… There are enough new things to wonder… And justifiably…

WebDev tries to minimize the 'web' development difficulties and succeeds on many levels (see my article on AWP mode)… But one fact remains: without a different way of reasoning, your WebDev sites are going to drive you crazy.

The BIG difference that you have to understand is that a web site is NOT a piece of software… Even if some code is executed!

A web site is more like a collection of software talking together:
- a big one (it's the server side application)
- a lot of small ones (each html page, executed independently on the client machine, in Javascript)

If you are able to see things that way, everything is going to be easier:
- all the browser code area of a page are in fact one small quasi independent program
- the server code area of a page are part of the big server side program

In fact, compared to a traditional software, it's a little like if each window was compiled in a separate executable, and that all the file access code (and other) was in a big separate executable. Add communications between them, et voila…

Lets take an example: info("MyMessage") is a very basic instruction… But if you use it without thinking about it in server code, it's not going to be executed… At least not immediately… It will be executed only when the next page will be sent from the server to the client machine where all displays are done!

So, when in WinDev an 'info' is blocking the program execution, it does not in WebDev server code… Surprise! At least if you continue to think in terms of standard software development.
On the contrary, if you think in terms of several programs working together, this behavior becomes normal.

Let's take another example:
In a page, you want to display fields depending of the user actions. To this point, everything is simple, we are in a case of browser code…

But your conditions also depend of the content of some files… This means that you have to send the user actions (checkboxes values by example) to the server to verify some file values. We are now talking about server code.

It's the mixing of both that is troublesome if you don't think about it the right way… How many times do you see questions on a forum about how to execute browser code from a server side code?

Here is how things should be done:

- the user clicks on checkboxes
- this information is sent to the server through a submit button or code hidden in the checkboxes fields themselves
- the server side code goes through the files, compare with the field values, and decides to make 3 new fields visible, to display a message in one of them, to send new values to a java applet, etc.
- the server side code prepares everything by setting flags (global variables by example), storing values to display, etc
- the server display the INITIAL page a second time
- in the page init code, the flags are checks and action are executed based on them, making the fields visible, etc

Of course, this made the web page flash (you all know what I'm talking about here)… And the answer to that disgraceful flash is … AJAX … But a little advise… If you still have trouble conceiving a web site the way I just described, forget ajax for now.

Build your site using the standard way, understand WELL how everything works, and we'll talk about ajax later…

Complicated? Yes and no.. Again, the philosophy of the thing is 'two programs talking via the web'… Imagine two computers linked via two modems and you'll get a better understanding of what has to be done where…

Let's take another example of the profound differences between a standard software development and a web site… In standard software, you can often see a min window opening secondary windows, each of them opening tertiary windows, and so on… If these windows are modal, the user can access only one at a time et will have to go back the way he came. If they are not modal, most of the time the user will be lost and you'll have to change your UI.

In a web site, you can open multiple browsers (or popup) too… But there is no MODAL idea at all… this has two consequences:
- the user can act on ANY of the open windows, with the clear risks that will have on your application
- if he has a choice, the user will go on ANOTHER website instead of yours, one with a simpler UI… Users just HATE popups…

So please think of your web site application like a software having only ONE window open at a time… Everything necessary for your process should therefore be in global variable (or classes variable of course) and in the current window… Don't count on anything else and everything will be better…