How do plan when starting a new project?

PHP
1

I’m wrapping up Kevin Yank’s course for beginning PHP/SQL (for the 2nd time ;)) and I’m curious how experienced coders plan/begin a new project.

When I first started creating websites some 15 years ago, I was taught to start out by creating a detailed flow chart illustrating the website’s pages and structure. Do others do something like that when designing a PHP/SQL driven site? Are there other ways to plan a project before beginning the coding? Do you just start coding with a loose idea in your mind?

I’m just curious how people start a project.

2

I always start with a plan. I typically layout a very high level of the workflow of the project, then start to define each step/process further and further until it is all defined. Once I have a complete definition, I ensure the business side agrees with it, and then begin programming.

That was a few years ago, now I do something similar, but as I define each step/process, I have the business side review it, and if they agree, I can go ahead and start the development of that step/process while still defining other segments (very similar to an Agile approach), which allows me to find issues sooner that may affect other parts of the project.

3

Glib answer: Does this site warrant using a Framework, if so which one and why.

Lets say you don’t want to use a Framework though, imagining the following scenario might help or hinder you, but it used to work for me:

Imagine your entire finished site is a bridge to a far off island. This land-side end of your bridge is the data (data store/database if you like) the other end on some island somewhere is the webpages (which include the necessary back-end admin pages).

You can see how far you can start building out from the land, so you imagine or start drawing up how your database tables will look.

You can then row out to your island and start building backwards from there too – this would be the wireframe flowcharts you already say you are proficient at. All of this so far might just be sketches on paper, but you are pretty confident that you can achieve those two ends.

Now, what is left in the middle (which might be in the clouds) connects the two ends of the bridge and that is now a bit hazier to imagine… a bit more threatening.

Now you look carefully and try and work out the bits you are not sure about, model those scenarios where you tackle the jobs you have never had to do before. I mean really put some data into some tables and try and get it out as you imagine you will be wanting it. Maybe this is all just SQL at this point, maybe you have some scripts.

Ideally at this point you might be isolating classes you will be needing, with luck someone else has already worked out how to do these jobs. This is where you can be more specific and ask pertinent questions on boards such as this.

So, you are kind of building from both ends and the middle at the same time.

Getting back to my “bridge”, though you start off thinking about both ends of the bridge you actually spend most time at the get go in the middle tacking the tasks which eventually dictate how you adjust those ends to suit the middle.

In doing so you should then hopefully have identified other benefits, other possibilities for how the data could actually improve the experience on the front end, and you can go back to the client either some new ideas, or at least some other addons you can offer over time.

The essential thing will be not how you store the data, but how do you get it out quickly and cleanly for the front end.

Despite all the tools and flowcharts and kit, we construct software in our heads first.

HTH