advice on starting new large project

Matt_Kindig

Hi everyone,

This is my first time posting to Echo Lounge, so my apologies if this is not the appropriate forum.

I have been a PHP hobbyist for about two years, learning the language and developing a few small scripts for friends but nothing big or commercial. I have recently been hired to develop what's now a relatively small website, but may grow in the future.

My question is: does anyone know of any good articles that describe considerations and advice for developing PHP sites? I tried searching this board and the web but couldn't find anything particularly useful. I looked into a number of CMS, but my client didn't want the rigidity of those solutions. Mainly, I'm looking for articles advising on:
-templates/stylesheet separation from content
-scalabilty (how to make sure I can add new features easily)
-security issues

Any articles that you can point me to would be much appreciated. Thanks for your help!

Matt

vaaaska

Typically people are full of advice around here...?

Form follows function. Use of smaller pieces of code/functions to do things seamlessly will (or it should) allow you to make smooth upgrades in the future.

That being said, tutorials are good and all, but they too need to address the issues you've mentioned to be worthy. I thnk you are on the right track by researching other (open source) applications and understanding how they get things done.

I'd recommend putting together database and validation classes first so you can focus on the details...probably an authentication class would be handy as well...

bpat1434

As far as templating/style separating: Creating a Template Parser Script

As for scalability: It's as easy as you make it to expand or add "mods" or code.

Security: Lots of issues, but only if you're stupid. Since you've been in it for two years, you should be fairly versed enough to know that you should RegEx any inputs to make sure it's not a hacking attempt. Watch the CHMOD and don't do the stupid n00b stuff.

Really, templating is easy, the hardest part is the design. Once you have the design, you code to what you want. Then worry about molding them together. If you want to see my templated website, visit http://www.winfieldvfd.org

Form does follow function; however, it can be vice-versa as well. Just be careful, and think about what you're doing before you do it.

~Brett

pohopo

If you want to do a larger project the most important thing you should do first is create a project plan and UML documentation. If you can't put what you want to build on paper first then you probably are not totally sure what you are building. (at least from an architects point of view)

AstroTeg

I don't know if a full out UML is needed. But I do agree you should jot some stuff down on paper. Things like a list of the main nav links/pages, hurdles you think you may have problems with, and any unique ideas you might forget when writing the code. Depending on your comfort level, jotting down the steps you'd take to develop the site may be useful as well.

Templating: I hit the Creating a Template Parser Script tutorial and its a start. Although I find ob_start() and its band of functions to be sluggish. You really only want to use that function in case of emergencies (in my opinion). I'm sure there's are alternatives out there and if not, come see me.

I do recommend implementing some form of a templating system. Be it one you write or an off-the-shelf-one - although I think if you write it, you'll have a very intimate knowledge of how it works and will be able to better plan sites around your system. Templating will allow you to seperate the PHP code from the HTML which will help greatly when it comes time to troubleshoot the one or the other.

If you're doing database work, make sure you spend some time planning it out as well. Figure out what data you need to store, where you'll store it, and what tools you'll use to shuffle the data around in your PHP code.

As others mentioned, spend some time figuring out how to validate user data to ensure it doesn't muck up your site or server.

You might spend a little bit of time poking around established PHP projects and seeing how they handle security, database connectivity, and templating. Determine what you like and don't like about their approaches. Then figure out if you want to build your own incorporating your ideas or go with the solutions found in the projects.

Lastly, test, test, and test the new site. Then have other people test and test the site. Get feedback. Make changes. Test again. Of course, a simple site won't need much testing.

Good luck! Just remember to take it one step at a time and your project will turn out fine...

dalecosp

Originally posted by Matt Kindig
Hi everyone,

This is my first time posting to Echo Lounge, so my apologies if this is not the appropriate forum.

I have been a PHP hobbyist for about two years, learning the language and developing a few small scripts for friends but nothing big or commercial. I have recently been hired to develop what's now a relatively small website, but may grow in the future.

"Have been hired ..."

My post will be OT, but this is the EL, after all: is this a salaried position or is the "hire" a contract? There are any number of folks here that might remind you to spec this thing out on paper real carefully (if it's the latter, and maybe if it's the former, too) due to the phenomenon known as "feeping creaturism". :eek:

As for reading other project's source: go for it, of course; but does anyone here feel like often times large open-source projects don't do too well at commentary? (phpBB stands out as a notable example in my mind....)

piersk

Just a minor point, for the templating, have you tried smarty?

pohopo

I don't know if a full out UML is needed. But I do agree you should jot some stuff down on paper.

Basic UML does not take long at all, yet the benefits you get from this are huge. Without it a project could easily take 2-10 times longer. I can't stress documentation enough. Too many people think that documentation is something you do after so the next person is not lost. That is secondary to the real need which is to help build the system. Imagine a person building a house without a blueprint? It is the same thing with building applications.

kitonia

Dynamic Web Forms by Dan Ransom, I think you can find it on AMazon. It's a good book that discusses everything you are looking for.

goldbug

Originally posted by pohopo
Basic UML does not take long at all, yet the benefits you get from this are huge. Without it a project could easily take 2-10 times longer. I can't stress documentation enough. Too many people think that documentation is something you do after so the next person is not lost. That is secondary to the real need which is to help build the system. Imagine a person building a house without a blueprint? It is the same thing with building applications.

And using a tool like ArgoUML, you can export empty PHP4 or PHP5 class skeletons to which you add your implementation, saving you even more time.