[BRENSH]

I am wondering to all you experts at web design. How long would it take you to develop a large social networking site or content managment website from the ground up.

Or do you develop your web applications from the ground up but use repositories such as PEAR for PHP. Im sure .net can be used also.

For some reason I always feel like Im cheating (only an amateur) when I use other peoples code (I never do though) even though it increases development time.

What do you all think about that.

Thanks

[forbairt]

I think writing your own code is all well and good BUT you are automatically adding a hell of a lot of bugs whether you think so or not.

One example i'll use is http://www.maani.us/charts/index.php the PHP/SWF Charts code. If I want to create something like that its going to take me probably months to develope. To bug fix is gonna take a hell of a lot longer. Sure I'd get a basic system up and running pretty quickly but it just wouldn't have the amount of features available with that.

Code reuse makes for less bugs in your code assuming of course you're using tried and tested code. So I'd have to recommend you not feel bad about using others code assuming the license allows for that. (in some cases of course it just won't).

PEAR / Cake / zend framework are great ways of speeding up your development time.

For something like a social network ... you've be talking about anywhere from a few weeks to having a working prototype to months.

Depending on how scalable you make it all will have a big influence on this. As well as all the nice little features we now expect by default from these types of things.

[ianhobo]

Every programming lecturer I ever had, always said....

...Don't go re-inventing the wheel!

Writing your own code is great for learning and appreciating how things are done, but theres no point using others peoples code if you don't completely understand whats it using.

Re-using code is pretty much a skill you'll learn as you move along

[CuLT]

Part of moving forward is embracing what those who have gone before have provided.
Sometimes the thought of building my own variation on something from the ground up is appealing (or necessary), but realistically, if everyone had excessive pride and did this all the time, we'd be carving our websites onto stone slabs

[BRENSH]

Yeah I see where you are coming from. I can understand code very quickly. I have an inate ability to do this. I have spent about 6 months learning PHP to this point. Not enough on javascript. I am fairly well equiped. However my ability to plan a project is useless. Developing classes and so forth. Especially when tiering. I assume usning the zend framework would really improve the time on a project.

Is Zend the best framework out there. Should I use others.

[forbairt]

personally I quite like cake ... its really up to you ..

Zend however has some form of commercial backing with the zend suite of software so the choice really is yours ...


I'm using Eclipse + Cake

[louie]

I prefer to create things from the ground up. This way if anything goes wrong I know where to look, but saying that, many times (e.g. blog) I used Open Source code as there is no need to invent the wheel again.

Time ... depend on the project - could be days, weeks or even months.

[BRENSH]

Oh my impatience. I do realise it takes so damn long to develop a website. If I may deviate from the point can I ask which system do you use when developing. I am currently in the process of adopting a 3 tier system.

This seems the easiest way. Develop a couple of relevant application classes, a fully fledged DB class and a bit of (html + css + php) in the presentation layer.

Seems the most logical route.

I am doing it for my new improved radioportal. Development time seems to have reduce 10 fold.

What do you all think.

[zardette]

Hi Brensh,

When developing a project, it can be better practice to model your system in a language independent form.

How did you capture your project requirements?

Why have you chosen a database over a flat text file or xml - sometimes simple is better.
However, if you determine that a database is more suitable to your needs, what type of one will you use.... Access, mySQL? How will you model the entries and mappings of your database?

Why are you using PHP over ASP.NET ,jsp or other web development languages..... what’s are your motivations for the choice of language?

..and if you use other peoples code please test it esp if its source from the internet... It may seem to have the functionality you need but alot of source code out there isn't fully tested and is bug ridden!

(oh and btw your are not cheating its called "software re-use"!)
Z

[The Corinthian]

Two points; first what do you mean by a social networking site? A bulletin board application? A Bebo-type home page directory? a networking or dating site?

The term social networking is a bit of a cliche at this stage, with some being relatively simple in design and others being far more complex - meaning there is no 'one answer fits all' to your question.

Second point I would make is while one should not have to go and reinvent the wheel, neither is there much point in also being a copy cat. If you reuse (modified) code, you're simply going to end up just being much the same as the other guys using the same software - albeit with perhaps a sightly different angle. At the very least it will bring down the value of your long term IP.

[BRENSH]

Lets say the application is going to be advanced. Thats all I am going to say about it. Need to learn a bit more about architecture though.

When developing programs do you all use MVC (model view controller) concept or are there other types of architecture out there.

Thanks.

[forbairt]

Maybe find yourself a copy of ...

http://www.oreilly.com/catalog/hfdesignpat/

it makes for some interesting reading

[zardette]

just a note on design patterns: while they are a good design practice,often a novice programmer (like myself) can force-fit them into a software design where there is no need for such complexity, and makes the solution more complicated. a bit like applying an Object Oriented Solution when a simple procedural programming might be a more optimal solution.

For me they just seemed to be very complex to understand but maybe thats just the book I used although the girl on the cover of that book seems a bit confused to me too......

I think they are something to consider when you have been exposed to using OO programming in larger projects, and are confident using them.

This article maybe of use too
http://www.perlmonks.org/?node_id=402070


Z