[eluser]Michael Ekoka[/eluser]
[quote author="esra" date="1183730592"]Impossible to gaurantee based on the simple fact that they won’t lose customers because of the glut of popular PHP4 open source applications bundled with hosting packages.[/quote]
Having a potential customer go to the competition because of lack of support is a loss. Having customers not renewing for the same reason is a loss. This week again one of our hosts had to write a long email to its customers requesting their patience, because some new features were coming by the end of the month, amongst which php5. Just a few months ago it didn't take me much to convince a host to install it for a client. So yes, hosts not supporting PHP5
are losing customers.
The main argument behind the continued use of php4 is that "
many host don't support php5", how true is it nowadays? or "
that's what the client is using", do clients really care what version of php you'll be using for new off the ground projects (CI is barely a year old)? I acknowledge that there are some legitimate constraints forcing the continued use of php4, but as I said in another thread relating to this, php is being a victim of its own popularity. Many people adopted it then because it was a simple language that required a low learning curve to get the job done. Today the language is trying to play catchup with its lack of features, but the community, which involves developers of various depths, is not following in unison.
[quote author="esra" date="1183730592"]
Why don’t those PHP5 frameworks exist now, given the long lead time since PHP5 was released? Symfony and Zend really don’t impress me as a OOP development foundation to base my future projects.
[/quote]
It's true that people include the language base as a criteria for the choice of framework they use. However I doubt that the fact that Zend or Symfony are less attractive options to some people than CI should be mainly attributed to language version. What pushes people away is the design and the degree of complexity. A framework is somewhat of a language in itself. How much more syntax, notions and convention can one bear to learn after learning php. Cake also is backward compatible with php4, but I find it a less attractive option than CI (or Zend).
CI has its own advantages. It is a light framework with a low learning curve, very good choice for minimal projects. However, frameworks like Zend and Symfony seem to be designed with experienced and somewhat demanding developers in mind. I'm convinced that after a couple of months tinkering with CI, these 2 frameworks will start being very appealing. Also, if you were to get a clue about architecture and OOP, I wouldn't cite CI as the reference and rather point you to these 2 frameworks that you say don't impress you.
[quote author="esra" date="1183730592"]
With an increase in the use of design patterns in PHP4-based CI applications and the use of Web 2.0 techniques, version 4 still has a lot of mileage left
[/quote]
From what I've seen across the different forums, I'd say that most people developing on top of CI and making use of design patterns, do so in php5. One cannot seriously talk about design patterns and willingly choose PHP4 over PHP5. Why would somebody go through such a pain. Pattern adaptations in PHP4 sometimes look like really ugly hacks. Sure PHP5 isn't perfect either, but it's a lot closer and makes many things so much simpler and clearer. Anyone serious about OOP or design patterns on PHP should move on, if they haven't done so yet. I have seen many interesting CI libraries from the community that make use of patterns and often they are "sorry! php5 only."
Now, I'm still at the stage where I'm going through the CI's source to see how things have been implemented. I've seen that in some instances the code forks when a hybrid adaptation is not appropriate (e.g. CI_Base class). Although my opinion about switching from php4 to php5 is unlikely to change, it is possible that after studying the code more carefully, I'd be more at ease that CI didn't sacrifice too much for the sake of backward compatibility.
