Future of CodeIgniter?

[eluser]nfx-nano[/eluser]
About CI 2, will it be PHP5 only? More OOP power? Will there be a huge overhaul of the framework (ie. structure), or will things remain in a general fashion as they currently are, except for maybe a few additions and changes? If I were to start a project using CI's current version, would it be easy for me to make the transition to CI 2? Or would I have to change and fix up a lot of code?

[eluser]xwero[/eluser]
CI 2 ??? I think next version will be 1.6.3 or maybe 1.7. Don't expect too many changes in the near future.

They were talking about looking into changing the validation and pagination library. The developers at Ellislab are not so keen on creating a non backward compatible version, i think because nobody likes to see apps/sites break caused by an update of the tool they use.

The update process of the framework has always been a breeze so i think that is not going to change. They even made pages with the instructions to update to each version.

[eluser]Colin Williams[/eluser]
I'm not sure I reflect the sentiment that the next CI release won't be big. I think a lot of things are changing, and for the better. The mantra is "if EE needs it, CI gets it." Imagine what EE needs!

I would say that right now, CI is a strong framework to be invested in. People don't love it because it necessarily does everything imaginable. People love it because of the way it does things.

And think about this: EE is EllisLab's flagship product, and they have made the decision to grow it upon CI. If it's good enough that they will build the future of their business (capital and reputation) upon it, don't you think it's good enough for you?

[eluser]xwero[/eluser]
One of the people at Ellislab mentioned EE2 is using stock CI for the most part. It would be stupid to base EE2 on a framework that will get a big overhaul, and as we know people at Ellislab aren't stupid.

I think the slogan is a look beyond the initial EE2 release. Now they will have a lot of work to get EE out of the door as a stable application.

I like CI because it's a low threshold framework. You can build your knowledge by working with it. Seeing the good and bad sides and try to come up with ways to improve it. I don't want framework developers to hold my hand while i'm playing with it. I think that is why people who like CI like the ZEND framework too. And the Symfony developers understand it too.

Back on topic now i would welcome a big overhaul but i keep my expectations to a minimum for the pre EE2 release. CI is still a framework other applications than a CMS can be based on so introducing CMS features to the framework would be the wrong way to go.

edit : EE is also a forum and wiki as mentioned in the footer, my mistake

[eluser]johnwbaxter[/eluser]
I don't normally get involved in these sort of discussions but i must say i'm intrigued how they have done it / doing it.

I'll admit that i haven't had much of a look at EE's codebase so i may not be the best to make these judgements, but i wonder how they are going go handle it all. I know CI is based on a lot of EE libs but i still think it can't be that simple. I wonder if there is some form of abstraction layer inbetween CI and EE to enable them to work happily together.

I would love to know how it is done, sort of the same way how i wonder how some magic tricks are done.

[eluser]xwero[/eluser]
i'm not sure what kind of things you think EE can do. It's just an application like the ones you are developing. EE is not some magical beast that flies through the sun

If i put my black hat on i'm wondering how they are going to prevent people from using pieces of EE in order to circumvent buying a license, if you look at an app you can't see its guts. I think it will be more temping for low budget developers to do this sort of thing once EE is based on CI.

[eluser]Milos Dakic[/eluser]
Some good discussion going on here, very happy to see that people are as curious as I am.

The thing that I expect to see in future releases is for CI to become more modular, so that developers can integrate their modules easily with other developers. This means, getting bits and pieces and putting it all together to get what you want. Might not be the best example but that’s what’s on my mind right now.

I like how ME5 lets your create modules with ease. This is what the framework should be doing IMO.

It would also be nice to see more libraries added that resemble what Zend is doing, but as I mentioned in my previous posts, it could be a bit too much for developers. There where I’m looking at gathering some info and making it easier to get the things you want out of CI without any hassles.

[eluser]Scriptor[/eluser]
Awesome discussion going on here. I like Sock Puppet's explanation of why EllisLab is still doing great, even though it distributes an unobfuscated (I assume) version of EE with all of their plans, and they have a free plan as well. Definitely a business model worth looking into. Besides, I don't think EE is intended for low budget developers. There will still be plenty of fairly well off clients who are willing to pay the extra for their developers to get all EE has to offer.

Now, to the topic of where CI's going. There will be some new libraries but I highly doubt that besides maybe a Javascript wrapper any of them will get in the core. EllisLab has done a good job with including almost everything a PHP developer would need to build a decent and fully-functional web app. If we're talking about 1.6.3 or 1.7, then (I hope) they will come out with a few tweaks to the validation class, mostly in displaying errors more nicely.

It's simply awesome that EllisLab is deciding to rewrite EE with CI. If and when they show that the framework works well with commercial-grade (enterprise?) software, AND they are willing to provide support for any problems...that tells you something. Although CI is fast already, there is always room for improvement (it's still got some overhead, after all) so the developers might find novel ways of speeding things up. I am sure that bugs and security holes will be fixed faster now since their own software depends on it.

[eluser]nfx-nano[/eluser]
Ok so if I get this right, EE2 is being rebuilt on top of the current CI framework, and things are being changed and added to the framework as EE2 progresses. Does that mean that they have started EE2 from scratch, using just a couple of libs from EE1? I'm just trying to understand how it's being developed.

I am wondering about how much code refactoring is being done, and the influence it will have on the workflow of it's users. If I have a project going on, it would be important for me to know how much will change in order to make the right decisions. I'm just trying to analyze the risks involved.

Right now the community doesn't know what to expect, and that's maybe because the EE developers have perhaps not set any expectations for it, or are aware of the fact that these expectations may change at an invariable rate, as is the world of web development, though it is always good to know the global direction of where things are headed. It's not necessary to know each and every specific detail of every adjustment, just the big picture, you know? There's just a lot of guesswork going on right now. And I know the developers might say, well if we tell you about feature x and adjustment y, and things do not go the way they were supposed to go, then you might get upset. Well, maybe it's better to upset 5 people, rather than leaving 500 people in the dark, not knowing what they can expect.

Example, we can assume that CI is a tool we can all use, so let's compare it with a flamethrower, since that is a tool also (ok, a weapon, you get the point). If I'm using a flamethrower to destroy a couple of buildings, how will I know that flamethrower 2.0 will not spit water instead of fire? That would be garbage. And how about if flamethrower 1.0 is designed for your right hand, and flamethrower 2.0 is being designed for your left hand, then it would be good to know this so I can train my left hand upfront. Perhaps 2.0 spits fire 10 meter while 1.0 spits fire 5 meter, so then suddenly 2.0 would have doubled it's scope. The same thing may be for CodeIgniter. Who knows?

[eluser]kylehase[/eluser]
Some of the posts in this thread mention the idea of "PHP5 only" for future CI versions. I know one of the core features of CI is backwards compatibility (PHP4) but at some point it won't be necessary to support PHP4 right? After all PHP3 is not supported.

My question is, if in the future PHP4 support is dropped would you expect performance improvements in CI? I assume there are snippets here and there that use less than efficient code to support backwards compatibility. ...wonder what will happen to Kohana at that time.