Professional PHP

My last WACT post drew a couple questions from Marcus Baker:

Do you have a release date planned for the next version?

Ah, the release date question. We are certainly overdue. Here is where things stand.

The core framework in CVS is in very good shape. We have 3332 passing tests and 5 failing tests. The failing tests are relatively minor. I would like to fix them before release, but don’t see it as absolutely necessary.

I feel the biggest obstacle right now is that many of the example programs that we have in WACT are either obsolete, broken or unfinished. They are mostly broken due to the incompatibility of changes within the controller layer from the last version. I would like to see working (and hopefully unit tested) examples for the release. I don’t want to make a release with code that errors out.

The current controller architecture in the CVS version of WACT is very different from what is in the previously released version. After writing what is there now, i have modified my thinking on what a controller framework should look like. I’ve written up some of the theory of this on this Model View Controller page. I’ve been working on a proof of concept of this, which I think will become the final WACT controller layer.

As such, I am reluctant to go through the examples now and update them to the current controller when I know that I will have update them again. I am also reluctant to expose a brand new API in the release which I already know will go away.

I’ve decided I want to work on the controllers and not the examples right now. That of course doesn’t preclude anyone else who wants to work on the examples from doing so and getting us into releasable condition. Thats just how I want to spend my time right now. If we are going to release with what we have now, which i am not opposed to, someone else will have to step up and push that.

Its possible that my controller ideas won’t work out. I’ve just reviewed the 3.0.0dev version of Mojavi. I have to say I am impressed. I have looked at most of the controller frameworks on this page and i have to say that Mojavi is the only one that i would seriously consider using myself. It has the right combination of support, maturity, and good design. If my experiments with MVC don’t pan out, i would seriously consider ditching the current WACT controller layer in favor of Mojavi integration.

So thats where I am at. I would personally like to see more of the controller issues resolved before the release goes out the door. I would like to feel as good about the controller layer of WACT as I do about the rest of it.

I should finish up with a project soon and have more time to devote to WACT.

When will the next WACT release be? I refer you here.

Marcus also asks:

The bit that I’ve never understood is, what is a DataSpace?

The DataSpace concept has evolved over time. in WACT, DataSource is an interface that allows an object to expose properties though a generic protocol. The DataSpace class is a convenience base class that helps implement this interface. The DataSource interface could be considered equivalent to the properties portion of the Java Beans specification. You could think of any object inheriting from DataSpace or implementing DataSource as a bean in java terms. However, I would probably compare it to the more dynamic Key Value Coding from Objective-C.