The Model is a complex subject. However, it is often boiled down to either a single model class or a full object relational mapping (ORM). I personally have never been much of a fan of ORMs as they tie models to the underlying database structure; I don't always use a database, nor do I want to rely on an ORM solution too heavily on the off-chance that I later need to refactor to use services or another type of persistence store. On the other hand, the model as a single class is typically too simplistic.
In my last post, I discussed using Zend_Form as a combination input filter/value object within your models. In this post, I'll discuss using Access Control Lists (ACLs) as part of your modelling strategy.
ACLs are used to indicate who has access to do what on a given resource. In the paradigm I will put forward, your resource is your model, and the what are the various methods of the model. If you finesse a bit, you'll have "user" objects that act as your who.
Just like with forms, you want to put your ACLs as close to your domain logic as possible; in fact, ACLs are part of your domain.
A number of blog posts have sprung up lately in the Zend Framework community discussing the Model in the Model-View-Controller pattern. Zend Framework has never had a concrete Model class or interface; our stand has been that models are specific to the application, and only the developer can really know what would best suit it.
Many other frameworks tie the Model to data access — typically via the ActiveRecord pattern or a Table Data Gateway — which completely ignores the fact that this is tying the Model to the method by which it is persisted. What happens later if you start using memcached? or migrate to an SOA architecture? What if, from the very beginning, your data is coming from a web service? What if you do use a database, but your business logic relies on associations between tables?
While the aforementioned posts do an admirable job of discussing the various issues, they don't necessarily give any concrete approaches a developer can use when creating their models. As such, this will be the first in a series of posts aiming to provide some concrete patterns and techniques you can use when creating your models. The examples will primarily be drawing from Zend Framework components, but should apply equally well to a variety of other frameworks.
I've fielded several questions about setting up an autocompleter with Zend Framework and Dojo, and decided it was time to create a HOWTO on the subject, particularly as there are some nuances you need to pay attention to.
Just about every day, I have an idea for a blog post, and most days, by the end of the day, I just don't have the time or energy to actually write anything up. The inner writer in me screams, "no excuses!" while the aging adult in me whispers, "time for bed, dear."
So, to keep my hand in the game, here are a few things running through my head, or that I'm working on, or that I'll be doing soon.
For this particular release, we tried very hard to leverage the community. The majority of new features present in 1.7.0 are from community proposals, or were primarily driven by community contributors. For me, this represents a milestone: ZF is now at a stage where fewer and fewer core components are necessary, and the community is able to build off it and add extra value to the project.
I've been getting a lot of interest in my Pastebin demo app — partly by those wanting to play with Dojo+ZF, partly by those just interested in the application.
I'm constantly trying to improve the application. I've done one webinar and one UnCon session showcasing it, and will be presenting it at Dojo Develper Day in Boston this Monday as well as at php|works later this fall, and want to keep the materials up-to-date and freely available. To this end, I've created a Github repository so you can track the latest developments, as well as pull custom tarballs:
All patches and feedback are welcome!
I'm a bit late on my ZendCon'08 wrapup; the conference left me both exhausted and with a backlog of email and work that has consumed me since it ended. However, this, too, is good, as it has given me time to reflect… and to finally get my slides up on SlideShare.
ZendCon was alternately exhausting, rewarding, educational, fruitful, infurating, and ultimately wonderful. I've been to every single ZendCon so far — I started at Zend a scant month before the inaugural event — and have spoken at each. My first time speaking was a fluke; David Sklar had just started at Ning and had to back out of his "Configuring PHP" tutorial session. Mike Naberezny and I were drafted to take it over, and we had N+1 attendees, where N was the number of speakers. Since that inauspicious beginning, I've gradually taken on more sessions and stuck around to participate in the conference more. I can honestly say that this was the biggest, busiest, and most community focussed ZendCon I can remember.
Now that Zend_Test has shipped, developers are of course asking, "How do I setup my test suite?" Fortunately, after some discussion with my colleagues and a little experimenting on my one, I can answer that now.
I have a number of updates and followups, and decided to post them in a single entry.
First off, you may now view my Dojo Webinar online (requires login and registration at zend.com). Attendance was phenomenal, and I've had some really good feedback. If you want to see it live, I'm giving the talk (with revisions!) at the ZendCon UnConference, at Dojo Developer Day Boston later this month, and at php|works in November. I hope to be able to show new functionality at each presentation.
Second, I've completed what I'm calling version 1.0.0 of the pastebin application I demo'd in the webinar. The PHP code is fully unit tested (though I haven't yet delved into using DOH! to test the JS), and incorporates a number of best practices and tips that Pete Higgins from Dojo was kind enough to provide to me. When using a custom build (and I provide a profile for building one), it simply flies.
The pastebin application showcases a number of features besides Dojo:
Zend_Test_PHPUnit was used to test the application, and
FireBug logger and DB profiler are used to provide profiling and debug
Finally, ZendCon is next week! I'll be around, but already have a packed schedule (1 tutorial, 2 regular sessions, an UnCon session, a meet-the-developers session… and that's just what I know about!). I look forward to meeting ZF users and developers, though, so feel free to grab me and introduce yourself.