Forgbook Progress Report

As you can see, there has not been any Forgbook updates for quite some time (a month), despite my resolve to post some. This is for a few reasons, some more interesting than others.

The least interesting is that I’m back at university, which doesn’t leave me quite as much time to work on things, though I’ve been giving what time I can.

More interestingly, I’ve recently stumbled across Pinax, a very smart extension of the Django application framework that offers a bunch of links to reusable Django applications, allowing the user to create the generic skeleton of a web site in a single line, along the lines of pinax-admin setup_project mysite. This provides you with a standard project layout and a collection of full-featured applications for basic capabilities such as user account management, notifications and wikis. Succinctly enough, “Pinax takes care of the things that many sites have in common so you can focus on what makes your site different”.

So I’ve been working on developing a reusable application for activity management within the Pinax environment, and have actually made considerable progress. It’s currently in a usable state where activities can be created, edited, viewed, deleted, completed (various amounts), paused, cancelled, and scored, and these operations propagate through the activity hierarchy. Relatively more advanced features such as categories, tagging, friendships, messaging, commenting, and email notifications are not implemented, but the beautiful thing is that reusable Pinax apps for these tasks exist already, and it should just be a matter of plugging them in with a little configuration. More Forgbook-specific features such as privileges and JavaScript-driven views will take a bit of work, but really Forgbook can function quite happily without them.

Having made that progress, I turned to investigating where such an application could be deployed and hosted so I could start testing it as a legitimate user. The traditional option is just to host all the files on a regular Web server, and connect the Python framework via WSGI. This is not going to be compatible with my current host NearlyFreeSpeech however, because of the rather idiosyncratic nature of their shared hosting setup. Furthermore, a basic Pinax project by itself typically sits in a virtual environment and weighs in at about 50MB, making it unwieldy to just copy, paste and host.

There are a few other options, but I was attracted to Google App Engine, because it’s Google (<3), and it’s pretty much free until the application gets particularly popular. The trouble is that the database used at the backend is a non-relational database, which is a new-fangled way of saying it’s not compatible with every other database. As such, Django is very difficult to support, and Pinax is even more divergent. There are few attempts at workarounds (django-nonrel seemingly the most promising), but it’s otherwise developer hell even attempting to monkey-patch the systems together.

Those remaining options then are a little more pricey but seem as though they will take most of the pain out of Django hosting. There are a few start-ups (Gondor, ep.io and Djangy, all in private beta) specialising in Django hosting which I’m keeping an eye on, as well as a few hosts that are friendly to Django, such as WebFaction (detailed as supporting Pinax too).

Currently the plan is to wait until a couple more of these are available (hopefully sneak into a beta or two) and see what cheap options can be had. In the meantime, development will continue, though maybe not as rapidly with uni in full swing. There are few other interesting activities and possibilities in the air as well, but I’ll discuss those a little later.

In summary, Forgbook is alive and well and even usable, but can’t really come out and show off until it finds a home. Hopefully that won’t be too far away.

Where Is Forgbook Going?

Of my large (and still growing) handful of projects that I’d like to be working on right now, Forgbook is taking up most of my attention. Hopefully I can get some kind of super-early release out in a few weeks, but until then it’s worth noting a few shifts in focus and philosophy.

As much as I love the Web, I find it hard to submerge myself in Web development due to the chronic pains of browser incompatibility, and perhaps the need to be constantly responding to user experience. Once a product is released, I feel an obligation to keep it updated, which can at times conflict with a desire to put a project to the side for a while. It’s much safer in term of user response to work on a project privately until it’s complete and polished before releasing it rather than releasing pre-alpha kind of work and iterating consistently all the way through to a final release. I guess I value my flexibility pretty highly, because after all this work is effectively my free time.

Forgbook is (or will be) my largest, most ambitious project in functionality terms, and I believe it will be of the most value to myself and others, so I want to get it right. I’ve been thinking about what I want it to offer for some time now, and I’ll be updating the project page with more details as I solidify them. From a technology perspective however, the interesting decision I’ve made is that I’ll be switching from a relatively unambitious attempt to combine Flex and PHP to a Web-native Django application leveraging the Dojo JavaScript toolkit for the more flashy UI components. This means however that I have to familiarise myself with the environment Django offers; it’s immediately clear that it’s extremely powerful and scalable, but the decoupling that is its strength results in a variety of idiosyncratic idioms that will take some time to adjust to.

Why the switch? One of the requirements I’ve come to be very passionate about is making Forgbook accessible from any common device – PC, smartphone, tablet, perhaps even consoles – so that it feels more like an encouraging companion rather than a chore you must log into every day. This is best achieved using a native Web interface (HTML, CSS, JS) over something requiring a plugin (Flash, Silverlight), as it should at least be accessible by any device with a Web browser, in the scenario that I don’t/can’t customise an interface for that device. It would be suicide to jump into creating a sophisticated Web application without a framework, and my love for Python (and mere tolerance of PHP) combined with the popularity and support for Django make it a perfect choice. The integration of Dojo via Dojango is a big bonus too.

I’ll document the progress I’m making with Django in the next couple of weeks, and hopefully get a stripped-down-but-working version (basic user and task management) out in the wild before March. From there, I’m hoping the development efforts will feel more rewarding, and I’m sure there will be many more fun things to handle (feedback and reflection followed later by advertising and analytics). I’m looking forward to it as much as ever.

P.S. If anyone can think of a quality name for a sophisticated event/task management application (Forgbook is very much a working title), let me know (comment or contact). If I like it enough that I decide to use it, I’ll be happy to give you free access to pretty much anything I release forever. Hopefully one day that’s worth something significant.