Mythology

A friend has noted on his blog the following test from an email sent to him by an industry recruiter:

"You must have solid commercial java software development experience, with good knowledge of development methodologies, and experience of the whole development life cycle, including agile mythologies."

My first response, obviously, was to laugh at the typo, but then I wondered whether someone was having a sly dig at the whole 'methodologies' industry.

Once upon a time we were all (allegedly) using the 'waterfall method'. Then there were ' bottom-up' and ' top-down (or the mixed-up ' spiral') alternatives competing for our behaviour. Now it is all about Scrum, Pair programming and other forms of alleged ' Agile' development systems for creating software.

Personally, I've always been a fan of incremental improvements based on an initial prototype and still work that way, but this is definitely one of those choose your acronym cases!

Diaspora

Wikipedia would have us believe that a diaspora is "the movement or migration of a group of people, such as those sharing a national and/or ethnic identity, away from an established or ancestral homeland" and, if you take the 'established homeland' as meaning Facebook, then the proposed Diaspora open sourced, distributed social media network concept meets that definition.

Ten days ago the four student initiators of this concept — Raphael Sofaer, Maxwell Salzberg, Dan Grippi, and Ilya Zhitomirskiy — released their first parse of the product to the world in the form of a " pre-alpha developer release". Written in Ruby, and providing its own web server (Thin; a fork of Mongrel) it professed to be able to provide the basic services required. In this case user registration, log in, picture upload, and a limited sharing of this information between different installations of the software (originally called 'pods' but now, it seems, 'seeds').

And, although this was clearly marked a developer-only release, so that people could in true open source fashion get their eyes on the design to find issues and improve the offering, a number of people started running sites and inviting public users. Problem was it very quickly became very obvious that the young designers had gone for the idea of a 'quick win' something to show off rather than concentrating first on getting the design and methodology in place. Indeed, early on it was apparent that many if not most of the people looking at the code were spending their time in 'prettying-up' the visual aspects of this alpha release (colour schemes, language localisation) than whether it actually worked.

Thing was, it didn't.

Oh, it appeared to do great things from the off. One reason why I 'jumped on the bandwagon' and installed a developer copy on a new site allocated for the purpose and commenced working on how to get it away from the stand-alone'ness of Thin on to Apache so that it could be commercially hosted, not just plonked on a user's home computer. But within a few days I took my service offline (and have since deleted it entirely) because the basic methodologies just haven't been worked out. Having sat on IRC for days and following discussions there, on the email list, and on their wiki, it is all too clear that at the moment too much is too unclear. There is, basically, no requirements document.

There is also no security process document.

Those college students have publicly stated that they want to make something work first, and then 'retrofit' the security to it. But as most programmers and designers have learnt by hard experience, you don't get there from here. Security has to be integrated from the beginning if it is to be effective.

The choice of Ruby has also proven to be problematic. That it is a scripting language currently in vogue there is little doubt, but that also means that there are fewer coders in the open source world who are experience in using it. And fast learning doesn't mean good coding. A number of people have blogged about the gaping security voids in the, albeit 'alpha', release. Patrick McKenzie has probably written the most detailed comments on the problems.

The concept of an open source distributed social media platform is one I've long been interested, and the Diaspora project managed a great start by garnering over $200,000 in contributions to the four guys when they first announced their idea. But their pitch said "We are four talented young programmers from NYU’s Courant Institute trying to raise money so we can spend the summer building Diaspora; an open source personal web server that will put individuals in control of their data." so basically they were planning on paying themselves the money they raised. Indeed, they went on to say "ready to trade our internships and summer jobs for three months totally focused on building Diaspora. We want to write code all the time, everyday. Once we have made our first solid iteration we are going to release our code..."

That they funded themselves for the summer is to be congratulated. That they built a solid codebase for Diaspora — or even made an adequate start — is most clearly not. In my opinion they have squandered their opportunity by believing too much in themselves and wasted the mindshare that the Diaspora concept had built up.

And while that is something that most of us do in our twenties we quickly learn that we can't do everything (no matter how much we'd like to) but have to work with others more experienced than ourselves from the outset if we want to do well.