In defence of Agile

Posted by Adam on April 27, 2011

Like several others, yesterday we read from a City Lawyer telling us that Agile is not suitable for Government.

Commentary on Twitter suggested that agile was a ‘fad’ and a gimmick. We wholly refute that, and wanted to explain why Government should be using Agile, and that it’s far from a gimmick.

Running Hack Days (as we do), and Agile Training, we’re more than aware that for a great many people, the idea of rapid development/planning, and even stand-up meetings (apart from those of the Privy Council) must be panegyric to them.

We’re used to this modern way of working being treated as something new, something untested, something untrusted. But PRINCE2 is hardly something to write home about. Do we want government contracts to only be available to a cartel (Q148) of ginormous outsourcing companies? Do we want to continue having seventy-seven weeks (Q88) as the average time for a government project to complete? Do we really want to hæmorage public funds unnecessarily? Do we want delivery on products that are no longer suitable, that have been ‘produced’ with such inept, out-of-date practices, that have been made obsolete by the time they’re eventually delivered?

The Guardian‘s coverage of Legal Practice Course enrollment figures is perhaps pertinent. We need lawyers who are able to wholly consider evidence, relevant cases, precedence/authorities, thence offering advice on the law, the standards, the practices/procedure and their counterarguments. Only then can only truly advise one’s clients.

Agile works. Three recent successes — hack days aside — can easily demonstrate this:

  1. Tom Loosemore & Pals’ AlphaGov
  2. The venerable
  3. BIS achievements

Responding to each of the critiques of Agile, from that post, then:

customers want to know up-front how much a system will cost.
A fair point. The simple answer to this is no one can. Look how often non-agile projects go over-budget, cost substantially more than they were estimated at, the on-going hidden costs, such as with PFI; at least with “we’ll pay you some money for something that does this” is clearer; there’s an objective to work to, and a contract can be formed around delivering a product/service that will do X; we’d like a solution to let people tell government departments X, Y, and Z information alpha, and information beta, now go away and build it this month” is quite clear. Isn’t it? I’d imagine with this there would be things like unit testing, consideration of people with apostrophes in their names, and even those funny accents some people have.

Departmental budgets are managed very tightly, and they must be approved.
By sub-dividing MEGAPROJECT into individual units, in a lot of cases, the cost for each can fall in under the 20thou’ sign-off ‘cap’. Sign-off can happen. There is scrutiny of public expenditure.

Agile implies that charges for time & materials should be open ended
No. Not open-ended. It’s delivery of a service; the only thing that has changed is the methodology. Most (if not all) of our developers would develop so they do not ship products that will need fixing later, and be an excuse to charge for fixing substandard products (we understand this to be more common with hardware). Many develop under open-source licences, making the code that makes the product work available for peer-review.

Government is also legally required to follow open procurement rules.
We’d like to draw the author’s attention to the Parliamentary Public Administration Select Committee, who have looked into this a bit; the UK would appear to be more cautious than our fellows in the Union: at the expense of small organizations; Rewired State are worried that outsourcing and procurement processes are not tailored (in the UK) for small organizations; who can afford spending 100k on working for a tender? Very few individuals, and a handful of SMEs.

Agile can’t give you a clear specification of outputs up-front … comparing different bidders on a like-for-like basis
By loosely defining the outcomes, and not the minutiæ, better products will come about; ideas not thought about emerge. The price for a typeface (“Nor can it give a definitive up-font price”) can easily be invoiced, mind. We suggest running microprojects as a commissioned hack-day competition; build this, this, and this, using these, these and these; to be used by Bob, Alice, and Carol, which does meet the criteria.

We challenge the complete ‘Unprotected’ scheme; as well as the notion that failure is a Bad Thing. The adage, “one learns from one’s mistakes” should be considered more in the UK — perhaps a symptom of the ‘compensation culture’ and, indeed Risk-adverse lawyers. Vint Cerf makes this point exceptionally well; look at the amount of people in Silicon Valley who’ve had at least one failure in the past; now look where they are.

As mentioned above, it’s not just the Public Accounts Committee who are concerned; but also the Public Administration Select Committee. Additionally, in many cases of agile work, the small nature of work could often mean that legal fees would cost substantially more than the work itself; if Her Majesty’s Government can spend millions on propaganda — for a scheme the public want — then surely, a few quid in Research and Development can equally be written off? Lessons, after all will have been learnt.

We’ll be responding to the critique around Agile and Decision-making in a further post.

From our experience in working with leading developers, we suggest that most of them:

  1. Don’t want dull planning meetings
  2. Don’t want to waste time
  3. Want to Just Fucking Do It
  4. Have quick turn-around
  5. Build lots of small things that do each task well, and then plumb them together (if appropriate).

We imagine that many don’t care, or even know about the more ‘traditional’ (ineffective) project management methodologies. We believe that most of those are anathamaic to modern cutting-edge developers.

PRINCE II, long projects, and the lack of interest/knowledge has wasted billions. Few citizens benefit. We’d imagine that the benefactors are accountants, auditors, lawyers, and shareholders; not even HM Treasury (these guys have good accountants who know about tax avoidance, after all).

The cabal has to be crushed. Procurement must truly be open. We need to encourage innovation. We need to terminate the culture of outsourcing. We need to keep the talent at home, contributing to our economy. Whilst the fate of Bletchley Park has been tragedious, the UK was the nation that cracked Enigma. That could not have been done with micromanagement, tickboxes, and project managers (admitedly, Chaps then had Common Sense and Proper Educations).

PRINCE II is moribund. Maranatha Agile