A Modern Methodology

 “When I have learnt the meaning of a modern methodology”

(With apologies to Gilbert and Sullivan)

My experience of contracting in the financial world.

 

The names and details have been changed to protect the guilty, and confuse the innocent.

 

I was bit dubious about the contract in the first place. It was through my current agency, and for various reasons I didn’t really want to go through them again. But the contract was coming to an end, and the alternative was the DHSS, on the wrong side of the counter, so I said yes, tell me more, and tried to sound bright and enthusiastic (it’s astonishing how bright and enthusiastic the mention of money can make me).

“It’s where you worked before, at Lightstone Circular Finance House. It’s a major new project, and they want analysts and designers. They are adopting a new design methodology, and they will send all the team, including the contractors, on a week’s course in the new methodology. The course will be run by Grabit  Run, the wellknown business consultants. It’s a very good chance.”

It didn’t sound that good so far, Lightstone had been my first contract. They had decided to go for a new Client Control System, based on the all singing, all dancing operating system from Yankee Computers Inc. They couldn’t find anyone with Yankee skills, so instead they took on a bunch of experienced people with similar skills, and gave us all (contractors and permies) a set of manuals to read.The permies all complained about the lack of training, the contractors all took the manuals home to read over the weekend. After about 3 or 4 weeks the contractors were therefore experts (ie slightly less ignorant than anyone else in the immediate vicinity).

My main memories were of burocracy, pig-headedness, and a standards manual 3 volumes thick, which no-one ever seemed to read, which gave no help at all on anything useful, but which was magically relevent whenever there was a need to blame anyone for anything.

“What about the money?” Of course I didn’t actually put it like that, I do prefer a more genteel approach. “Are you in a position to say what sort of rate they will be offering?”

“Well, they do feel that as you did learn Yankee here, that will be reflected in the rate” It was as well! Still any money is better than none, and a free course is certainly a bonus. I did insist on a six month contract though. That turned out to be one of my better decisions.

 

The course was at a posh hotel halfway between home and the client. About 20 of us sat there for a week while several smart young men from Grabit and Run explained how their system was carefully designed to ensure that each stage of the project, analysis, design, coding, and testing, would be so carefully done that each stage would be clearly documented and fully specified, and right.

I remember that I did feel a little cynical at the time, but I decided that perhaps I was being a little too pessimistic in my view of human sinfullness. Then I found out a bit more about the background. The last few projects (well the last many actually) had all been wildly overbudget, overtime, and overstaffed. So the powers that be had called in a group of consultants, Grabit and Run no less, who did a full evaluation of the situation, and explained that what Lightstone really wanted was a modern methodology, like for instance theirs.

The Grabit structured approach didn’t seem too bad actually. The smart young men knew their stuff, and to be fair, they knew and admitted its limitations. A pity therefore that some of the Lightstone management on the course didn’t listen to them.

 

We went back to base and started design. First problem was that the system required hundreds (literally) of forms to be filled in. Each form started with a diagram showing the various data flows in the system. Then underneath, a set of little boxes to be filled in, and lastly a text description of the system.

The forms were available on the workstations, but the diagrams presented a major problem. The edict came down that they had to be done by hand, in ink. “We do not expect that mistakes will be made.” was the clearly stated implication.

The ink policy lasted about 3 or 4 weeks, then reverted slowly to pencil. The mounds of paper grew. My cynicism started to grow in sympathy.

The system seemed to me to be fairly straightforward. Applications from prospective clients came in one end, various checks were done on their suitability (depending exactly what they were applying for), and out of the far end came either new records for the Client file, or polite letters saying “don’t ring us, we’ll ring you”.

Somehow it didn’t seem so simple when we started to apply the methodology.I think that the first thing to worry me was that the diagrams seemed to make the system more difficult to understand, not easier. I had trouble following anything I’d done a few weeks ago, let alone anything someone else had done.

I tried to work out why, and came to the conclusion that we were going down to a ridiculously detailed level at far too early a stage. Management however disagreed, and they were in charge after all.

There was so much paper though. After six months work the final production came to an inch thick manual of complex, hand drawn diagrams, for a system which I could have described in words in about 5 minutes. I was the one who had the job of taking a pile of these round to a bunch of busy managers in Client Control, to ask them to give us their approval, or otherwise, by next Tuesday week.

I felt a fool. There was no way those men and women were going to find time to plough through all that. If they did, I do not believe it was the kind of clear unambiguous specification that they required (and deserved for the money it had cost). All that would happen, I was sure, was that they would accept it, and then complain six months later when the final system turned out to be completely useless. (Assuming that it was ready on time that is.)

I was nearly right, but not quite. What we did get was a large BASIC listing from one of the managers.

“Well actually, we got fed up waiting, when there were so many delays, so I wrote my own system at home on my own PC over a couple of weekends. We’re going to use that as a temporary measure, for the next six months until the proper project is ready. But this might help you to understand exactly what it is we want.”

I analysed the BASIC. It was crude, unstructured, and full of spaggetti code. It didn’t have any of the automatic interfaces which our system would have. It was code done by an amateur, and frankly it showed. But it did have two overwhelming advantages over our system. It worked, and it worked now.

I left soon afterwards. The last I heard it was a year late, and still going. I never did find out for how long they used the manager’s home written BASIC program, but I wonder what his opinion of us was.

I’m not saying that I don’t believe in methodologies. I’ve seen too many hideous examples of unplanned disasters. On the other hand I’ve never seen any super large project, implemented in one stage from scratch, which wasn’t pretty embarrassing when you looked at it closely.

I’m no expert on structured design, maybe the smart young men from Grabit and Run were right. After all they did warn us, “It’s a good technique, but use it intelligently, and with a bit of common sense. And remember any methodology has its limitations, and one of them is that it’s only as good as the people who use it.”

On the other hand perhaps my worst paranoia is true. Maybe all the emperors of structured methodologies really don’t have any clothes on after all.