Better than complete non-delivery... sure .devilsadvocat wrote:Wez seemed to think he'd do a better job out of his garage!
The primary reason these projects fails is because there are simply too many people involved. Just because something is a 'big system' doesn't mean you have to throw 500 people at it from day one.
I wasn't *quite* suggesting I would do a better job 'out of my garage', rather that I would be able to put a small team together and get something useful delivered. Last year, me and 1 other developer built a fairly complex website (think gumtree but more advanced), for a client. This went into production across the entire Indian subcontinent and worked just fine..... 2 developers! (and I was the only full time)
Pete is spot on when he says that most projects fail because the client can't make their mind up what they want. The trick is.... basically to ignore them. Get the general idea of what is needed and fucking build it. Trying to get every detail right up front kills your project... build *something* first and make alterations later. Software is mutable, which is why we don't 'engineer' software the way we engineer bridges.
I am sure that accenture know this..... but all the while the government are paying them by the hour to sit there and listen to the requirement change, they are probably quite happy to watch the project flounder around. It's all profit.