[ Date Index ][
Thread Index ]
[ <= Previous by date / thread ] [ Next by date / thread => ]
Steve Marvell wrote: > > Call me a hcker if you will, but there have been no good process > driven projects I've ever worked on. Hmm - In the sales role we saw all sorts of teams, and all sorts of processes, and only those with decent processes were going anywhere fast. Decent processes aren't necessarily heavyweight, which is where the military/government often fell down. Some sites had good developer involvement in decisions, realistic deadlines that they usually achieved without excessive midnight oil being burnt, everyone knew the process, and their own roles. They were few and far between, but they do exist, and the developers smiled more at those sites. We saw some really wacky abuses of SCM and process - the funniest was one military contractor who had fought for an opt out on their previous project from the project SCM tool, but obviously lost the political battle this time and had some hideous heavyweight process tool inflicted on them, which was about 3 years behind the current release of said process tools because the vendor had introduced some hideously complex upgrade procedure - no one was going to get much software written on that project anytime soon. As a large government project it also got hit by loads of other process related requirements - any methodology appropriate for as long as it is SSADM - I pity the poor development teams in that kind of environment. So there is good, and there is bad, and there is totally over the top. Why do I suspect you've seen too much of the third. I agree complex to configure tools are a problem, which is why I emphasised "lightweight" nature of the tools I had liked. The biggest problem was they were almost all Win32 only interfaces - which was a pain for us as we were mainly supporting cross platform stuff, so Java would have been a god send. Simon -- The Mailing List for the Devon & Cornwall LUG Mail majordomo@xxxxxxxxxxxx with "unsubscribe list" in the message body to unsubscribe.