2 ms·

Crunching against tight deadlines is not abnormal in big IT projects. Which is of course why big IT projects are such horrible things. The risks (some would sa

by generj 13y ago

Crunching against tight deadlines is not abnormal in big IT projects. Which is of course why big IT projects are such horrible things. The risks (some would say inevitable results) are well researched at this point. There is a reason the world has switched to Agile or Agile-like processes. There is no reason a project with a 3 year timeline didn't have a build and testing starting at year 1 and continuing periodically. There shouldn't have been any development by the start of September 2013. None, except fixing bugs found by testers.

GVIrish 13y ago
The problem is that if you have a terrible customer holding the purse strings, no development methodology would save you. Maybe if CGI used an Agile process to develop this, they could've at least gotten a good start even though the requirements were horribly late and were changing all the way up to launch. But we don't even know if they received enough requirements to start development before this year. We also don't know if the customer would've authorized them to start development before this year. Bad decisions and failings from the customer can doom projects like these very early in the process. And even if CGI had used Agile and gotten as much done as they could, if you didn't get all of the requirements in time to implement them, you still probably couldn't launch. Politically HHS probably would not agree to launch without certain functionality completed. In a sane project if you're getting big requirements changes late in the game you'd either push them to another release or you'd push back the release date. Neither happened and predictably, disaster ensued. Then there's the whole idiocy of allocating one week for full integration testing. Of course that test failed but again, no one sounded the alarm (at least not publicly).

by generj 13y ago

GVIrish 13y ago

The problem is that if you have a terrible customer holding the purse strings, no development methodology would save you. Maybe if CGI used an Agile process to develop this, they could've at least gotten a good start even though the requirements were horribly late and were changing all the way up to launch. But we don't even know if they received enough requirements to start development before this year. We also don't know if the customer would've authorized them to start development before this year. Bad decisions and failings from the customer can doom projects like these very early in the process. And even if CGI had used Agile and gotten as much done as they could, if you didn't get all of the requirements in time to implement them, you still probably couldn't launch. Politically HHS probably would not agree to launch without certain functionality completed. In a sane project if you're getting big requirements changes late in the game you'd either push them to another release or you'd push back the release date. Neither happened and predictably, disaster ensued. Then there's the whole idiocy of allocating one week for full integration testing. Of course that test failed but again, no one sounded the alarm (at least not publicly).