Re: Big demo. Should we test?
A demo is an integral part of the development process. A functionality "proof of concept" demo should be a part of the initial planning phase of any 'waterfall' project, even if it's just dummied up for the dog and pony show.
not sure what the article meant by referring to 'that' as a 'a classic waterfall affair', when there's apparently NOTHING to demo progress along the way, for 12 months even!
In just about every project I've worked on, especially my own, there's a demo that must be done early on as a "proof of concept", and some occasional functionality tests to make sure integration is working properly. And in one specific case, a team of 3 (minus me) went off to do something that was initially called 'agile' but was later deemed to be 'cluster-@#$%', and spent a YEAR on something I'd dummied up and demo'd in about 3 weeks... [and my version was used in several subsequent potential customer demos to show what the company was working on].
The 3 man team version was a re-write of what I'd done, with my version as a 'guide' of sorts, and the specs kept changing. They _NEVER_ implemented the functionality in the 3-man version until after one member was lost from the layoff cycle and I was brought in to help finish the thing. That project was always chasing the mayfly of "features" without FIRST getting the core to work.
In any case, what was described in the article is NOT a 'waterfall' process. A 'waterfall' process would focus on the big stuff up front as part of the overall design spec. And, a _PROPERLY_ done waterfall process would have a "dog and pony show ready" demo of the features as a proof of concept. I'd actually use that to test the system along the way, from time to time, swapping in 'real features' for dummy ones as needed to test things. [this differs from 'test driven development' in that I'd just make sure it all fits and works, rather than doing 'unit tests' all of the time and wasting effort re-testing trivial things over and over]
Typically my bosses/managers/customers would want to see this kind of demo from time to time to make sure I was getting work done. It makes them happy to see something, to see "progress", and they usually gave feedback which helps me to make them happy. Then you can tell them 'yes' 'no' 'it will be expensive' or 'it will take too long' and discuss stuff without having to move the target a whole hell of a lot.
Anyway, if the project went on for a year without any kind of dummied-up demo to at LEAST keep the client from asking too many questions, that's not 'waterfall'. that's more like 'poorly managed'. And from what I understand, 'FRagile' projects are well known for 'poorly managed' more often than not.
I would say that 'Agile' should look like 'waterfall' the way I described it, with only occasional 'scrum' meetings and more frequent one-on-one's with the project manager. But just like your average cluster-blank isn't Agile, it isn't waterfall, either.