So if I cheat – just a little bit by rounding predictions to the nearest 1% – and go with the “reversion to the mean” approach to state swings (see the discussion in the previous post), my model doesn’t actually look too bad:
Previously I was extremely dismissive of UML and back when I was doing the masters degree at Birkbeck I really loathed it.
But now I have come to end of my first working year as a software engineer I can really see why people use it on complex, collaborative projects.
PlantUML in particular has been a revelation.
