3D V-model
Earlier this year I enthusiastically accepted a keynote. In no time I had a working title, “From Computer Aided Engineering to Software Aided Engineering”, which reflects my 30+ years in the business well.
The V-model needed addressing and I have been struggling with (not really, but) it for years. The V-model starts with requirements elicitation, requirements break down, design and implementation. From implementation, different verification stages are followed with final validation methods. In that order.
However, I have never experienced a project that follows the V-model process as depicted in sources with a timeline under the V (just google V-model images, and you’ll see). On multiple occasions I have met engineers waiting for requirements and their breakdown that ‘never’ come when (re-) designing systems. The interesting part, the design engineers had already a good picture on what needed to be done without formal documentation.
In practice, engineers start to experiment, be it physical or virtual (simulation), with design alternatives. These design alternatives help to find formal requirements. All the V-model stages will be performed in the end and the found set of requirements will constitute the delivery contract once validated.
The original V-model has a timeline from requirements down to implementation and then further to validation, resulting in possible anti-clockwise iteration loops. My experience let me believe the timeline is orthogonal to the V-model methods (the number of methods may grow over time and the set of requirements evolve too) and the iteration loops are clockwise.
Any thoughts?
Neue Freiheit
9 个月I think, starting with top level requirements is realistic. Derived requirements are added, as solution (sub-)alternatives are chosen. So, a timeline for the V-Model is not necessary, because information of different types is added in parallel step by step at various places of the V-model
CATIA Systems R&D Technology Director at Dassault Systèmes
9 个月One aspect of the development model is reuse of earlier designs. Either as experience ("we know this usually works") or as partial designs and solutions ("let's hack the old code"). For better or worse.
Sharing insights on MBSE & System Simulation | Tech Cluster Lead | Lecturer
9 个月Edo, I agree with the statement (: The way I see it, there is no timeline because it is not “predictable” how many iterations you will need (or said in a French way: how mediocre your design methods are ??). However, what I tell the students is that the “horizontal axis” of the V depicts the progress of your development. And yes, sometimes the progress goes backwards when you find a requirement not being met. A couple of more points: 1. your rotating arrow goes - for my brain - in the wrong direction. I would expect it to go from left to right and I see it in the other direction. 2. In a 3D-V model it could be show also that there are many Vees are done (for each subsystem basically).