Mail received from the mentor today:
Hello Loïc and thanks for your feedback and proposal.
I agree with putting communication strategy in the agenda for July meeting.
I would however keep the item business case as I feel it would be a strong argument fuelling the communication towards the Forge providers.
The business case should be kept simple :
-
Hypothesis on the average numbers of bugs reports per lines of code per year (or month if more appropriate)
-
Hypothesis on the average code size expressed in number of lines
-
Hypothesis on the average number of programs maintained and in development per developer
-
Hypothesis on the average time currently spent without Fedeproxy to report bugs and follow up bugs solving up to trace back to what issue generated it.
-
Hypothesis on the average time which could be spent with Fedeproxy in place to have the same bug tracing back and forth
-
Hypothesis on the number of developers for each Forge
Calculation step A : hypothesis 1 x hypothesis 2 gives the number of bug reports per program
Calculation step B : That step A result x hypothesis 3 gives the number of bug reports per developer
Calculation step C : That step B result x hypothesis 4 gives the average time currently spent by a typical average developer per year without FedeProxy (if hypothesis period is month, then x 12 months)
Calculation step D : That step B result x hypothesis 5 gives the average time currently spent by a typical average developer per year with FedeProxy (same as above for periodicity)
Calculation step E : Step D result – step C result gives the average time saved by a typical average developer per year
Calculation step F : Step E x 50 % as only part of the developers may use FedeProxy (of course ideally we would there make hypothesis on ramp-up rate over several years, but let’s keep it simple at this stage).
Calculation step G : Step F x hypothesis 6 gives the time saved by the community of developers of each Forge
There is a bias as each developer may be part of several Forges and adding time saved of each Forge for the same developer could be refined, but let’s keep simple the calculation as this stage.
Note : if that ease the thought, hypothesis may result from ranges having each each its hypoithesis level, and each range being allocated a % share to rebuilt an global average
My 2 cents,