On 2020-07-29 05:03, Richard Owlett wrote:
[A suggested] approach is literally orders of magnitude more than I want.
Consider these idealized cost functions for solution technologies A, B,
and C:
fA(t) = t*t + 1
fB(t) = (t/3)*(t/3) + 10
fC(t) = (t/10/*(t/10) + 100
Observe:
- fA has the lowest initial cost, but the highest cost over time.
- fB has an order of magnitude higher initial cost, but an order of
magnitude lower cost over time.
- fC has the highest initial cost and the lowest cost over time.
- fA and fB intersect at time tAB.
- fB and fC intersect at time tBC.
The obvious strategy is to estimate t and pick the technology with the
lowest cost:
- For t < tAB, pick technology A.
- For tAB < t < tBC, pick technology B.
- For tBC < t, pick technology C.
The debate is over which cost function applies to which technology, and
estimates of t.
I would put Excel and scripting language solutions into A. I would put
Access into B (unfortunately, there does not appear to be a FOSS B), I
would put .NET, Java, CORBA, LAMP, etc., into C. Using Perl, I estimate
your app should be t < tAB.
David
