launchpad-dev team mailing list archive

[Robert Collins <robertc@xxxxxxxxxxxxxxxxx>]

In general, I think spending significant effort to categorise things is
'Waste', unless a majority of the things so categorised are touched for
the last time as a direct result of the categorisation.

So for instance, if 2 man-months of tech debt work is permitted in the
next macro cycle, stop tagging bugs at about that value - more
categorisation beyond that isn't going to get more debt paid off, and is
using time that could be paying debt off.

Further to that, I think tech debt in particular is best paid going
forward - that is, say that you have two items of tech debt, Foo and
Bar, a project Quux that is on the immediate road map, and Bar makes
doing Quux harder then:

- doing neither Foo nor Bar retains the current tech debt (though you
pay for it by Quux being hard to do)
- doing Foo reduces some tech debt, but leaves Quux hard to do
- doing Bar would make Quux easier to do, and pay off tech debt.

So in this situation I'd do Bar, given that I had time to do do tech
debt. (Note that time(Bar + Quux) > time(Quux alone) for the purpose of
this illustration).

-Rob

Attachment: signature.asc
Description: This is a digitally signed message part