[O] Multiple notions for what's a day

[Sebastien Vauban]

Hello,

For 6 weeks now, since the commit a00a7b2 of Toby, things have changed when
generating clock tables. For example, instead of getting this (for some sample
files I have):

    | File  | Headline       |  Time |
    |-------+----------------+-------|
    |       | ALL Total time | 30:19 |
    |-------+----------------+-------|
    | A.org | File time      |  2:59 |
    |-------+----------------+-------|
    | B.org | File time      |  2:30 |
    |-------+----------------+-------|
    | C.org | File time      | 24:50 |

    (edited for the sake of simplicity)

we do have, as of today:

    | File  | Headline       | Time    |
    |-------+----------------+---------|
    |       | ALL Total time | 1d 6:19 |
    |-------+----------------+---------|
    | A.org | File time      | 2:59    |
    |-------+----------------+---------|
    | B.org | File time      | 2:30    |
    |-------+----------------+---------|
    | C.org | File time      | 1d 0:50 |

That is, the presentation *default* has changed: it now displays big amounts
of hours in days (of 24 hours).

I don't welcome the change of the default as such -- as that could impact
computations made on those figures; OK, that's not such a big deal --, but I
clearly like the ability to generate such formats, much more readable.

However, there is now a problem of interpretation for the value of "days";
until now, a day was valued as "8 hours"... see `org-effort-durations':

    ╭────
    │ Its value is (("h" . 60)
    │  ("d" . 480) --> 8 hours a day
    │  ("w" . 2400) --> 40 hours a week
    │  ("m" . 9600)
    │  ("y" . 96000))
    ╰────

In the current situation, you'd express days of 8 hours for the Effort
property, but be presented with days of 24 hours for clocking display. Weird,
no?

I guess that the clock tables should use the values defined in
`org-effort-durations' for their display. Or, at least, any other solution
where a day has the same value when used for clock estimates and for clock
reports...

Just wanted to show a problem, and start a discussion on what should be
done...

Best regards,
  Seb

-- 
Sebastien Vauban