"Julian Scarfe" wrote:

But I think you forget where you came into this, Jukka. The thread
is entitled "units of measurement on altimeters".

It's part of the very idea of the SI system that a single unit is used
for each physical quantity, in a unified manner, not varying the system
by application, country, or phase of the moon. It is clear that the
system is not always optimal when judged from a narrow perspective of a
specialized field, but if we go that way, we'll end up with expressing
quantities in incompatible ways - there's _always_ at least some reason
to deviate from a system.

The pascal is a very small unit in many areas of everyday life,
technology, and science. This is handled, as usual in the SI system,
using a systematic set of multipliers that correspond to powers of
1000, so that the numeric values can be scaled to a reasonable range,
[0.1, 1000). In some situations it might be, at least due to historical
reasons, marginally more convenient to use 100 or 42 as a multiplier.
But that's not a good approach. (It is true that some additional
multipliers exist in the SI system. But this is due to historical
reasons and discouraged in many standards, and tends to create
confusion because prefixes like h or da are not widely known outside
some specific areas of application, like the hectare.)

The quantities
that need to be expressed are in the approximate range of 970 to
1040 hPa, with a precision of 1 hPa.

It's against the principles of the SI system to select units according
to the range and precision that you have in some special situation.
We don't invent new units every time we encounter a new situation.
That was the old way.

Quantities in the range 97 kPa to 104 kPa can easily be expressed to
any precision you need or the current technology permits. Surely people
who work with such things can be expected to be able to work with
numbers with a decimal part.

(If it becomes relevant to work with a precision of 50 Pa, would you
insist on inventing a unit that equals 50 Pa, so that you can keep
using integers only? What about 42 Pa?)

The hPa is the right unit for that job.

No, the hPa is not a unit in the SI system, any more than 100 Pa is.

--
Yucca, http://www.cs.tut.fi/~jkorpela/