[
Lists Home |
Date Index |
Thread Index
]
Bloated and leads to problems of scale in low power
devices. Carrying an object-oriented context around
in every file is an academic solution to a non-problem.
The problem of context is viewpoint. These naming specs
assume maximum independence of local data. That is never true in real
systems. The environment of interoperation which sets
the energy budget is more important than the theoretical
lifecycle.
There is a tradeoff between naming conventions and namespace
conventions in the hierarchy of the document. GJXDM is an
excellent example of naming conventions become deranged. It
takes too long to sort out the intended semantic just to
find the int at the bottom of the deep well.
len
From: Bryan Rasmussen [mailto:bry@itnisk.com]
In our case we pretty much just refer to ISO 11179 as the origin of our
rules
and then enumerate the rules used which comes out to probably 20 pages. So
we
don't necessarily have bloat. The problem is dogmatic use of these rules
which
I've seen lead to some pretty unusable names. I think Jonathan Barwell had
some
sensible suggestions on those lines.
That said the thunk factor is obviously important in the context of how much
benefit one derives in relation to the number of pages in a spec. The
benefits
of reading and using the ISO specs are all pretty much soft benefits and I
doubt
they will very appealing to people for these reasons.
|
