Jonathan Leffler wrote:
<snip>
actually, I'm not fussed which option you take
as long as what's there is readily explained (learned). But I would
head towards minimalism - I don't work with a language with binary
notation for numbers, and have seldom really missed it (I use hex or
occasionally octal - the latter mainly in relation to Unix permissions).
However, every feature has a cost. There's the implementation cost; the
documentation cost; the learning cost. The first two are borne
primarily by the developers - the latter affects all the users, though.
And you need to balance those costs against the amount of usage the
feature will get. And, frankly, I think that the usage will be minimal
outside the 4 bases already identified - binary, octal, decimal,
hexadecimal. So, unless the cost is similarly minimal, I would not add
the general feature.
It's a rather pragmatic view of things; in Agile Development, it is
known as YAGNI - you ain't gonna need it.
Actually, YAGNI is a very good point. And as some people have said,
completeness for its own sake isn't always justified.
So barring any other strong reason to keep them, I now intend to remove the
M;NNN (numeric) and M;'NNN' (blob) syntax entirely and just retain 5 specific
numeric syntaxes (N,0bN,0oN,0dN,0xN) and 3 specific blob syntaxes (Ob'N', Oo'N',
Ox'N').
Literals in any base can still be expressed in the "N/N" and "N*N^N"
formats/expressions where N is any of the prior 5 numeric formats.
That should keep things a bit simpler.
If anyone actually needs more they can instead make a case for addition and/or
define a user function.
-- Darren Duncan
_______________________________________________
muldis-db-users mailing list
muldis-db-users@mm.darrenduncan.net
http://mm.darrenduncan.net/mailman/listinfo/muldis-db-users