
ISO CD ballot for CD 14882, alias C++
Ben said:
"Undoubtedly their main problem was the controversial decision to use
the confusing term "class" for the OOP construct, versus the
more intuitive term "tagged" :-)"
Clearly the market-drive argument ("that tagged was preferable because of
greater familiarity with all the Ada 95 programmers out there") just wasn't
strong enough to carry the day :-)
Little piece of history for those who don't know the reference here.
Ada 95 has always used the term tagged, but during the design phase,
several people, notably Jean Ichbiah, suggested switching to the word
class, based on the familiarity argument.
There was actually a lot of sympathy for this idea, and when the DR's
voted on the general idea of replacing tagged by class, the ayes got
a clear majority (I forget the vote, it could be looked up, or
remembered by someone else). However, there was no such consensus as
to what the syntax might be.
The issue heated up, and came to a boil at the Salem meeting of
WG9 -- am I remembering the location right? if not, it wuld have
seemed an appropriate venue :-)
Both Tuck (adamantly opposed to class), and Jean (adamantly opposed to
tagged) gave extensive presentations, and then we voted, the vote
was initially 2-2 with most delegations abstaining (probably to be
interpreted as meaning that they thought it was not such an important
issue). The US delegation had in advance agred to join the winning
side, whichever it was, to help settle the matter, and since a tie,
especially with so few votes, was not enough to mandate a change,
the US added a negative vote, and the final vote was 2-3 with a
bunch of abstentions.
So we cam pretty close to writing
class a is ...
instead of
type a is tagged ...
but in retrospect, I think the abstainers probably had the right view :-)
I don't know if people want to rediscuss this issue -- you can be sure that
every possible argument stone has been unearthed alread on this particular
issue :-)