8 May
2010
8 May
'10
4:59 a.m.
Fred writes:
In terms of being noticably more primitive, let me
offer a hypothesis of
an honest way that Dan could be mistaken: It is not at all uncommon for
universities to proudly use software, including compilers, that were
written by students there. Such student written materials may be quite a
bit cruder, and lacking many refinements. Did your first attempt at
writing a C compiler include floating point? It may not always be
feasable to differentiate between earlier software V modern early efforts
by talented programming students. After all, that seems to be how we
ended up with the UCSD Pascal and P-System, in which certain
characteristics are noticably more primitive than much of the commercial
software of the time, while certain other characteristics may be novel and
innovative.
Many languages of the 60's and 70's were horribly over-specified and
the "full language" in fact had a lot of features that were not desirable.
Some of these were developed at schools as local products that were
never released, and others did become products with a life outside the
school. Others were developed by minicomputer companies because the
"full mainframe language" was not practicable on the mini.
e.g. WATFIV was like FORTRAN but eliminated the separate complie/link/execute
phases and could collapse them. It extended the language in some directions
while collapsing it in others. Later on came RATFOR which went in a
different direction, adding structured elements onto FORTRAN.
e.g. BASIC was supposed to be a simplified version of FORTRAN or ALGOL, but
in fact it was simplified even further into things like PILOT.
e.g. COBOL is, even today, cantankerous but some concepts from it went
into the much streamlined DIBOL. (I hava a sweet spot for DIBOL.)
I think a very important concept, is that having more features in a language,
especially when those features were specified by a committee, results in
an overly complicated overly large cantankerous language. Too bad
most people still get sold on features :-(. It's like the DVD players
with 141 buttons on the remote - people buy it because it has more buttons
or functions, to the point where putting more buttons on becomes a marketing
function, not a usability function.
Tim.