Re: Resistance to change (was free format question)

Let me put it another way. Consider FORTRAN. You HAVE all at least HEARD
of FORTRAN, whether or not you can read it or code it, right? Now
consider ANSI-66 FORTRAN IV, as compared with primitive FORTRAN on the
one hand, and FORTRAN-77. They're all distinct dialects, with some
mutual incompatibilities, but they're all recognizably FORTRAN (unless
they're coded for that dreadful CDC free-format FORTRAN "standard"!)

Or consider BASIC. It is itself a recognizable derivative of FORTRAN,
with greatly simplified syntax, and free formatting of source (but you
pay for it with mandatory line numbers, and no way for separately
compiled pieces of code to call each other.) Old Dartmouth BASIC, IBM
VS-BASIC, DEC BASIC-Plus, Altair BASIC, Radio Shack Level II BASIC,
Apple Integer BASIC, Applesoft, PET BASIC, and GW-BASIC are all distinct
dialects of BASIC, but they're all recognizably BASIC.

Now consider some things that happened some years ago at ETH, mostly by
the hands of Niklaus Wirth: He wrote a teaching language, designed to
force incorrigibles to write structured code. Indeed, it was so fussy
about structure that there were a number of constructs that were
perfectly valid from a structured programming standpoint, that couldn't
be implemented in this language without either using backdoors (like
using a DOWHILE or DOUNTIL loop to do the job of a FOR loop) or
violating structured programming rules. But then again, he never
intended Pascal as a production language, only as a teaching language.
When people started clamoring for a Pascal-like language for production
use, he came up with a language that added a lot of new features,
particularly aimed at code re-use. Did he call that new language Pascal
II? No; he called it Modula, and the most popular implementation of it
was his second try, called Modula II.

Consider an area outside of HLLs. Consider word processors. WordPerfect
was a marvelous word processor, and an industry standard for many years.
It was lean, and it ran like the wind even on fairly old, fairly small,
fairly slow hardware. But it wasn't a sophisticated typesetting system,
and didn't pretend to be: if you wanted that, you got one: PageFaker, if
you wanted slickness, or were just jumping on the bandwagon because
everybody else was on it (and if everybody else jumped off a cliff . . .
?), or Xerox Ventura Publisher, if you wanted the text to stay in the
original word processor file (WordPerfect, Wordstar, MS Word, or
whatever), and you wanted the finished document to look like people who
knew what they were doing had spent hundreds of hours setting it in real
type. Then along comes Microsloth, and decides that the very idea that
other companies are making money selling typesetting software offends
them. So they start building pseudo-typesetting capability into Word for
Windoze. What happens? People stop using separate typesetting
applications, and the standard of quality for computer-generated
documents (already rather low, thanks to PageFaker) takes a nosedive.

I see something similar happening in HLLs: the die-hards who seem to
actually believe that their language is truly the one to render all
others obsolete set out to remake their respective chosen languages into
something they're not. And in the process, the unique strengths of those
languages often get cast aside. And you end up with people claiming to
be RPG programmers, who don't know what "The Cycle" is, much less how to
make productive use of it in ways both conventional and unconventional.
More than likely, they also know very little about RPG's file I/O. They
end up writing programs that they may as well have been written in Old
Dartmouth BASIC, for all the advantage they take of RPG's strengths.

Then, too, how many people are pushing ILE, and yet ignoring ILE's
greatest strength: the ability to achieve static linkage between modules
written in different languages. You don't do that, and you don't even
think about using service programs where appropriate, and you may as
well go back to OPM.