On Mon, Sep 4, 2017 at 1:29 PM, Paul Winalski <paul.winalski@gmail.com> wrote:

I was once a firm believer in the virtues of a record-oriented I/O
interface such as RMS, but after using UNIX and Linux for several
years I became a convert to stream I/O as the base primitive. One can
always implement record I/O on top of stream I/O if you want it, but
it's difficult to do it the other way round.

Amen - this was my point with Dave.

One thing, I'll grant him though from the VAX-11/C experiment which I was wrong at the time. In those days, I hated that he had ignored Dennis' 'register' keyword. As a programmer and thinking in terms of C being a better assembler, I liked being able to control the compiler. But in my years of eating lunch you and the other folks in the compiler group I admit, I admit you have won me over. You guys can do a better job than I can, particularly with random architectures. If code is to be portable, then let the compiler to the register assignment.

Dave was right about about that one, but maybe we shouldn't tell me him one of the old UNIX guys I said that.

Clem