Turbo Pascal on CP/M, MSX-DOS and MS-DOS

I wrote Pascal for 8080 in 1976 and Finnish Army even paid some money for it.

I was truly disappointed that Turbo Pascal did not copy my handy PORT-metavariable. You can write & read IO-ports without any extra hässle:

https://photos.app.goo.gl/qeJpYb94XzLBCE139

You don't need to wait for someone to use Claude Code for this as it already exists and existed for many years now:

https://www.lazarus-ide.org/

Long before the term IDE was coined, there were source code editors from within which you could run the shell commands required to build a software project and you could examine their output, e.g. the error messages.

Moreover, interpreters that offered a complete interactive environment for editing, running and debugging programs, had existed for decades for languages like APL and LISP.

Nonetheless, the availability of Turbo Pascal on cheap computers has probably introduced much more people to such a programming environment than the older alternatives, which were available only on big and expensive computers, so at most they were accessible through time-sharing services in big companies or in universities.

I have used Turbo Pascal on 8080- and Z80-based computers running the CP/M operating system, before becoming able to use IBM PC compatibles, where the Microsoft and Borland C and C++ compilers were a better alternative, and it was a good progress over what I had used previously, but before that, in the university I had used a "Remote Job Entry" system which accessed by time-sharing a computer mainframe from a serial video terminal, and you could do with it everything that you could do with Turbo Pascal (for programs written in Fortran), except that its source code editor was far more awkward (it was a line-oriented text editor, where you could see the source code, but to edit a line it was extracted into a buffer at the bottom of the screen and merged back into the source code when editing it was complete).

On CP/M, Turbo Pascal was the most convenient programming environment, but for serious work, where run-time performance mattered, I was using the Microsoft Fortran compiler for CP/M, enhanced by replacing a few of the subroutines from its run-time library with optimized variants. Pascal is really much too weak for number-crunching problems, even if it may be acceptable for non-numeric problems that are not too complex (the original Pascal does not support multi-file programs, Turbo Pascal was extended in a non-standard way to support this).

Must be later improvement, because I whined about it in Usenet.

Originally you had to write inline assembly.

Interesting, given the prominence of Casio calculators in Portugal to this day.

However it was always BASIC in the models we had access to.