June 21, 2026
Code, chaos, and comment carnage
My 1992 view of the problems of computer programming in 1992
Old coding rant resurfaces, and the comments instantly turn into a nerd civil war
TLDR: A revived 1992 essay argues that by then, the biggest problem in software wasn’t the tools but the fact that people still didn’t really know the best way to write programs. Commenters turned that into a lively fight over IBM history, old languages, and whether coding is still basically organized wizardry.
A dusty 1992 essay just crashed back onto the internet with a brutally simple message: making the tools better doesn’t magically make programming easy. The writer points to old IBM history and says that by the early 1990s, the real problem was no longer building better code translators, but figuring out how humans should actually tell computers what they want. In plain English: we got better hammers, but building the house was still chaos.
And wow, the comment section was not about to let that slide quietly. One camp basically nodded along at the line calling programming a “black art,” treating it like an eternal truth that still hits hard today. Another camp immediately hit the brakes and started fact-checking the history lesson, with one commenter flatly saying IBM’s in-house world in the 1970s was really ruled by COBOL, not FORTRAN. Translation: the post wasn’t just philosophical bait, it was also timeline-fight bait.
Then came the classic internet seasoning: nostalgia and jokes. One commenter swooped in to remind everyone that while IBM was grinding through business code, Alan Kay at Xerox PARC was busy inventing the future with Smalltalk — then landed the punchline, “Sadly, it didn’t sell paper.” Another tossed in a deep-cut flex about pre-textbook compiler wizardry using “heroic techniques,” which is the kind of phrase that makes old-school programmers sound like mythic warriors. So yes, the essay asked whether coding is still half magic, half confusion — and the replies answered with correction, nostalgia, and just enough smugness to keep it spicy.
Key Points
- •The article says IBM used FORTRAN for much of its in-house programming in the mid-1970s and relied on a compiler called FORTRAN G.
- •IBM is described as having invested millions of dollars and hundreds of programmer-years to build the FORTRAN H compiler.
- •The author states that compiler construction had become far easier over the following fifteen years, to the point that undergraduates could build passable compilers in one semester.
- •The article argues that compiler quality was no longer the main bottleneck in software development.
- •The conclusion is that programming problems were primarily about methods and language design, not the absence of a better FORTRAN compiler.