Hacker News new | past | comments | ask | show | jobs | submit login
IO Monad Realized in 1965 (2012) (okmij.org)
42 points by dgraunke on Sept 30, 2015 | hide | past | favorite | 8 comments

Everything old really is new again.

This reminds me of a paper by Gary Kildall published in 1970 called "APL\B5500: The Language And Its Implementation" [1], which I think will fascinate many by just how modern it reads. Just about all of the terminology and concepts were refined by then and remain identical to the present day, when the conventional understanding of the software industry -- where everything allegedly becomes obsoleted every 6 months, would dictate that it must be hopelessly anachronistic by now.

We're not as innovative as we think we are.

[1] https://www.cs.washington.edu/tr/1970/09/UW-CSE-70-09-04.PDF

It's rather odd that the idea that the software business changes rapidly is still the conventional wisdom. After all most of the really big ideas were created by people of my generation and the one before (I'll be sixty shortly and I wrote my first programs in Leasco time sharing Basic in 1969 at the age of fourteen). I still see and hear people talking about object orientation as something new yet it wasn't new even in the late 1980s when Turbo Pascal got objects, after all Simula was released in the mid 1960s and had pretty much all the basics of objects and classes.

I often get the idea that there's been sort of this slump of scale that occurred somewhere in the 80s where the pioneers of CS found themselves suddenly flooded with new people. The challenge moved from thinking of better ways to "do CS" to teaching what already existed.

Sometimes I think we're just now getting back around to where people were at in the 80s—albeit at much greater scale. Sometimes I think we're not there yet. Sometimes I think I'm just sort of thinking wistfully about some golden age of yore.

Forgot to say thanks for the link. I was immediately stuck by the technical report name format 70-09-04, very nearly an ISO-8601 date.

... and one that is only a few months before ISO Recommendation R 2014-1971 formalized such things. The "everything old" idea applies to ISO 8601 too. I myself was using these date formats, complete with the century number, several years before ISO 8601:1988. I had to touch nothing of what I had written to deal with the Century Bug.

Ten years after the turn of the century, I was still dealing with the interoperation between my approach and softwares that used 2-digit years: http://catless.ncl.ac.uk/Risks/25.94.html#subj11

it just shows that language design is part of IT that moves slowly, because they are good enough. Also, we do happen to have a few millenia of experience with languages, whereas most people get by quite alright without solving numerical puzzles of all kinds.

the ancient Greeks actually had a fully-developed LISP, there's really nothing new under the sun.

This reminds me of the tutorial named "You could have invented monads" [1]. So apparently, somebody did :)

[1] http://blog.sigfpe.com/2006/08/you-could-have-invented-monad...

Guidelines | FAQ | Support | API | Security | Lists | Bookmarklet | Legal | Apply to YC | Contact