cross-posted from: https://lemmy.cringecollective.io/post/9716
Maybe some ALGOL 58 while we’re at it too.
Ada sitting at #22
EDIT: COBOL community adopting a dinosaur for the language’s logo
I still use Ada daily for my personal projects after having used it at work. I find it compliments my thinking patterns well. My only gripe with it is that they ate too much of their own dog food at AdaCore and now it can be hard to install Ada and gprbuild (due to a circular dependency). Plus gprc stole libgpr and broke some stuff too.
I wonder, if this is another effect of LLMs. Maybe it’s just really easy and lucrative to generate+sell books for these old languages.
Most of the COBOL material never made it into the internet. Like the actual instruction manuals for the languages. Also a vast majority that do have it on the internet have it under paywall. I notice that anything that is under paywall, the LLMs suddenly dont do as well. I think its because they only train them on the “open” internet.
> Thinking the TIOBE Index is worth anything beyond the 2000s.
COBOL is actually not that bad. It can work with SQL, it can have unit testing/integration testing. It can even go on the web (LOL).
But in all seriousness, the bad part about COBOL is lack of context. Most code that is in COBOL has not been touched in decades. And no one is willing to modify because of serous consequences (AKA job ruining errors) that can occur.
I worked with it in insurance and transportation. In both cases, the COBOL was actually pretty solid…but we didnt know WHY we were doing the operations.
But didn’t they have to retrofit structured programming into COBOL? As in if-else, loops etc. didn’t exist in COBOL originally, it was all just GOTO.
I guess, what I’m asking is: Does “not that bad” mean still pretty awful, but perhaps not as awful as one would expect for its age…?
If you want to laugh: http://www.coboloncogs.org/HOME.HTM
the old languages still have their fans – and COBOL, Fortran, Ada, and Lisp are still holding strong in their respective niches
Lisp variants like Clojure are being used for new projects (e.g. Logseq) but I’d be surprised to hear of anyone choosing COBOL for a greenfield project.
Yeah the only reason someone should learn COBOL is job security and potentially making a living moving things over. No reason to start a project in the lang. You can make flat files into ODBCs nowadays.
I suppose the ability to be left alone because everyone is afraid the COBOL person leaves and the company goes under is a good reason :)
I took a principles of programming languages course a while back and got to touch on a lot of these old languages. My professor had huge hard-on for Lisp. Don’t get me wrong. The simplicity of the language is admirable. But reading and parsing that shit gave me headaches. No me gusta.
I think I was the reverse, I found it easier wrapping my mind around Scheme than C …
Lisp has always been the future >.>
And tomorrow is always just a day away.