Cake
  • Log In
  • Sign Up
    • 95% of ATM transactions pass through COBOL programs, 80% of in-person transactions rely on them, and over 40% of banks still use COBOL as the foundation of their systems.

      The governor of New Jersey put out the call on live TV that he is desperate for Cobol programmers: New Jersey's system for processing unemployment claims runs on COBOL and has fallen down under the load.

      What’s a “dead” computer language that’s still in use at your company or at one of your customers?

      Shout outs to @RogerVerhoeven @Factotum @jpop @mbravo @Chris @Ravi @ChrisJenkins @Linux @yaypie

    • I'm afraid most of my customers and/or jobs are pretty modern, sometimes on the leading edge of it. However, in the background I do see a certain resurgence of interest in dialects and rethought applications of LISP (think Clojure or, to a lesser extent, Scheme - both fascinating and useful programming languages). In fact, re-learning and leveling up Clojure is one of my personal goals this year.

      In scientific circles, I believe Fortran is still very much alive.

    • I don't have any good examples of dead languages, but Legacy systems abound. Trying to integrate a web platform with old server based versions of Sage, for instance, is a nightmare compared to your standard JSON style integration.

    • Nothing really ancient in use, but one of the basic data packets in our virtualised, dockerised, distributed cloud system is still called a 'modem' because it traces its origins to the first version of the system that was developed in Clipper, and said data packets were distributed to POS locations by dialing up the central location using 'plain old telephone line' modems.

    • In scientific circles, I believe Fortran is still very much alive.

      Yep...very much so. It was used as our primary data processing language for 30+ years. The "Fortran to C+" conversion was well underway when I retired a couple years ago.

      Government is highly resistant to change...business, perhaps a bit less so, because change is very expensive.

      Enough legacy knowledge was available to keep the "Bubble gum, Band-aids and Scotch Tape" viable for all of these decades on those legacy systems.

      It's painfully obvious now, but when politicians were told a decade or more ago "update the systems!", they kicked the can down the road. Thus the state, via the taxpayers, will now foot that bill for the neglect.

      My sympathy meter isn't registering at all on the issue.

      😄

    • I agree in principle, even though I disagree on some particulars :) as in a) I would never consider changing from Fortran to C++ for any number crunching (the former is so much better at it than the latter, in all aspects including gigantic amount of reusable code and libraries) and b) I believe it's never advantageous to skimp on sympathy :)

      Yes, tech debt is eventually debilitating. I haven't worked in a single place where tech debt wasn't piling up and I've been around the industry since late eighties. And in a bureaucratic environment, multiplied by mainframe (and similar) environments multiplied by, let's not ignore this, a rock-solid stability of these legacy systems, moving forward towards changing the whole ecosystem is not only mega expensive and likely perennially out of budget, but also very hard to find enough motivation to energize multiple layers of decision makers. So sympathy I have, and a way to magically solve these types of problems I do not. It takes some great leadership and an uncanny dose of luck to be able to push such changes through.

    • Also, can't help but share this bit of some exquisite geekery (read the poster's follow-ups too) - https://twitter.com/BrianRoemmele/status/1257832168455208963?s=19