People insult Cobol, but is the state of play going to be any better in 40 years

[gleepower]

Cobol is often insulted for being "legacy" code that should be replaced with "modern" applications. The thinking is that these banks are using 40 year old codebases running on ancient IBM mainframes that have so many embedded assumptions and dependencies, most of which the people who created them are dead or retired, are becoming increasingly costly to maintain.

But what was once modern will one day become legacy, and I fear the hundreds of languages, technologies and frameworks will make this problem much worse! Cobol at the end of the day is quite simple. But what about legacy code using haskell, clojure, Java tomcat and several frontend frameworks, back end frameworks and distributed microservices, of course written in several languages each. In forty years, what hope will we have of understanding these things? Google has more than one billion lines of code, and it's only a few decades old.

Does anyone else share the same worry?

[Jason Knight]

A lot of "legacy" languages get static from younger know-nothings, particularly those full fanboy gushing over the latest and greatest hotness.

See how JavaScript goes from joke language, to outdated language, to hot and trendy. Or how Ruby was going to destroy all other languages. Or how "PHP is dead".

Oh, you'll hear that last one all the blasted time. As much as I dislike the steaming pile of ineptitude, Wordpress is one of the big bears of the Internet, and I don't hear them running and screaming "Oh we have to get off of using PHP!"

But you talk to node.js fanboys and they talk about PHP like it was Cobol.

I deal with that for my favoring Wirth family languages. The idea that Pascal is dead and nobody uses it for production software anymore.

Except for Skype -- which only just moved away from Pascal and the new version sucks as a direct result of it... but how about Winrar? Winzip? CloneDVD?  InstallShield? Nero? UltraISO? FL Studio Pro? Panda AV? Spybot Search and Destroy? Aida 64? IcoFX? Avant? Dolphin?

ALL are either Delphi or Lazarus/FPC, and that means Pascal. But tell me again how "no viable commercial software could ever be written with them"

But yeah, in banking Cobol is alive and well and not just because "it's dealing with legacy data". There are certain types of financial operations it's just easier to use Cobol than something C releated. Cobol often has cleaner and easier to understand code than C syntax's painfully and aggravatingly cryptic nature. More so for many of the "latest hotness" languages where security, efficiency, and ease of use are basically told to go **** themselves.

Yet because nubes don't use it, or "names" in the industry are looking for a clickbait headline, classic legacy languages -- Pascal, Ada, Modula, CoBol, DiBol, Fortran, Perl, etc -- as well as many far newer but still over a decade old languages -- like PHP -- are getting dumped on with lies by people who desperately want their new language to be the ONLY language, even if the newer language is often not crafted to be a "one size fits all" solution or is utterly mismatched to the task being performed.

[gleepower]

Some good points. ADA seems very interesting, but I don't hear much about it. Mind you "how much you hear about it" seems to be a distorted signal - you rarely hear much about Cobol, yet it's probably the most used programming language today (although it's hard to know for sure whether this is the case).

Except for Skype -- which only just moved away from Pascal and the new version sucks as a direct result of it... but how about Winrar? Winzip? CloneDVD?  InstallShield? Nero? UltraISO? FL Studio Pro? Panda AV? Spybot Search and Destroy? Aida 64? IcoFX? Avant? Dolphin?

As for C, it's syntax definitely has some warts, but it's unchanging nature and it's universality and robustness of its tooling has the great advantage that programs written in it will also have people who know how to maintain them.  Unless a language really is miles ahead, I feel the cost of fragmenting the ecosystem is worse than the cost of using a slightly outdated language. And I don't feel many languages are that different then others (maybe Haskell is since it bars fundamental concepts, but is PHP really that much better than Perl or Java or C# and vice versa?)

Perhaps though all the languages will stay in the current human knowledge though, who knows. And even if they don't, they are all pretty similar? Maybe it's the complexity of the programs themselves we should be worried about.

<edited by ian inserted [quote]>

[Jason Knight]

Some good points. ADA seems very interesting, but I don't hear much about it. Mind you "how much you hear about it" seems to be a distorted signal

One has to remember there was a time simply admitting you knew Ada landed you on government watch lists because that's who used it nearly exclusively. It was created for mission critical applications where ANY form of error was intolerable.

... and why the government allowing C++ to be used instead being part of why "government security" from a tech standpoint is inferior to what the majority of off the shelf wireless routers provide out of the box.

Since C, C++, and their kin/kine seem to have been created from the very start to not only let programmers make obvious mistakes, but to encourage them to do so through a laissez-faire attitude and pointlessly/needlessly cryptic syntax, naming conventions, etc.

Choices that (might, maybe, almost) made sense in an age where people were typing in and working with code on 150 to 300 baud dialup over dumb terminals, but has no place any time after -- or even during -- the microcomputer revolution of the late '70's and early '80's.

Part of why I can't believe C is still going strong, much less so blasted "influential" in the design of 90%+ of all languages since.

But what do I know? So far as I'm concerned JavaScript and PHP are just C in a frilly dress and CFM pumps, tarted up with theatrical makeup, and tucking back Mr. Winkie for a night on the town.