Freepascal

[marcelomaurinmartins@gmai]

I noticed that Free Pascal hasn't seen any changes in two years; has the project stalled? I see that Lazarus gets frequent updates. But is it possible to advance Lazarus without its Free Pascal foundation? 

"Latest News

Jan 4, 2024
The creator of the Pascal Language, Niklaus Wirth, passed away on January 1st. Free Pascal would not have existed without the work of Niklaus Wirth. We mourn a pioneer and a source of inspiration."

link: https://www.freepascal.org/

[marcelomaurinmartins@gmai]

After looking into it further, I saw that Free Pascal has migrated to Git. I apologize—I was worried because I love Lazarus.

[DavidL]

No, you had it right the first time.  From a release perspective, Free Pascal is about as stuck in the mud as it can be.  Bug fixes and (generally overwrought and completely unnecessary) new features appear on trunk and a couple of branches from time to time but as a whole the overall project has become such a mess that it may never see another actual release.  The language itself is no longer clean and simple Pascal as it's run wild over the years with OO creeping-death "features" and "my-favorite-thing-C/C++-does" clutter.  Wirth would be appalled.  If they live long enough, all languages atrophy and rot - this is no exception.  (Except hacktoy Python and trashcan language C++, both of which started out rotten in the first place.)  Good luck.

* DavidL dons firesuit and awaits the wrath of the Believers.

[jamie]

Yeah, I gad you added that last liner.

You know, it's not like the old feathers of the language have been disabled, they are still there, you can at any time while using whatever version you have installed to not use those extra features.

 I guess that would be like owning a car with air conditioning, you elect to roll down the windows instead, at least most cars I know of still allows you to do so. I said most, we aren't talking about those rich man ve-hic-cals with fancy flip up doors etc.

Jamie

[Curt Carpenter]

The view that you can simply ignore new language features if you want to is of course true.  But I think that view reflects a code writer's perspective over and above a code readers, and I've always believed that readability was a Pascal priority.

Out of curiosity, does anyone know how many innovations have been introduced to COBOL over its lifetime?  Or has it survived largely because it has been an extremely constant, stable platform for almost 70 years and is readable (maintainable) even now by the COBOL programmer of 1965?   

(I understand that there is now "Object COBOL"  which seems weird, like people are writing novels in New Latin.  COBOL++?)

[Warfley]

Both models for programming languages exist in very successful languages. C, Go or Java are extremely stable, new features are added sparsely and only after thorough consideration. C++ or Rust are the complete opposite, there new features are added regularly, sometimes those features turn out to be dead on arrival and are never really used.

But the main reason with FPC is less ideological and more pragmatic. There are lots of new features already done, but we simply had no new release for 5 years because creating the release takes so long.

The view that you can simply ignore new language features if you want to is of course true.  But I think that view reflects a code writer's perspective over and above a code readers, and I've always believed that readability was a Pascal priority.

FPC also has a different problem, size of the compiler. FPC already supports multiple different Pascal dialects with different features to be enabled and disabled. The code will only get more complex and the core development team is quite small.

TurboPascal, MacPascal, ObjFPC, Delphi, ISO Pascal, etc. can be at times very different from each other already. So there are already places in the code that are very complex due to this. Any new feature is technical debt for the future.

[H₂SO₄]

Out of curiosity, does anyone know how many innovations have been introduced to COBOL over its lifetime?  Or has it survived largely because it has been an extremely constant, stable platform for almost 70 years and is readable (maintainable) even now by the COBOL programmer of 1965?

COBOL has survived solely because of the enormous amount of legacy code still underpinning a shocking amount of essential digital infrastructure (think banks or the United States' tax authority). COBOL is maintainable only by "the COBOL programmer of 1965" since nobody wants to work with it/learn it outside of lucrative (or hideously expensive, depending on which side of the employment contract you're on) maintenance work.

The language itself is no longer clean and simple Pascal as it's run wild over the years with OO creeping-death "features" and "my-favorite-thing-C/C++-does" clutter.  Wirth would be appalled.

Most of Brian Kernighan's objections to Pascal (many of which are absolutely justified) have become irrelevant solely because of innovation in the language. If all you want is to avoid OOP look no further than  {$mode}/{$modeswitch}.

Here's a question for all the "believers" who quote the "prophet" time and time again: Why would you consider the original incarnation of Pascal the pinnacle of programming language design when even Wirth himself created not one, but three successor languages (Modula, Modula-2, Oberon) with the express goal of fixing Pascal's shortcomings?

[ALLIGATOR]

FPC also has a different problem, size of the compiler. FPC already supports multiple different Pascal dialects with different features to be enabled and disabled. The code will only get more complex and the core development team is quite small.

TurboPascal, MacPascal, ObjFPC, Delphi, ISO Pascal, etc. can be at times very different from each other already. So there are already places in the code that are very complex due to this. Any new feature is technical debt for the future.

💯%!

I wish there were some way to deprecate outdated platforms)
And keep one or two dialects that are still active, IMHO: Delphi & ObjFPC. It might also be possible to streamline some of the mode switches by making certain ones enabled by default

Do something similar to what’s done in the Linux kernel: if no one is willing to maintain a particular target, that target gets removed. Later on, if someone is interested, they can always update the target to the latest version

[Thaddy]

COBOL has survived solely because of the enormous amount of legacy code still underpinning a shocking amount of essential digital infrastructure (think banks or the United States' tax authority). COBOL is maintainable only by "the COBOL programmer of 1965" since nobody wants to work with it/learn it outside of lucrative (or hideously expensive, depending on which side of the employment contract you're on) maintenance work.

Yes, I can confirm that. COBOL seniors get £2200,- + a day in the City of London and around €1800,- + a day in the Netherlands.
I often get more, around  €300 an hour without even asking for short projects/changes. several a year even if I am a pensioner. COBOL programmers are often offered work uninvited.... The circle is small, you meet the same faces all over Europe and almost all over 60..
And noboby does fixed price. Note every year there is more phased out, less and less work.

[LeP]

Here's a question for all the "believers" who quote the "prophet" time and time again: Why would you consider the original incarnation of Pascal the pinnacle of programming language design when even Wirth himself created not one, but three successor languages (Modula, Modula-2, Oberon) with the express goal of fixing Pascal's shortcomings?

Programming languages ​​evolve, a bit like spoken languages.
We can always do better: evolution teaches us that... what was perfect yesterday is now "to be thrown away" because there's a better version...

But the essence, the form, and everything that Pascal brings together is unique. It may not be perfect, but it remains the most readable and easiest language to learn.

Someone might think a Ferrari is the best car in the world... let's try driving it across the fields and see if it is... or maybe an off-road vehicle is the best car in the world... let's try driving 600 km on the highway and then we'll see...

I don't think there's one language that's absolutely better than another, but I think Pascal embodies one of the best.

[gidesa]

COBOL has survived solely because of the enormous amount of legacy code still underpinning a shocking amount of essential digital infrastructure (think banks or the United States' tax authority). COBOL is maintainable only by "the COBOL programmer of 1965" since nobody wants to work with it/learn it outside of lucrative (or hideously expensive, depending on which side of the employment contract you're on) maintenance work.

Maintaining very old Cobol programs, stratified by generations of programmers often not so competent,  is one of the more boring (and sometimes disgusting) activity in software development.
But entirely rewriting that monolithic and unintelligible mastodons in modern languages is a very risky enterprise. Maybe IA will be the tool capable of accelerate Cobol migrations.

[marcov]

Both models for programming languages exist in very successful languages. C, Go or Java are extremely stable, new features are added sparsely and only after thorough consideration. C++ or Rust are the complete opposite, there new features are added regularly, sometimes those features turn out to be dead on arrival and are never really used.

But the main reason with FPC is less ideological and more pragmatic. There are lots of new features already done, but we simply had no new release for 5 years because creating the release takes so long.

Such reasoning has been debunked in various "release" discussions already several times, so I'm not going to repeat it.  Sufficient to say that a compiler is much more than just the parser.

[Thaddy]

Maintaining very old Cobol programs, stratified by generations of programmers often not so competent,  is one of the more boring (and sometimes disgusting) activity in software development.

Me and others like me experiences are different: we talk the same language and are highly proficient. It is the noobs that let us down, so the old guard can repair it.  I don't object.  It is merely bad management that caused that. Recurring cost "savings" by not migrating. The penalty we receive, but not incur is welcome.
There is nothing monolithic about COBOL. It is very structured so maintenance should be easy even for noobs.

[gidesa]

Me and others like me experiences are different: we talk the same language and are highly proficient. It is the noobs that let us down, so the old guard can repair it.  I don't object. It is merelely bad management that caused that.
There is nothing monolithic about COBOL. It is very structured.

I think that high pays compensate high boring, not high pleasure 
Anyway, some old standards, as Cobol 68, were published not before OOP, but before procedural programming! Writing monolithic is almost inevitable.

[Thaddy]

@gidesa
Trust me: it is not boring, it is a challenge. Just like any other assignment is. Wrong attitude.