The future of Free Pascal

[lazie]

“Only the best live twice” 

In earlier times the big success of Pascal was concentrated in the company Borland.
- Thas was a problem:

Borland was renaming the company many times (some lost the connection to Pascal within the 90s)... - Some decisions of Borland did not be good for Pascal/Delphi. The Company was loosing developers to Microsoft (big money offers...)....

One could get other languages/compilers for free. - Turbo Pascal/Delphi was only a product for windows - so other languages which where costless and platform-independent could extremely easy win from that situation.

TODAY the situation of Free Pascal/Lazarus is much better for the future !
It is in hands of the community....
.. – it is not really important – that people (first trend-orientated) talk and talk to much theoretically about some details of the “Pascal language of yesterday” (before OOP).
There is ever a different point of view to talk about – til u USE it.

Lazarus is a big thing to build usefull applications for Linux... (- like they do it for Windows).

If there would have been Free Pascal/Lazarus in earlier times – I would have been with Pascal since the early 80s without interruption.

[jacmoe]

Coming from C, C++, PHP, Python, Lisp - even though I actually did start out in Pascal in the 80's - I am pleasantly surprised to return (after 20+ years) to a language that is modern, and in many ways ahead of the 'competition'.

I am also extremely happy to see that FreePascal and Lazarus are being worked on actively.
That makes me excited!

Object Pascal is not really that old:
PHP, JavaScript, Ruby and Java all went public in 1995, when Object Pascal did.
And Python is from 1993, AFAIK.

It was unfortunate that Borland screwed up, and that Microsoft screwed Borland by stealing their chief engineer and 30% of their staff.

What Lazarus/FPC has, that Delphi hasn't, is that it is community built and fully open source.

So, here's to long life! 

[eMko]

The problem is using it in the newer environments of web and clouds, etc. So I only have one problem, getting the Pascal compilers and IDE to develop run time code for the newer environments.

What's your problem exactly? I've been writing server-side web apps on clouds using FPC just fine. For example, you could try CodeAnywhere's VPS service (it offers free account with some limitations but you can use it without problem), install FPC on it, and write some web apps. Here's my account on CA, cak.lebah.web.id where I experiment with FPC and Swift.

Since FreePascal and C are both compiled into native code, there is no reason why their position in this environment should be that different from the technical point of view (if we won't think of number of programmers - staffing Pascal project (or any other which is not in a mainstream language) can be quite a nightmare for every corporate manager (trust me - last year we had a research project, we used F# for that - project was a success, but now it will have to be rewritten into C# because nobody here can understand functional programming)).

In my previous company (which I left 2 years ago) we had an old information system (early 1990s) with core written in C and PL/SQL (with tiny bits of C++) running on Unix with later addition of Linux support (mid 2000s). User interface was VB6 and VB.Net (Windows.Forms), web services in C# (Windows Communication Foundation) - that was my part. Systems like these are not unusual. Calling a native library from .Net is quite an easy thing (no matter whether it's C or Pascal if you stick to primitive types/pointers to structures with primitive types and think about memory in advance; C++ can be a little bit problematic), the same with messaging between Java/.Net/whatever on one system and a native code on another.

Would you write a complete web application (including a front-end) in C? Well, you definitely can via FastCGI. Does anybody do that? Of course! But 999 projects of 1000 are done in other languages. For most of the projects, the language or run-time inefficiency doesn't really matter that much. What matters more is either time-to-market ratio and cheap development with a ton of ready-made frameworks (PHP, Python, Ruby) or manageability of an over-grown over-complicated over-engineered system reflecting corporate business rules which nobody understands so much they can give you a "big picture" (Java). Compiled native code gives you some advantages, especially efficiency. The price is slow development and changes taking more time (compared to scripting languages) and worse maintainability (compared to Java and C#, projects in those languages and environments are quite manageable even if some programmers are "average at best"). This is sometimes a very good trade-off.

When comparing C++ and FreePascal, the latter gives you more readable code in a language with very few "dark corners and traps" and fast compilation times. Downside is smaller community and harder to find workforce. No pointy haired boss wants that, because those are no benefits for him. So there is very little reason why you couldn't use FP in your project even in cloud environment or on the web. However since the biggest benefit of doing that is comparative rapid development and a native code performance, i.e. things which are not a top priority in those environments, you can have quite a hard time finding a job for that.

[Thaddy]

Object Pascal is not really that old:
PHP, JavaScript, Ruby and Java all went public in 1995, when Object Pascal did.

Wrong. Object Pascal dates from 1986. https://en.wikipedia.org/wiki/Object_Pascal.
I had a first view of Delphi in December 1994. It was public indeed in Februari 1995. Maybe you confuse Delphi and Object Pascal.

[jacmoe]

Wrong. Object Pascal dates from 1986. https://en.wikipedia.org/wiki/Object_Pascal.

I said that it went public in 1995, which is correct.
I am not confusing Delphi and Object Pascal here because Delphi is indeed the foundation of modern Object Pascal.
With emphasis on 'modern'.

It's like C, if you treat it like you treat object pascal, you would include A and B, wouldn't you? :p

That is my definition of modern Object Pascal anyway.
I am aware that some form of object pascal went before it, but I don't count that as a public specification of the language.

[Thaddy]

Wrong. Object Pascal dates from 1986. https://en.wikipedia.org/wiki/Object_Pascal.

I said that it went public in 1995, which is correct.
I am not confusing Delphi and Object Pascal here because Delphi is indeed the foundation of modern Object Pascal.
With emphasis on 'modern'.

It's like C, if you treat it like you treat object pascal, you would include A and B, wouldn't you? :p

It is most definitely NOT correct. Object Pascal went public in 1986.
TP5.5 also pre-dates Delphi with a good many years. That's the stupidest answer I have seen for decades. <grumpy mode firmly on  >
TP5.5 is the foundation of modern Borland flavored  object pascal as we know it. The Delphi language extensions merely bolted on to that.
Object oriented Wirthian languages were also available way before 1995. Don't be silly. Look at the facts.

[jacmoe]

Don't be a fucking moron, Thaddy.

Modern Object Pascal := Delphi.

What went before that leads up to the language that we use now:

After 9 versions of Turbo and Borland Pascal compilers, which gradually extended the language into the Object Oriented Programming (OOP) realm,Borland released Delphi in 1995, turning Pascal into a visual programming language. Delphi extends the Pascal language in a number of ways, including many object-oriented extensions which are different from other flavors of Object Pascal, including those in the Borland Pascal with Objects compiler (the last incarnations of Turbo Pascal).

That's flavors, precursors, in my book.

And I really don't think that it's stupid.

[Artlav]

To add an outsider perspective to the discussion, i would have liked to see a clean and modular pascal ecosystem.

Lazarus the IDE and Lazarus the rapid GUI designer should be two separate things, since all the desktop/RAD stuff is a rarely used ballast by now.
It's nice to see the IDE part had come a long way (since http://forum.lazarus.freepascal.org/index.php/topic,29210.msg183997.html#msg183997 ), it's almost as good as Delphi 7 with productivity mods now, with just a few annoying quirks.
But looking at it's code reveals a cheetah dragging a dead elephant of all the forms, components, objects and so on.
It would have been nice to be able to separate the two.

On the compiler level, a cleanly defined interface and implementation part - being able to have a completely agnostic compilation part separate from an OS- and environment-dependent I/O and UI part.
That is, being able to include some units unto a program, and have an interface that you can feed a string and a buffer pointer to, and get a compiled code spewed out into the buffer.
This would enable embedded compiler, just-in-time compilation, high-performance scripting, CGI-like web stuff and so on.

On the language level, i would have liked to see all the object oriented stuff dropped. It was always painfully ugly in pascal, and practically useless.
The parts i would have kept are abstract types like string and operator definitions.
Since these are my personal preferences, i don't expect that to ever happen, but if the compiler was modular it could have been a matter of some compiler definitions to build a "lite" version and "object" version, once again giving an option to unharness the cheetah from the cartfull of dead elephants.

[jacmoe]

I am not sure what you are after, but FreePascal can be configured to use different modes:
http://www.freepascal.org/docs-html/user/userse33.html

Perhaps FPC mode ?

all language constructs except classes, interfaces and exceptions are supported. Objects are supported in this mode.

[Artlav]

I am not sure what you are after, but FreePascal can be configured to use different modes:

Not exactly.
I'm looking at the compiler and IDE as a piece of code, not as a binary to use, and easily 95% of that code implements features i don't need or won't ever use.
If it was more modular, removing the unused parts would have been practical.

[howardpc]

Seems to me you're looking for a (modular) DE (development environment), not an IDE (integrated DE), perhaps an editor with macros etc. for compilation and intellisense tailored to the language in use.
It is a feature of Lazarus' integration that nearly every part of the DE knows about (and hence to an extent depends on) the other parts, (else it would be a dis-integrated environment).
For example, Alt-Up takes you to the declaration of the class you are in, whether it is in the same unit, the same package, another part of the LCL, the FCL or a third-party package. That does not just depend on clever CodeTools as a modular plugin. It depends on fairly deep integration between the 'modules' in the IDE, and the whole 'project' and 'project group' concept.
That said, you can actually use CodeTools quite independently from the Lazarus IDE in your own projects, because it has been written in a modular way.

[Akira1364]

almost as good as Delphi 7

Uh, I think you mean "way, way better than Delphi 7 in every conceivable way." Seriously, have you tried to use Delphi 7 recently? It's not a good time. Open it and even just try dragging one of the dockable windows across the screen. Smooth, huh?  And don't forget about the great and useful hints and warnings provided by that version of the Delphi compiler! Or how it totally didn't need FastMM to be viable for any sort of moderately high-performance application! (Not, on both points.)

i would have liked to see all the object oriented stuff dropped

Ah, that explains everything then. I suppose in your perfect world we'd all still be using Turbo Pascal on DOS?

Seriously though, this kind of highlights what I think is a real barrier to the continued improvement/modernization of Object Pascal: there are simply far too many people who for whatever reason view ancient IDEs/Compilers through rose-tinted lenses. I've even seen people try to advocate for the use of Lazarus by talking about how it's "basically a free Delphi 7", as though that's somehow impressive, or something anyone would get excited about in this day and age. (It's not.) If you want to get people using Lazarus, then talk about the numerous features that make it unique, or more generally about the various high-quality libraries available for FPC that aren't available for Delphi, or so on and so forth. Don't compare it to a 15-year-old IDE that flat out sucks by all modern standards!

[Bostjan]

Pascal and then Delphi was my first programing language/tool and I am still nostalgic about it.
I am impressed with how good the FPC/Lazarus become and I am using it to maintain few older applications ported over from D2007 and D7.

But truth be told FCP/Lazarus is no match for modern day Visual studio, C# and .net, especially now that you have free community edition of VS and .net being open sourced.
If you absolutely must use language compiled to native machine code then I would say that FPC/Lazarus is still the best option for general purpose development since I dislike C++ and other compiled alternatives are to obscure to to find developers.
But now days the managed languages are good enough for almost any task. And for really low level stuff like drivers, kernel, low level libraries the C is still de fact standard.

Just my 2c, don't shoot the me    

[tr_escape]

But truth be told FCP/Lazarus is no match for modern day Visual studio, C# and .net, especially now that you have free community edition of VS and .net being open sourced.

Just my 2c, don't shoot the me

How can you describe the "Modern Day" ? I think you mean popular culture of desktop/backend(Server side) programming.

I am still using delphi and fpc for production lines , as scada solutions (pascal scada) , testing softwares in very hard conditions.

Ok we won't shoot you

[Thaddy]

Don't be a fucking moron, Thaddy.

Modern Object Pascal := Delphi.

What went before that leads up to the language that we use now:

And I really don't think that it's stupid.

Well. Commercial interests are interesting. I would rather see that you stick to the facts.
The fact is that Delphi just introduced some syntactic candy and RTTI. I still use Freevision daily and also use it for KOL.
Object Pascal dates from 1986, not 1995. That's silly.

The reason I use it (and many others do) is that öld school objects are stack based and more efficient than heap based classes. Classes are just another tool in the toolbox. The concept of Object Pascal dates from 1986. Period. And even TP5.5 dates only from 1998.(beta) and 1999 (release). I know. I worked for them.