The Future of FPC

[PascalDragon]

Strange. As there are directions clearly stated. Such as widechar-rtl, sub-targets, managed types, ....
What is all that, if not directions?

The closest I see to a list of detailed goals v3.2.4 is going to meet is found at:
https://wiki.freepascal.org/User_Changes_3.2.4

Honestly, if that is the list, I view that as rather inadequate.  It also seems to be an "after the fact" list, not  something that was planned beforehand but simply "this is what happened".

The point of that list is to highlight changes compared to 3.2.2 that affect compatibility. It's not a “this is what we plan to do” document. In fact there is no such document. That's because the project as a whole does not decide what is to come. It's decided by whatever specific devs work on - like MvC's preparations for the Unicode RTL.

[Chronos]

Marcov tells me that the last official released of the compiler (FPC 3.2.2) came out on 20th May, 2021 (uploaded to sourceforge 9th July, 2021) . that was over FOUR and a half YEARS ago. while there may have been the most earnest of "directions" in place since then, there has been no UPDATED package for FPC released to sourceforge over these past 4.5 years. and let us be realistic - the FPC/Lazarus homepage points to sourceforge as the place to get FPC from, so sourceforge will be where most users go to get their FPC install from.

Imagine alternative universe where FPC team releses a new version lets say once a year but with just small updates. Then the argument that time since the last released version is too long would be erased. So it is not just about cadence of new releases but also about support for current moder technology. FPC is a tool and if it doesn't solve your problem then it is useles for you.

It needs not just manpower but also goals and leadership. Successful open source projects are either backed by big companies or they have a good leader. Sometimes even both.

[440bx]

The point of that list is to highlight changes compared to 3.2.2 that affect compatibility. It's not a “this is what we plan to do” document. In fact there is no such document. That's because the project as a whole does not decide what is to come. It's decided by whatever specific devs work on - like MvC's preparations for the Unicode RTL.

IMO, that's a big problem.

What project is managed by "let's see what this group of people do... some parts of it will become the next product" ?

This is why I say there is no direction to this compiler... the "direction" is "whatever the devs feel like doing, if they have the time to do it and, the inclination to complete it."    This is a novel project management methodology I had not seen before (probably because it was known not to produce the "best" results in the long run.)

[LeP]

Only the people that read too many IT management rags.  Try to break down what those paper guarantees are really worth.
Basically I only look out to the coming next years up to maximally 5 years, assuming I need 1-2 years to rewrite, and ignore all the doomsaysers.

I don't read any management "paper" or things like this. I'm a tech not a real business man.
But to choose I should view the past / present and project the future. I cannot construct nothing without the present consistent and the past history.
And the future is only a thinking (with many and many solutions).

But for sure it's the past and the present. And the team service is a reality (like in the past), like is a reality that there is not this service in other solution (like in Lazarus).
And the same I'll apply with other environments.

This is not only for my "tranquillity" but also for my customers. When I present the tools and environments I work with, no one has ever complained.
In the past, when I presented tools and environments that weren't "guaranteed" by companies, typically even prestigious open source ones, there were always discussions about the "guarantees"....

In the past I had some issues with Delphi and I told one of my customer that the company is working on (with e-mail sent and received) and all will ended right ... with Lazarus (or FPC) what I should said: "wait, the community is working on"...

[LeP]

It needs not just manpower but also goals and leadership. Successful open source projects are either backed by big companies or they have a good leader. Sometimes even both.

Open source projects like FPC and LAZARUS should have funding plans, such as agreements with groups (e.g., industry), donors, advertisers, research institutions, and universities.
But to achieve this, they would need to present a credible development plan.

This is my personal opinion. I believe the Lazarus/FPC project has become truly important, and without this, I see consistent and continuous growth as difficult.

[Joanna]

. With Lazarus and FPC, being open source projects without official, "due" support, an entrepreneur rightly wonders what the risks are in using them as tools for their work.

And I'm not talking about quality here, just opportunity.

Fpc/lazarus are a good starting point IF the entrepreneur is willing to have full time employees that are capable of fixing bugs and improving it.
It would be a bad idea to rely upon free Fpc support from volunteers because people are very good at permanently disappearing and the entrepreneur will be left with an unsupported project that doesn’t work right.

It would be an even bigger mistake to use non open source software in my opinion.

[creaothceann]

Commercial software is like a car dealership that sells various models. Open source projects are more like DIY shops that sell you the tools for smelting your own steel, and books on color theory for the paint job.
It reminds me of those humorous "what if software companies were aircraft manufacturers" lists. Under "OSS/Linux/Unix" it would say something like "passengers bring parts of the plane to the airport, assemble it on the runway, and argue about the design and how to operate the plane."

Adding features to a mythical future version of the compiler isn't my definition of direction.

You could perhaps get a list of things that the team is currently working on, but without any guarantee that it'll work out. It can work in practice, as shown by the Dolphin emulator project, but even they have issues known for years until they're suddenly fixed, sometimes by chance.

[Joanna]

At least with open source software and your own employees working on it you don’t have to worry about the software that your company relies upon suddenly being switched to an expensive subscription model  or your confidential company data being exfiltrated to who knows where…

Opensource code that is missing features is the starting point not the final result.

[kupferstecher]

Actually, I wanted to compile a list/inventory of arguments why FPC will have a long life, for many years to come,

Its the DNA of freepascal and Lazarus to support many targets. And you can actually see it, i386, x86, arm, arm64, riscv, riscv64 (and some more). For me this brings confidence in using pascal in two ways, one is there are already plan-Bs, e.g. in case a target gets dropped from outside (e.g. Windows switching platforms). And the other is that the fpc developers already showed, that they deal with new targets regularly.

[LeP]

Fpc/lazarus are a good starting point IF the entrepreneur is willing to have full time employees that are capable of fixing bugs and improving it.
It would be a bad idea to rely upon free Fpc support from volunteers because people are very good at permanently disappearing and the entrepreneur will be left with an unsupported project that doesn’t work right.
It would be an even bigger mistake to use non open source software in my opinion.

Of course, your opinion is valid.
But I don't think this applies in all cases, on the contrary...

My experience is completely different: I use software to build something else (usually something industrial).
And this something isn't a piece of software that you can change at will, but something that must work without any ifs or buts.
I don't think anyone is happy if their car has a bug, say, in the ABS system, or in the brakes, or even in the acceleration system.
When you build something that works thanks to software you develop, and that is supposed to make its users earn money... well, guarantees are never wasted.

And just to draw attention to the upcoming news at the end of the year: a European law will come into force (obviously in Europe) for all producers of software dedicated to "consumer" products that will regulate some important characteristics and obligations, and open source is included in this law in a seemingly positive way (the law exempts open source from some obligations), but I am more than convinced that this will actually damage it, as these obligations will fall on those who use open source without any possibility of recourse.

[Ten_Mile_Hike]

Code: Text  [Select][+][-]

1. I would love to contribute to Laz/FPC codebase development; unfortunately my coding
2. skills are only one step above a spider monkey.
3.  
4. If anyone desires hundreds of program crashes, Windows BSOD, Infinite loops, buffer
5. overruns, and error codes I will then gladly "help" with development  :(

[n7800]

Stop asking what the language will offer you in 10 years - look at what it can already do. These are more than just guarantees - they're proof. A huge number of different types of programs have already been written in it: a compiler, an IDE, text editors, file managers, players, browsers, games... There are already several forum topics listing the most popular programs.

Therefore, the language is quite full-featured. Everything else is just improvements and fixes, but they exist too.

Moreover, open-source software is more sustainable than proprietary software. If Delphi disappears, no one will be able to continue the project due to closed source code (especially officially).

Furthermore, FPC/Lazarus is free, while Delphi cannot be used without a fee (depending on the license, which is subject to change). Some entrepreneurs will probably consider financial guarantees much more important...

This discussion could go on and on, but I'd rather get back to the patches and contribute to the future of FPC ))

[LeP]

Stop asking what the language will offer you in 10 years ....

Since I think you are speaking about me, I don't mind about next 10 years of developing. I think about past and present to construct a future.

... look at what it can already do. These are more than just guarantees - they're proof. A huge number of different types of programs have already been written in it: a compiler, an IDE, text editors, file managers, players, browsers, games... There are already several forum topics listing the most popular programs.

And ... what means ? All these things are available since decades, is not a FPC, Lazarus or Delphi privelege.

Therefore, the language is quite full-featured. Everything else is just improvements and fixes, but they exist too.

I quote from the page:

Note that since svn trunk is by definition still under development, some of the features here may still change before they end up in a release version.

really anyone want to construct something professional with this ? Neither Linux wants to deploy "NOT RELEASE" distro version ...
One can use them for test or hobbist projects, but I don't think that can be used to make something to sell or part of money market.

Furthermore, FPC/Lazarus is free, while Delphi cannot be used without a fee (depending on the license, which is subject to change). Some entrepreneurs will probably consider financial guarantees much more important...

Only "some entrepreneurs"  ? May be there are more. This doesn't mean that the alternative is Delphi, there are a thousand other programming environments...
What's the problem if you pay a fee (maybe a bit high on Delphi) for an environment that you work with and that makes you earn money?

This discussion could go on and on, but I'd rather get back to the patches and contribute to the future of FPC ))

It's not just patches; there's more. Or do you think that by continuously patching, you'll significantly improve the future of FPC? And how many people, like you, are capable of patching all the topics, perhaps in the FPC compiler?
I've already shared some thoughts on how to improve the future of FPC.

[n7800]

In this and the previous comment, I'm responding to the topic in general, not to anyone specifically.

FPC is for writing programs, and yes, that's been possible for a long time. So what do users want from the "future" of the language? Even large companies that announce goals don't always achieve them, or they take forever to achieve them. These aren't "guarantees" you can rely on in business. However, you can rely on proven products with a long history and extensive documentation. Business needs stability rather than new features.

This thread discusses a certain "future", and I've included a wiki page for improvements to a future version, not for current use. However, this is mostly code that already exists in trunk and will likely be in the next version.

Unfortunately, paid products also offer neither a "bug-free" guarantee nor a guarantee that they will be fixed. Windows and GitHub are examples. Has anyone tried creating bug reports? Are they guaranteed to be fixed? Even simple bugs can persist for years. At the same time, bugs in FPC/Lazarus and GitLab itself are often fixed by contributors, as the code is open source. Open source software offers a higher chance of bug fixes. And what will fix bugs if not a patch?

I could often add a temporary patch immediately and continue developing the app. This way my application is more free from the tools used. Workarounds aren't always possible, and with closed source code, I'd be locked in until a fix was released in the next release.

As for "support", there are forums, mailing lists, a bug tracker, and many other websites/channels dedicated to FreePascal. Frankly, I almost never asked questions there, as the documentation, manuals, video tutorials, books, wiki, sample projects, and source code were enough for me to study and solve problems. Of course, such contributions are just as welcome as patches!

Therefore, as developers, we can tell our application users that the FreePascal community is huge and quite active. It has been around for about 30 years, and a large number of programs have been developed. And they're even better supported than some open-source projects.

[PascalDragon]

This is why I say there is no direction to this compiler... the "direction" is "whatever the devs feel like doing, if they have the time to do it and, the inclination to complete it."    This is a novel project management methodology I had not seen before (probably because it was known not to produce the "best" results in the long run.)

That is in fact how open source projects without financial backing(*) work. It's even how bigger open source systems with financial backing work, like the Linux kernel, FreeBSD, NetBSD, etc. There are no fixed goals, no plans what will come, it simply depends on what the people contribute.

(*) Yes, I'm aware that nowadays we also have the Foundation, but it's mainly sponsoring features that they believe are relevant which is not necessarily shared with the believes of us devs.

It needs not just manpower but also goals and leadership. Successful open source projects are either backed by big companies or they have a good leader. Sometimes even both.

Open source projects like FPC and LAZARUS should have funding plans, such as agreements with groups (e.g., industry), donors, advertisers, research institutions, and universities.
But to achieve this, they would need to present a credible development plan.

No. That's the point of it being an open source project that there is no need for such things. If there'd be such things as a “credible development plan” I'd be much less inclined to contribute.

At least with open source software and your own employees working on it you don’t have to worry about the software that your company relies upon suddenly being switched to an expensive subscription model  or your confidential company data being exfiltrated to who knows where…

Then you probably missed the one or other project where the company decided to move their Open Source software from Open Source to proprietary (though that did then result in a fork that continues as an open source project). I can't remember the name right now (because it was a project that doesn't affect me), but it was some time last year...