[fpc-pascal] Fwd: What to do to get new users

[Stefan V. Pantazi]

Hi all,
Interesting discussion that was bound to happen at some point.

As a long-time (+30 year) user of the language and many years of 
FreePascal and Lazarus, I did not mean to pile on too much, but it looks 
like I may. Sorry.

 From what I read so far, my views match well those of Nikolay and Luca. 
Maybe age has to do with it or that my projects are rarely GUI apps and 
mostly low-level API programming where I barely make use of generics and 
where try to avoid classes and managed data structures as much as 
possible. I value less the looks and new language features but more the 
low-level programming power, the stability of the compiler, build tools 
and the the accuracy of the documentation. I am very grateful that 
FreePascal and Lazarus do exist, as imperfect as they may be!

My higher level, less API-oriented, dynamic programming needs are 
currently met by the Python+VSCode combination as we gave up on NodeJS 
a while ago. This also matches the needs of many of my students that are 
doing data analytics, data visualization and occasionally simple GUI 
apps, employing rather cumbersome tools based on quirky frameworks such 
as Tk and maybe Wx. Here is where RAD tools are helpful and why some may 
take a liking to C#. But as much as I would like them to, many I know 
are less likely to program embedded devices or to build low-level or 
complex GUI applications that need to squeeze every ounce of 
performance. Yet to me personally the need of low-level programming 
remains relevant. This is why I would still prefer a stable fpc compiler 
with macros, build tools and easy to configure targets, similar in power 
to gcc but hopefully less cumbersome and with faster compile times. It 
would also form a solid enough foundation to allow many other things to 
be built with it.

Stefan


On 10/17/24 06:47, Luca Olivetti via fpc-pascal wrote:
> El 16/10/24 a les 5:00, Hairy Pixels via fpc-pascal ha escrit:
>> Even if there is a market for Lazarus type apps people in 2024 don’t 
>> want to use a massive legacy IDE and prefer better tools like VSCode.
> 
> 
> I'm biased because I've been programming in pascal for almost 40 years, 
> starting with line editors/command line compilers and linkers (*), using 
> clipper for the user interface (anybody remembers it? the syntax for 
> defining input fields was way better that doing it in basic, yes, plain 
> old ms dos basic), marvelling at what was possible with delphi (even if 
> at the time it had no autocompletion or other smarts) then switching to 
> lazarus 20 years ago.
> So I mostly do "old style" gui applications (which, btw, have to work 
> 24/7 unattended), but not only.
> I recently had to use other ides for other languages (personal projects 
> in C/C++), one of them VSCode and I find it a steaming pile of s*it, 
> bloated and its "smart" features barely work (it's probably me though). 
> Good luck trying to design a GUI with it (I think that even clipper was 
> way better for that).
> Same experience with eclipse and netbeans (I had to do some java work, 
> ouch), again, it's probably me, but good luck designing a GUI with those 
> (but at least the "smart" features work somewhat better).
> So give me a "massive legacy IDE" that actually works any day, please.
> I'm sure that this "massive legacy IDE" will allow me one day to design 
> web interfaces with the same ease that it now allows me to design an 
> "old style" gui application.
> 
> 
> (*) programming embedded systems, so no debugger, not even writeln.
> 
> Bye