I belong to a mailing list for a defunt open source chemical process simulator (Sim42). Members of the list are now showing interest in restarting the effort. It was originally written in python which cause some speed issues. Several of the list members (including me) suggested freepascal and lazarus. The gentleman spearheading the effort sent the following and I would like to know what is the best way to respond. I do not know what features define a "modern language" and would like to know what points to bring up.

"My personal objective is not just to put out a simulator, but a fast and efficient simulator. Furthermore, personally, I do not consider a program portable if it is written in a language which very few can understand. A modern language such as any of the .NET languages will meet the efficiency objective but portability remains an issue. While I do have the Visual Studio .NET and I am happy with it, I understand that not everybody has it and it is not cheap. I looked at the Lazarus project and (at least at a first glance) it is indeed very "Visual" and will likely do the job. It will however, limit us to Pascal which is not really a modern language. For those of you who are in favor of using Lazarus, can you assure the rest of us that Pascal has been modernized? "

Thanks for any help
Matt Henley

Mild Pascal rant:

Thank you for considering Pascal. Contrary to popular belief, Pascal is indeed a modern language. Though developed some time ago it has evolved the to include most of the extensions seen in 'modern' languages. It should be noted that modern is not synonymous with recently invented.

Pascal is fully OO. Many would say the implementation of OO is better and more complete than C++. Performance is on par with C and C++. Any differences usually related to automatic and predictable garbage collection present in Pascal. Accessing C libraries is also simple.

Coming from a long background in large scale application design and management, Pascal has other advantages. The syntax is somewhat simple than C and Java. As well, experience and studies have shown that the same skill and effort applied to a C and Pascal project usually results in about 50% less productions bugs in the Pascal code.

FPC specifically has more advantages. First is the active (and somewhat rabid:) development community. It is centered in Europe, where the 'language du jour' does not hold as much sway as it does in North America. Also, there is a large collection of libraries, tools and utilities available.

FPC handles many different flavours of the language. It is portable (as is Lazarus) over many operating systems and processor types. It is far more portable than Java and .Net (for different reasons)!

A thought-out design of your application can be as fast and portable as you wish.

Thank you for considering FPC and good luck with your project.

</RANT>

P Davidson
Corax Networks Inc.
http://CoraxNetworks.com

Most people out there probably think of Pascal as still being in the state it was in when Niklaus Wirth first designed it.

Object Pascal to date is fully OOP, and misses nothing that C#, C++ or Java has: Interfaces, Exceptions, Classes: you name it, Object Pascal has it.

Michael.

Matt Henley wrote

understand. A modern language such as any of the .NET languages will meet the efficiency objective but portability remains an issue. While

I believe the with .Net, you will have the same speed issues that you have with Python since it compiles to CLR and not native machine code. So he is mistaken in his initial premise. Java will have the same issue. Plus, when running a simulation, you don't want the program to suddenly decide it has to do garbage collection, thus slowing down. Any .Net or Java program will have these problems due to their very architecture.

What does he consider a "modern language". You can't begin to quell his anxiety or disperse his ignorance until he can give you a definition. Ask for specifics.

Is he aware that Lazarus is written in Free Pascal? I would show him Pixel... it is very impressive.

Mark

That's simply not true. C++ has multiple inheritance, templates, classes in shared libraries; all things which FPC does not have. FPC does have metaclasstypes (and virtual constructors) in the OO area, which C++ does not have.

I do consider FPC to be a modern language, FYI ;-).

Micha

"multiple inheritance" what for?
"templates" is a specific resource, but ok.
"classes in shared libraries", delphi does have it.

Solved by interfaces in a much cleaner way.

templates,

Agreed, but absolutely not essential. It just saves typing.

classes in shared libraries

So does FPC if you so desire ? The RTL can be compiled as a shared lib, and that includes Classes...

Michael.

> That's simply not true. C++ has multiple inheritance, Solved by interfaces in a much cleaner way.

That doesn't solve the same problem. MI is much more powerful, but also much more complex, and easily abused so that code maintainability goes *down* instead of up.

> templates, Agreed, but absolutely not essential. It just saves typing.

That's an opinion, not a fact ;-).

> classes in shared libraries So does FPC if you so desire ? The RTL can be compiled as a shared lib, and that includes Classes...

This is *very* recent stuff. Is it in 2.0.x yet? Does lazarus work properly with a shared LCL library ?

Micha

but not losing the focus, Pascal is a very modern language.

I'll stick to facts, and skip opinions, as you rightly point out...

This is not recent stuff, I did that back in 1998 already on Linux (the machine called tflily) . Admittedly, it was manual work using ppumove, but it worked perfectly.

What is new is that the compiler does all the work for you.

What is still missing, is Win32 support. A DLL is a different beast than a shared lib on linux, because it's usually self-contained, and because it can't export variables. Mainly, this is package stuff.

The new internal linker should make this possible...

Michael.

Thanks for all the replies, i will try to formulate an agrigate response.

Since when is this a language feature? It is a implementation feature.

Hi,

What about the "new internal linker" ?
Is it already available in the development tree?
What has changed? Any documents to read?

Thanks and best regards.
chrom

IMHO the fatal flaw in this reasoning is that this opinion simply regurgitates some IT management blurb, and doesn't really tailor a choice of language to your needs.

There are three different arguments that I would mention in your response:

1) While not nearly as bad as Python, there are potential performance issues in using managed languages. This is not just raw calculating speed, but also startup time, memory usage (not unimportant in scientific calculations with large datasets!).
Worse, doing something about it often means doing speed dependant calculations in a non managed language in a DLL. So you potentially force contributors to learn a new language, and later have to partially back out again.

2) The only somewhat jusitifyable choice for "modern" programming languages in the IT sector is hiring. One can debate if .NET and Java are new generations, or just a glorified old hat, but the main point is that they _are_ prolific.

However that is not a 100% simple situation:
- First availability must be seend relative to demand (C# programmers are the only programmer on the US top 10 most wanted list, J2EE has been so in recent years). A lot more supply, but also a lot more hiring.
- Also, these languages are mainly business (read DB apps) oriented, and much less scientifically. Pascal has been a scientific language for years.
- Are you going to be hiring anyway? Otherwise I would inventorise first which suitable language is most common in your community and choose that (Pascal, Java, C# or not). It would be stupid to e.g. offend your most worthwhile potential contributors with a wrong language choice.

IOW, don't be fooled by a simplistic mantrum, but do the research what language is most suitable, and what's available in your community.

(your actual question is pretty much unanswerable till you define "modern")

Yet another lost soul wanders out of the C-Language (C, C++, C#, Java, Python, et. al.) forrest and into the fertile fields of Object Pascal.

You know, in all my years of using Pascal (starting with Turbo 3.01A in
1986) and Object Pascal (starting with Turbo 5.5 in 1989), I have heard this story on more than one occasion. However, I have never heard a story describing the opposite, ie. a person going from Pascal to a C-Language and having the same epiphany. I've tried to make the transition myself on several occasions, but I always end up coming back to Pascal.

Congratulations!

Mark

Please note that I am advocating Lazarus for the project.. I was responding to the project leader's (defacto at this point) call for pros and cons of each language. I and one other gentleman suggested FPC/Lazarus. I posted here precisely because I do not know what constitutes a "modern" language. I am not a programmer, just a chemical engineer who has done a lot of programming (mostly in Fortran), but who switched to delphi -> kylix -> Lazarus for my own projects.

This is a community project... no companies are currently involved. The project is a continuation of an existing codebase which was written in python. Few of the people want to continue it in Python. Several want to use C#. Two of us know and love Lazarus.

The application would involve calculations, database access, and a graphical user interface (visio like diagramming interface)... for reference you can look at
http://www.aspentech.com/industry_solutions/oilgas/product.cfm?IndustryID=23&ProductID=274 (scroll down for screen shot)

Thanks for all the input

Matt

Matt Henley wrote

Several want to use C#. Two of us know and love Lazarus.

On Open Source projects the Developer is King. Talking doesn´t matter much, what really matters is the code that is written.

If they don´t agree on making it a Lazarus software, ask them to have 2 versions. a C# for windows and a Lazarus one for cross-platform. Talk to them about targets that c# does not reach, like *BSD, Mac OS X, Sparc, etc =)

Then work very hard on the Lazarus version.

They will soon drop the c# version if favor of the Lazarus one =)

This is exactly like I expected. Still the end conclusion stands. Go with the language of the (group of) people you expect to contribute most.

Also be wary of people (in the community) suggesting languages they do not daily use.

Ah, Aspen. Yeah, used that during my studies, though not "oil and gas". If you need help with from FPC from time to time, just drop me a note. Would be fun to do something with chemical engineering again. (did it for 3 years before I switched to IT)

I also have some experience with numeric calculating in pascal, and inherited a pascal lib of relevant numerical routines from my Uni, so be sure to also run numerical math problems over this list. (it's in packages/extra/numlib, but unfortunately the docs are still untranslated in Dutch. I can pick out specific routines and translate/document them if you need them)

chromdildo wrote

Hi, What about the "new internal linker" ? Is it already available in the development tree?

It is available in the 'linker' branch in SVN.

What has changed? Any documents to read?

No documents that I know of.
The current status is that the compiler can now
work without any external tools: both assembler
and linker are internal in the compiler for at
least windows.

'make cycle' runs, cross-compilation and linking
works, the compiler produces a lazarus binary
which is functional AFAIK.

I have not seen any timings yet, but I have
seen reports that memory usage has been
greatly reduced.

Peter Vreman has done most, if not all, development on this. He can say more about it than I can.

Michael.

Yes, scripting language. Every bigger program I saw so far in python was basically a big hack. I don't consider Python as a general purpose language. The lazy typing makes it a maintainance nightmare. It's nice for scripting but not for a full grown project.

Why don't you just write the speed critical parts in C or C++. That's how everyone does it in python. There is no need to port to a language which needs 5-10 times the efford in getting things out.

Not wishing to make this OT, but...

Talk to them about targets that c# does not reach, like *BSD, Mac OS X, Sparc, etc =)

Um.. *cough* Portable.NET, *cough* Mono. Both will get you on to most of those platforms. Don't assume Microsoft needs all the answers. You can certainly compile under Windows and run under some of theose OS with little changes. Others might mean slight GUI tool kit issues. No more so that a Delhi to Kylix or Lazarus transition though.

However, I have never heard a story describing the opposite, ie. a person

going from

Pascal to a C-Language and having the same epiphany. I've tried to make

the transition

myself on several occasions, but I always end up coming back to Pascal.

This will bring a tear to your eye then... I have been a Delphi/Pascal programmer for 10 years. I recently began a transit process to C#, and .Net and I absolutely love it. It is a complete breath of fresh air. I can't think of a single reason to stay using Object Pascal, not one. And there are only small bits of the VCL I miss (DB access components.) I have used Delphi 2005, and Object Pascal is just funky and nasty when used with .Net.

M

Matt Emson wrote

Not wishing to make this OT, but...

I'd like to ask everyone interested in continuing this topic to take it to the fpc-other list.

Thanks,

Jonas
FPC mailing lists moderator.

Matt Emson wrote

This will bring a tear to your eye then... I have been a Delphi/Pascal programmer for 10 years. I recently began a transit process to C#, and .Net and I absolutely love it. It is a complete breath of fresh air.

Not so fresh, since it's the people that designed Delphi which designed C# and .NET, so actually you are doing more of the same.

:-)

Michael.

Not so fresh, since it's the people that designed Delphi which designed C# and .NET, so actually you are doing more of the same.

Nothing is fresh in IT. OP originated (so I understand, at least partially) with additions to Pascal for MacOS.

God only knows what will happen now that Borland is going through its changes. Without the Delphi jump point, making OP an important language will be a lot harder. Let's hope the new company doesn't go down the toilet ;-)

Wasn't it called Clascal for a while? And it might have been on the Lisa.

Matt Emson wrote

This will bring a tear to your eye then...

Oh no, please, I absolutely don´t want to know about your life.