New FPC IDE

[qdos]

Insulting? Go back and look at your posts. You are insulting and nitpicking from your second post out. You ruined the whole thread by fragmenting the message. For what? Attention?-- and you call me insulting, you are being a complete tool.

Yeah design dont matter, here from ubuntu developer:
http://developer.ubuntu.com/apps/

Design comes first

All Ubuntu apps share a dazzling design and superb functionality. From top to bottom, they look and behave alike, regardless of the implementation toolkit. A great first step in Ubuntu app development is visiting Design  and then specifically checking out its Apps section. This is where you learn the fundamental principles and GUI building blocks that make Ubuntu apps sizzle.

Did you get that? "the fundamental principles and GUI building". You should call the Ubuntu consortium straight away, inform them that good looking UI's are rubbish. Not to mention that you alone could cook something together in a day which is better than all of Microsoft and all of Ubuntu combined. Being such a source of inspiration and motivation as you are when you describe the work of entire teams as "bull". Yeah hardly insulting at all..(sigh).

Oh and i got the point about m$ buying up retail space for win95, no argument there, that is common knowledge. But m$ likewise won the hearts of its users with their modern, fresh and colorful UI -- so i just dont understand why you homed in on one tiny aspect, instead of focusing on the importance of a good UI: which was the topic. If you want to air another topic, start your own bloody thread. But this type of behavior is what trolls do i guess, the keep nitpicking on things until they ruin it.
So to sum up, the question of marked reach for Windows 95 can be dealt with like this: "Is it true m$ spent a lot of money on gui?" Yes. Done.

Im here to inform you about an upcoming ide. Thats it.

Sadly it seems to have been wasted thanks to sam707, but now people know about it.

[marcov]

To all: please keep discussions to the point. This is going nowhere.

[qdos]

To all: please keep discussions to the point. This is going nowhere.

I agree 1000% -- but i dont think i have more to say.
Always one guy who has to ruin it for the rest.. Sigh*

[airpas]

good luck , don't let people discourage you , "Design comes first" that's truly true

[sam707]

ruin what? u r cussing someone who disagree with u, every reader can see that and anyway  if that post is a 3rd party anouncement it has nothing to do on the "general" tab of the forum. someone else wrote this as answer too... So if I agree with most of the people on there and disagree with u then u must fight me ? lemme ends up with that = if you write software as good as u use forums, I dont want to know about your softs

[Mocte]

ruin what? u r cussing someone who disagree with u, every reader can see that and anyway  if that post is a 3rd party anouncement it has nothing to do on the "general" tab of the forum. someone else wrote this as answer too... So if I agree with most of the people on there and disagree with u then u must fight me ? lemme ends up with that = if you write software as good as u use forums, I dont want to know about your softs

Well I think you indeed are ruining the thread just count the number of posts you have in this thread trying to discourage qdos efforts,  looks like you make your point clear enough, you can let others to express their opinions I myself think is it a great idea and I don't care if I'll have to pay for the IDE maybe others can get inspiration from this and we can have something open source on the future, I'm not as talented as  qdos  so I appreciate his initiative and the best part is it won't stop as an idea but it will be something we could use.

Greetings to all.

[qdos]

I dont want to waste more time on this (and thank you for the support, those of you who gave it). There are far better things to do. But you should remember Sam707 --it is possible to disagree with someone, without undermining their effort or ridiculing their motivation. I might not agree with you, but why should that result in smug remarks in your face or undermining your ideas? (especially since my so-called stupid ideas have already been realized and proven). I would not do that against you, that would be rude and to some extent selfish of me.

Anyways, feel free to investigate my work any time you like. Here are two open-source projects:

ByteRage:
https://code.google.com/p/byterage/source/browse/trunk/brage.pas

PixelRage:
https://code.google.com/p/pixelrage/source/browse/trunk/pxlrage.pas

PixelRage-Lite for C#:
https://code.google.com/p/pixelrage/source/browse/trunk/pxlrage.cs

I have hundreds of posts about object-pascal on my blog:
jonlennartaasenden.wordpress.com

I also run the biggest Delphi group on Facebook, with 2000+ active members:
https://www.facebook.com/groups/137012246341854/

You will also find some of my products described here, but only a fraction:
http://jonlennartaasenden.wordpress.com/products/

As for my quality of product, I have done work for Bang & Olufsen in Denmark, Norwegian Hydro Oil; My code has been used by Nasa (Norway supply several key technologies), and i have roughly 20 years of high-profile jobs and projects behind me; including work for the Norwegian government. Take your pick. But what you wont find is low-quality code.

You can also download the RTL for Smart Mobile Studio, written from scratch by yours truly. As was the IDE and utilities.

Considering the amount of donors to LCL, perhaps you should pick your battles more carefully, instead of being completely counter-productive and off putting for the Lazarus/FPC community.

Now, the derailing began by me talking about UI's, and how I found many of the alternatives out there hard to work with - because, brilliant as the authors are, they tend to focus more on code than looks. There is nothing bad about this - and it boils down to preference and expectation. CodeTyphoon is a fantastic alternative, but coming from Delphi and C#, I find myself spending a lot of time looking for the simplest functions.

And I dont like the multi-window layout (perhaps it can be docked, but well -- I want a Delphi XE + mono IDE). No disrespect to any author of alternative IDE's, there is more than enough room for everyone to co-exist. In fact, that is the great thing about open-source, namely that it's a win-win situation for FPC/Lazarus.

As for competition -- there is none.
Lazarus is very good at what it does, and that is not going to change. Also, Lazarus has many programmers working on it at the same time, expanding different application types and technologies. Lazarus will continue to be the flag-ship of FPC, just like Delphi is the flagship of DCC. But there is room enough for both RemObjects and Smart Pascal.

My adventure into HTML5 and native apps will be very different. I have absolutely no wish or desire to copy-cat Lazarus. My interest is primarily in robotics, embedded systems, operative systems, javascript (which is more and more popular on HW controllers and micro-processing units) and mobile solutions.

The concept of a Generic IDE is simply to expand the object pascal community toolkit. We have the compilers, we have the script engines, the servers and the text-editor components -- but there is no open-source and generic IDE. And those few IDE's out there are usually "hobby" themed, written poorly or just put together using some component package (like TMS).

What I bring to the table is clean, object oriented, stable, tried and tested code which many will benefit from.

What people may like or dislike about that is up to them -- i make this first and foremost because it's a tool i need for my own products.

[Leledumbo]

Good work, qdos. Actually I'm looking for a lighter IDE that only deals with code, much like Lazarus editor (improved SynEdit) + CodeTools and compiler calling only, and is capable to compile a file directly (optional project is ok). Well, a little debugging facility won't hurt.

No need to reply those who make you down. This thread is about your product, so anything that doesn't relate to it doesn't need response

[qdos]

The IDE I enjoy the most is probably monoDevelop.
And you see exactly what I have been trying to propose - namely that monoDevelop has been used and included in other projects to great success due to it's open nature.

The guys behind "Unity 3D" used monoDevelop to allow for C# script in their platform. For those of us who play around with C# that is good news. It means that I can now design my game/reality in Unity-3D, and then fire up monoDevelop for Unity and code my  effects, movements and other details that requires a fine grain of attention.

And that is what i would like, namely to have a standard IDE that can be injected into any projects who needs it. Written in clear-cut, anti-spaghetti, object oriented code which is easy to use and understand.

Studing others

I also want to take some inspiration from python and ruby. These languages and their tools were relatively "unknown" to the masses, used primarily by technical people who needed to automate tasks. But suddenly these two languages had a huge "make over" and have become unbelievably popular. Mainstream even.

FPC needs that. It needs a make-over and a new approach to problem solving. A programming language is a way of thinking, and the VCL is to old for modern problems. We really need a simplified RTL which try to learn from the past 10 years of "delphi" rule.

Focus for the new RTL would then be:

 ** Always write platform independent code if possible
 ** Use proper exception handling, deal with errors dont just exit
 ** QTX is not backwards compatible: no TP code. Not even "sysutils".
 ** Everything is an object, or should be
 ** More standard classes, use the .net framework as an inspiration

[marcov]

FPC needs that. It needs a make-over and a new approach to problem solving. A programming language is a way of thinking, and the VCL is to old for modern problems. We really need a simplified RTL which try to learn from the past 10 years of "delphi" rule.

Focus for the new RTL would then be:

 ** Always write platform independent code if possible
 ** Use proper exception handling, deal with errors dont just exit
 ** QTX is not backwards compatible: no TP code. Not even "sysutils".
 ** Everything is an object, or should be
 ** More standard classes, use the .net framework as an inspiration

People wanting to totally re-setup FPC on new principles (usually the darling language/paradigm of the day) has happened before., but I think if you really want such radical changes, better start over from the ground up, in a new, legacy-free project. Write compiler and RTL from scratch,  like Remobjects did.  Design in the basis for the (scripting) features you need, decide if you want to make allocations movable (to improve GC potential) etc.

I guess I would try to piggy back on some major (Java,C#) VM, and then later try to slap something self-contained under it. Maybe even use the FPC  JVM backend. That is already more Delphi legacy free by necessity, and could use contributors.

Trying to change a massive project as FPC/Lazarus inside out into something radical different will only get you mired in details and endless discussions. (see the over 1000+ (not kidding!) mails about just some unicode RTL details). And even aside from the discussions it wouldn't be productive. At the very least you would have to work in a fork till you really can demonstrates that it works and you have an interested audience for it.

[qdos]

People wanting to totally re-setup FPC on new principles (usually the darling language/paradigm of the day) has happened before., but I think if you really want such radical changes, better start over from the ground up, in a new, legacy-free project. Write compiler and RTL from scratch,  like Remobjects did.  Design in the basis for the (scripting) features you need, decide if you want to make allocations movable (to improve GC potential) etc.

The thing is, the change wont be so radical as people might expect.
Python is still python even though it had it's libraries re-arranged and modernized.
Ruby is still ruby even though it got "rails" now.
Same with perl, it was pitched in a new theorem but behind the hype it's still the same old script.

There was another discussion on this forum, regarding "is pascal losing markedshare" (or something along those lines), and the fact is that people regard object pascal as "old and outdated". This is exactly the same jargon that was used about perl and python a few years back. People regarded it as "old stuff that old coders user on servers". Well, that has changed practically over night.

I visited the Raspberry PI forum a couple of months ago. When I proposed that some of the examples be re-written in FPC for better speed (they used python) I was practically kicked out. People were actually hostile against FPC and pascal! I could hardly believe it!

I am not proposing a new FPC or Lazarus. I am proposing that we re-assemble and to some extent, re-package what is already there.

For instance, how would the RTL look if we took all those old procedures and stuffed them in helper-classes instead?

So instead of writing:

Code: [Select]

Procedure test(value:String);
Begin
  if length(trim(value))>0 then
  //do something
end;
We would write:

Code: [Select]

Procedure test(value:String);
Begin
  if value.trim().length()>0 then
  // do something
end;
Moving length() and trim() from intrinsic into a helper:

Code: [Select]

TQTXStringHelper = Record helper for String
  function trim:String;
  function length:Integer;
  function startsWith(acompare:String):Boolean;
  function endsWith(acompare:String):Boolean;
  ..
  .. all string operations here
  ..
End;
The functionality will be exactly the same as "old" delphi/pascal -- but the way people interact with the codebase is modernized. Let's face it -- there are tens of millions of .net and java coders (young coders) -- which dialectic layout do you think will fall natural to them? The objectified, class helper driven approach? Or the old function calls?

The old function calls will still work, especially those that are intrinsic to the compiler, but the presentation of the RTL and language will be greatly simplified.

In Smart Mobile Studio we did exactly this. We also moved all helper functions to types into helpers -- which really turned out to be extremely powerful. Especially for manipulating structures like RGBA, Rect, Point and others. Building on point we continued to array of TPoint, which then turned into TPolygon and TPolygonF -- which each gained helper classes as well, dealing with rotation, scaling and much, much more.

We also added a helper for handles and handle management.

Old way:

Code: [Select]

if (handle<>INVALID_HANDLE) then
//do something
New way:

Code: [Select]

if handle.valid then
Begin
end;
The helpers truly allows you to write more readable and much more effective code. Especially when dealing with arrays of typed data. For instance, You could use the polygon functions on an in-memory array of TRect's, because a TRect can also be defined as two-TPoints.

As for discussion -- they can discuss as much as they want. I will do as I please and change whatever I desire. FPC is an opensource project. FPC wins no matter what due to the open-source nature and license.

As long as we get FPC and object pascal out there -- all is good. Because that means more potential users of object pascal, which in turn means jobs and projects for us.

[marcov]

IMHO that is a pointless exercise. Why? Because such anti pascal sentiment, even though presented differently is simply ancient anti-teacher, anti-establishment sentiment.

They had to use the language for school, school is stupid, so language is stupid. They rather use what Thorvalds used, since Thorvalds is cool. Simple as that, and no superficial dressing up of the pig will change that. 

[BigChimp]

IMHO

However, qdos indicated he will go ahead anyway:

As for discussion -- they can discuss as much as they want. I will do as I please and change whatever I desire.

... so I wonder what that IMHO is worth... Especially after the previous posts in this thread.

At least he seems like somebody who can pull it off which is a refreshing change from all the people who want other people to change the language etc.

Good luck, qdos and hopefully you will increase FPC visibility!

[Mike.Cornflake]

@qdos:  You go ahead and do whatever you want to.  I wish you the best.  When the time comes for testing, drop me a line.

[snorkel]

Nice Qdos.
Any plans for a form designer?   I wouldn't mind seeing Free Pascal renamed to eliminate the "Free" part. 
many many corporate manager types won't allow FPC to be used simply because the name contains the word "Free"