[skipping over most of the last 3 pages because I don't have time]

One thing that irks me about this discussion (and most other open-source discussion) is that open source is equated with Linux and Linux is equated with Unix.

In my experience the FreeBSD community for example is much less dogmatic (though smaller) than the Linux community. There seems to be a trend in general towards less GPL zealotry though, in my experience grade-A fanatics are rare even on slashdot lately. Of course there's still a group of old fossils on the one hand and clueless wannabes on the other that treat free software as a dogma, and they are the most vocal ones. I don't think these sentiments are representative for the open source community as a whole anymore though, fortunately.

Anyway, I'm excited about this IDE of yours Yann. I'm still using Emacs whenever I have the choice because other open IDE's annoy me to no end. I have never had the pleasure of using Visual Studio (I'm poor), but I've recently been forced to use Borland CBuilder and that hardly seems better than KDevelop.

Taking a practical approach to this open source issue: I think the product would be helped in the end by being entirely open source, because that way people can...
* Make contributions
* Take over development if you abandon it
* Port it to their favorit obscure Unix

Reacting with hostility to a closed prototype is entirely clueless though. It's a gift. Do not bitch about gifts. Going back to the AP's analogy near the start of this thread: I am a vegan too, but if someone offers me meat I just politely refuse.

[edit] Oh yeah I forgot: The GPL needs to die. I much prefer BSD-style do-what-you-want-with-this licenses.

Yes, there are GPL monsters out there, but come on. I have no problem with downloading closed source linux libraries, but you are walking over the entire linux paradigm. What do you expect people to say when the entire operating system is revolving around an open source, free software solution. If you're going to make it free anyways, why not release the source you wrote, and keep YOUR in-house libraries closed. Also, I think this also is a bad trend for linux programmers. Once you start closing source, people will start selling, and that just crushes the hopes of linux.

I seem to recall Yann saying he cant just release his in house libs as closed because they make use of templates and as such you are required to give out all the code to the templated part.

I dont see why 'selling' crushes the hopes of linux? Unless free software is that bad that you think that as soon as commerical software you have to pay for gets on the scene everyone will jump ship to that? [smile]
Infact, the introduction of commericalism is what could save it and push it on, as commerical apps tend to demand more uniform libraries and install systems, making it easier for the end user and making linux easier makes it more accasible (sp?) and in theory more popular ("hey, have you tried linux? yeah, its as easy to use as windows and has lots of free software!"), thus pushing it on wards.
Unless ofcourse by 'the hopes of linux' you mean keeping it in the server room only and on a tiny market share of desktops? [grin]

Yann: What I don't understand is the "fixed widget" problem. Unfortunately, you haven't given much details, but it sounds as if you made some changes to Gtk in order to fix some bugs. Now apparently these changes are minor - if they came near to a full rewrite, you wouldn't complain about the perceived necessity to do so because of the license.

So, you've got a patch for Gtk. What is it that stops you from pushing it upstream to the Gtk developers? All you have to do is publish that patch - it has no effect on any other obligation you might have when it comes to linking or distributing source code.

The linking issue is a completely independent topic altogether, and as others before me have pointed out, there are more or less well-established ways to get around it. You could just package the "optimal version" of important .so's together with your IDE. That's what ut2004 does. You could also supply one dynamically linked version (with the modified .so) and one "user-friendly" statically linked version of the executable. That's what Loki did.

Apart from that, good luck with your project, it may turn out to be interesting. I'll definitely give it a try, though I have to agree with bani that it's probably a waste of effort, especially since you don't intend to continue its development. I guess that's just how game developers work - always trying to reinvent the wheel :) (I know I've made that mistake far too often)

cu,
Prefect

Quote: Original post by CoffeeMug
Well, to be fair your schedule can be very granular and you can release a new version after adding only a few new features as opposed to hundreds (or thousands).

Yes of course, you adjust the granularity of product feature updates to fit your specific target market. But you still operate on a well defined feature set - the new featues required for the next version are sketched out, their design finalized, implemented, and tested. That's the cycle I was talking about. In most OS software this is not the case. They just add any feature at any time, as they drop in, often without even questioning the actual useability and conformance with the products philosophy. The result are inconsistent products, with conflicting and sometimes incompatible features that change at every nightly build.

Quote: Original post by Doc
What real advantage is there for anyone to create a common set of non-GPL base libs? A company would not be able to retain any sort of competitive advantage if it released its non-GPL base libs. The "Linux community" can generally see little point in it.

I seem to recall, that the "Linux community" wanted their OS to become popular on the desktop market ? Well, if they still follow that goal, then you have found your point: with only "license challenged" libraries, there won't even be something remotely similar to a desktop Linux revolution.

Quote: Original post by smokes
Please keep up the good work and don't let religious dickheads make you quit linux development :)

Heh, thanks :) Don't worry, I'll certainly not give up Linux development - many of our commercial customers run Linux. Of course, that's not open source software either...

Quote: Original post by cb007sax
Also, I think this also is a bad trend for linux programmers. Once you start closing source, people will start selling, and that just crushes the hopes of linux.

I'll tell you what - this is the moment I'm waiting for, because this is the turn around that will make Linux a real competitor to Windows on the desktop. We need commercial software on Linux, this is actually the only hope for its future. I really love Linux, I find it technically far superior to Windows in many aspects. But if you want it to become popular (and I guess we all want this), then it has go commercial. The open source community would even benefit from this: by getting direct commercial competition, they coudl target their efforts and improve the quality of the open source products. Look what the direct competition of Internet Explorer did to the quality of Firefox ! If it wasn't for the strong will to take market share away from IE, open source browsers would still be on the level of Lynx (yeah, I hate Lynx).

Quote: Original post by Prefect
Yann: What I don't understand is the "fixed widget" problem. Unfortunately, you haven't given much details, but it sounds as if you made some changes to Gtk in order to fix some bugs. Now apparently these changes are minor - if they came near to a full rewrite, you wouldn't complain about the perceived necessity to do so because of the license.

So, you've got a patch for Gtk. What is it that stops you from pushing it upstream to the Gtk developers? All you have to do is publish that patch - it has no effect on any other obligation you might have when it comes to linking or distributing source code.

Well, I will certainly send it to the Gtk developers. But first, I don't want to wait until it finally gets included in the CVS - if it gets included. Second, I cannot require everyone that wants to try out the IDE to install the latest GTK version, especially since there seems to be a lot of backward compatibility issues with Gnome. So yeah, I'm left to distribute the source of all my widget modifications with the product. It's no big deal, but it's still awkward.

Quote:
Apart from that, good luck with your project, it may turn out to be interesting. I'll definitely give it a try, though I have to agree with bani that it's probably a waste of effort, especially since you don't intend to continue its development. I guess that's just how game developers work - always trying to reinvent the wheel :) (I know I've made that mistake far too often)

a) I'm not a game developer
b) If all available wheels are broken, then you don't have a choice other than reinvent it
c) From many of the responses so far ("I've given up on Linux development, because of no good IDE"), I don't think the effort is wasted ;)

---

Oh yes, and a quick progress report: we have now fixed all the important widgets (especially the file chooser, a new gnome independent version of an icon view, and some strange behaviour with the general combo boxes). GtkCellRenderCombo is still a major pain in the ass, because it reacts - well - in a very unusual manner. A non-GPL HTML widget is still out of reach.
Well, I guess I'll be able to post a first demo pretty soon now.

One thing Visual C++ Express is missing is refactoring. Since I have two versions of my engine, .NET and C++, I often code things first in C# to make use of its refactoring and better intellisence and then copy the code over to C++ and change it as needed. It might be a particularly hard thing to implement in your IDE, but Visual C# Express-like refactoring would be a great feature.

Quote: Original post by Yann L
In most OS software this is not the case. They just add any feature at any time, as they drop in, often without even questioning the actual useability and conformance with the products philosophy.

This problem is related to inexperienced management, not open source. The reason it's so apparent in open source projects is that *anybody* can start one, while commercial software has considerable entry level barriers. At least with GPL experienced people can (and often do) fork successful projects and polish them due to superior management skills.

Not to hijack Yann's thread, but for all you who complain about that emacs hasn't got a class viewer and intellisense, go download cedet already.

Now, about this IDE. Will it have an integrated debuger even comparable to VS? Because that's the only thing I've really missed since I switched from windows/vs to linux/emacs(+cedet)/gcc.

Quote: Original post by Luctus
Because that's the only thing I've really missed since I switched from windows/vs to linux/emacs(+cedet)/gcc.

So the whole loss of sanity due to autotools, managing makefiles, and C-x C-u C-e wait 372ms M-s C-r M-t business didn't bother you? [smile]

Quote: Original post by bani
whats interesting is that microsoft's eula forbids the obverse, which is using visual c to develop anything gpl.

FUD, or, more accurately, bullshit.

Carry on.