Builder: A New Development IDE Being Built For GNOME

[anda_skoa]

Why am I saying that? A good IDE does understand the code you write. To do so it has a code model for all the languages it supports. That model is what powers refactoring operations (like renaming methods, etc.). It is also what powers the syntax highlighting, the code completion, for code navigation -- and basically everything else! In a good IDE even the commit editor is hooking into the code model so that it can offer to complete function and class names! That is the "integration" in "Integrated Development Environment". Whether or not e.g. the debugger is run in a separate window and/or looks slightly different from the rest of the UI does not matter:-)

Indeed!

Years back an IDE was just the integration of different tools under one "roof", i.e. withing one user interface, but nowadays integration has to go way deeper.
E.g. as you put so well, understanding the code being written.

In the special case of Vala even more difficult: e.g. understanding which piece of the code was most likely responsible for a warning or error of the C compiler.

However, the current concepts could be a like a prototype stage, to test what can be done in tihs kind of system. If it turns out that certain things need tighter integration that should still be an option.

Cheers,
_

[anda_skoa]

Apple isn't divided between KDE, Gnome and few other DEs.

GNOME isn't devided between different DEs either, the only DE by GNOME is GNOME Shell, no?

Cheers,
_

[anda_skoa]

Market share impossible to convert to money is not worth much

You are most likely referring to monetary worth, but worth is a widely applicable concept.

The worth of market share could be influence, recognition, pride, gratification, etc.

Cheers,
_

[petah]

hopefully http://www.codelite.org will get a cursory glance before gasoline dousing starts. I use it mostly on debian and OSX (didn't try Win), my 2 main targets are clang trunk with libc++/cxxrt/LLDB and gcc 4.9.1/GDB, using every C++11 feature possible. It interfaces with make/cmake or whatever, has a RAD plugin (for wx), it's tiny, stable and fast as hell with an sqlite ctags background engine. Maintained by a couple of awesome guys apparently waiting for wx's Android port to cover that too.

-- p

[Karl Napf]

Indeed!

Years back an IDE was just the integration of different tools under one "roof", i.e. withing one user interface, but nowadays integration has to go way deeper.
E.g. as you put so well, understanding the code being written.

Thanks! I have not hung out on Phoronix too long yet but had frankly not expected such a positive reply from this site

Usually I get told that a real hacker is faster with vim than all those wannabes with an IDE and that I should return to Visual Studio -- which I never used by the way.

However, the current concepts could be a like a prototype stage, to test what can be done in tihs kind of system. If it turns out that certain things need tighter integration that should still be an option.

Sure, it is just software, software is easy to change. But I do not have high hopes for a project that starts by running off into the wrong direction

I am using Qt and Qt Creator myself and did never feel at home in the gnome community. But I do care a lot about having great tooling for all free software out there, so I do follow all kinds of IDEs, compilers, debuggers and whatnot that are getting developed as free software. It is depressing how far we are behind commercial offerings in that space and how ignorant vast parts of our community -- across all projects -- are on that topic.

As an example of how bad the situation really is: I spoke to somebody actually showing of a couple of prototypes for a new IDE a few month back and was told "this does not need to be full-featured, this is only for beginners. I expect real developers to move on to use vim".

[anda_skoa]

Thanks! I have not hung out on Phoronix too long yet but had frankly not expected such a positive reply from this site

There are all kinds of people reading here

Usually I get told that a real hacker is faster with vim than all those wannabes with an IDE and that I should return to Visual Studio -- which I never used by the way.

Recently?
I I remember that mostly from the late 90s, early 2000s, when IDE were often very bad.
Many of them lacked support for buildsystems other than their own, had nor or very limited support for commandline building (a nightmare for automated builds), had text editor components that lacked most of the advanced text manipulation features that made vim or Emacs so powerful.
Heck, some even lacked such basic functionality as regular expression based search.

Back then a good programm, using the tools written by programmers for programmers, could easily be more productive then anyone stuck with an IDE.

If you got a comment along these lines recently then maybe by someone stuck with old knowledge about IDEs.

Sure, it is just software, software is easy to change. But I do not have high hopes for a project that starts by running off into the wrong direction

True, there is the risk of that.

Cheers,
_

[gufide]

I have not problem with C++11 but it will take a lot of time to create an IDE. I think there are many IDEs out there that support C++11. AFAIK C++11 compliant compilers are limited to two, clang/gcc. But vala in theory can use many c compilers. Personally I think learning C++ is a big investment in time and effort. Vala is for everyone. Even for me.

I agree that C++ is one of the most complex one out there, but when you know your tool well, you can beat anything you could do with other language. The downside is a high learning curve but C++11, despite this downside has became very easy to use, and you can even get automatic memory management and more!

[matt_g]

There are all kinds of people reading here


Recently?
I I remember that mostly from the late 90s, early 2000s, when IDE were often very bad.
Many of them lacked support for buildsystems other than their own, had nor or very limited support for commandline building (a nightmare for automated builds), had text editor components that lacked most of the advanced text manipulation features that made vim or Emacs so powerful.
Heck, some even lacked such basic functionality as regular expression based search.

Back then a good programm, using the tools written by programmers for programmers, could easily be more productive then anyone stuck with an IDE.

If you got a comment along these lines recently then maybe by someone stuck with old knowledge about IDEs.

Do you have any examples of a good modern IDE? I started programming in late 90's /early 2000's using EMACS I've been using it pretty exclusively for almost 15 years now. I guess my emacs is pretty heavily customised but it is what I am used to and I've found its the most productive editor I've used. I can understand someone new not wanting to learn it as until you set it up how you like it is hard to get value out of it but once you have all the macros setup I think its fantasic EMACS is pretty much the one tool I can't be without for my work.

It's what I learnt to use at University and I have used it ever since. I can use VI in a pinch and some of my work colleagues prefer to use VI.

Occasionally I have needed to demonstrate something on a colleagues PC, he uses Eclipse and I cannot stand it, a lot of it is probably unfmailiarity/inertia but I find it slow and laggy and just frustrating in general.

I signed up to post on this forum because I guess I'm someone like you mention who is only familiar with old IDEs but what are the complelling advantages of a modern IDE. At this point in my career I doubt you will be able to persuade me but I'm willing to listen.

Mostly I write C (and some Fortran) and make heavy use of GDB.

[Karl Napf]

Do you have any examples of a good modern IDE?

I always find it hard to recommend any specific IDE: It depends a lot on the languages you use as well as the workflows you are accustomed to. Basically the important thing to look out for is whether the tool has a code model for the language you want to work with.

Good commercial IDEs are of course X-code on mac, visual studio on windows. There is also everything from jetbrains (intelli-j), which are really great for the languages they support. The only good free software IDEs I am aware of are Qt Creator and KDevelop.

Qt Creator is the one I prefer personally, it being targeted on C++ development (which is what I use 95% of the time). It supposedly also works well for C, at least I met a couple of kernel developers using it.

I would recommend to try different tools, giving each a couple of days. And I strongly recommend skimming the manuals (or other documentation if available). All IDEs I ever used are complex tools and spending an hour to skim over the documentation can save you days working with that tool later.

[blackiwid]

I always find it hard to recommend any specific IDE: It depends a lot on the languages you use as well as the workflows you are accustomed to.

Workflow ok, but language? I think a ide have failed its job if its only good for one language.

I am also a emacs user, sadly not 15 years I curse my schools and even the university of applied sciences I was taht they never showed me emacs my live would be different I am pretty shure. I would have done so much more work in so much less time And would be today even omre efficiant with using it. My emacs config wouldbe 5000 lines long not only 500 lines

no really emacs is so strong because there is a modul for everything, EVERYTHING. and some own killerfeatures like orgmode.

Thats the problem of every other editor, u would need at least 10 years to have that much integration in the hole mostly foss community.

It would have to be something like emacs maybe written in python with also a package manager like u have in emacs with 1000 modules in github. And then u only would have replaced the internal main language and archived not much more, maybe more sane defaults.

A editor that only is written for one language could never get so much friction that u get so much 3rd party support than u have with emacs.

But ok maybe for that one small task u get a somewhat ok ide. But the cost to have a good designed ide with good shortcuts with some kind of configurability etc only to support one language or one framework is insane. Like some kids trying to develop a mmorpg from scratch.

I think a ide lives from integrating so much tools as possible in a good way. to not have several apps for every small task, to not use a seperate console to as example compile stuff or something like that. So u have stuff like tramp in emacs that gives u transparent access to different file paths/backends... etc. creating all that stuff for a very special ide for one special case, is at least no easy task.

But I think its maybe for some hobby programmers, that try to start hack something, if u continue into that u need something more "integrated" I think.