End user experience with D

[Ludovit Lucenic]

On Sunday, 1 September 2013 at 22:21:10 UTC, Ramon wrote:
> Ludovit Lucenic
>
> While one may discuss about IDE or cli, I feel that an IDE is 
> at least an immensely helpful starting point.

I would add to it, that IDE could also solve some of the problems
discussed e.g. on the Panel talk in DConf. I mean the code
upgrade automation once the compiler comes in newer version. We
solved this (in the job mentioned) by having an IDE that did an
automatic upgrade on the project opened. I absolutely agree with
Walter that it seems weird to have a compiler which modifies the
code :-) But having an IDE that does this is pretty much
convenient. Not talking about the whole bunch of other minor 'D
UX' issues that could be managed easily with a little help of an
IDE.

> I agree that one should be able to do programming with no more 
> than a simple editor, a compiler/linker, and a debugger (!). 
> Newcomers, however, (and even many seasoned programmers) 
> handily profit from an IDE (and I don't mean Eclipse which I 
> count as a moloch that, to make things worse, is based on java).

I did not meant Eclipse is 'the way'. I said eclipse-like
debugging user experience is still something worth taking
inspiration from. Anyway I am persuaded having IDE written in D
fits the needs the best. Just look at all the relevant IDEs
around. Every and each of them is used primarily to target the
language the IDE is written in :-) MonoDevelop for Mono, Visual
Studio for .Net, Code::Blocks for C++, Eclipse for Java... Have
you noticed it ?


> Finally and possibly most importantly, basically not having 
> fully working  debugger support is a very serious lack. I may 
> be wrong (actually I hope to be wrong) but from what I know, 
> this issue doesn't seem to be seen as having very high priority.

I got to admit this is the most surprising thing about the D I
came across. Well, let's not complain about it, let's do
something with it. I assume the GDB is cross-platform solution
for debugging that seems to be furthest. Therefore I chose
MonoDevelop to improve its debugging capabilities (at least I had
something to start with) and add the visualization needed by some
of us.
However, I don't think I can move things much forward on Windows
unless we have ready to use GDC/LDC executables (that one can
download and install easily on Win platform) which would by-pass
the DMD/GDB/Win problem then. Moreover, I don't like to code in
C#, when there's D around :-) but for the beginning...

> I'm more and more getting the impression that D could be a 
> programmers dreams come true if only it were practically and 
> productively useable. Sure, it's not D's fault. But in the end 
> we need a fully working debugger, some good editor support 
> (say, on a Scintilla/Scite level), preferably at least one 
> cross platform non-java IDE (like Code:Blocks) and easy access 
> to C libraries in order to actually and productively work with 
> a language.

Could not express it better!

> D's bias toward Windows doesn't help either.

Absolutely disagree. My experience is quite the opposite. I mean,
you might be referring the VisualD stuff, but regarding the
debugging, it is not mature enough anyway.

> Looking at it from a marketing perspective, i.e. looking for 
> what potential users (read: "the large crowd out there") would 
> look for, I see that the creators of D have done an excellent 
> job that, however reminds of a Porsche engine on a makeshift 
> chassis.

I am idealist myself too, but got to admit: perfect analogy.

> Frankly, if we want D to attract developers and to get the 
> position it undoubtedly deserves that will not be achieved so 
> much by finetuning D or even phobos but by being able to offer 
> a well-rounded and well useable package (and soon after good 
> docs for what we have so far).
>
> A+ -R

The work with D has to bring joy, but sometimes it leads to
confusion. We're strong, we will come through, but have anyone
already see the tunnel ? Not speaking of the light :-)

Ludovit