my demise

[Kyle Furlong]

Fredrik Olsson wrote:
> Kyle Furlong skrev:
> <snip>
>> The OP was about the failure of D as a good fit for the poster's 
>> project. I was merely pointing out that "good enough" obviously wasn't 
>> good enough in this case.
> 
> And my points are that
> 1. Emacs (or any other text editor) and gdb (or any other debugger) is
>    more than good enough. Even a bonus, as tight integration with a
>    "standard tool" would ruin it for users without access to the "right
>    standard tool".
> 2. Richard Koch was well aware of the lack of:
>    * "Visual Studio"-like integrated tools.
>    * GUI-libs, with design tools.
>    * No 1.0 release.
> 3. An "Visual Studio"-like tool is not required for a language to be
>    accessible to beginners, or for a language to become popular.
> 
> Childish to complain about that 4 months later, even stupid to waste 
> that much time when it should have been obvious it was not what he 
> wanted from the start.
> 
> 
> I have no clue what he means by "usable libs, extensions and standard 
> algos". There are many, many resources for grabbing and reusing existing 
> code for D out there, for a wide variety of purposes. And I find them 
> usable, and when I don't I write my own.
> 
> 
> I would never come up with the idea to make an application with a 
> graphical user interface using D today. So for those apps I choose 
> another tools. I would be willing to invest some time fixing that 
> deficiency if I saw a suggested solution I liked, writing a new GUI-lib 
> is a huge task, and unless the involved love and believe in the idea 
> 110% the effort is wasted. Plus anything that is not platform 
> independent is not interesting at all for me, and unless it takes 
> advantage of the features of D, why waste the effort?
> For the time being I use D for what I know it is good at: command line 
> tools and server software without user integration.
> 
> 
> // Fredrik Olsson

Like I said, people have disparate ideas of "good enough." If you are 
content with no IDE project management, with using command line tools, 
and manual debugging, thats fine for you. But many, many people out 
there enjoy the ease of development that niceties like Visual Studio 
give. I realize that many of you come from a C++ background, and this 
might be the norm, but in the Java/C# world people dismiss D for the 
lack of these things. So if D is trying to woo that crowd, just throwing 
your hands in the air and saying "ITS GOOD ENOUGH!" is NOT good enough.

Like you said, I think its a case of the right tool for the right job. 
At this time, D is not suitable for anything other than code that 
doesn't interface with anything but a command line, i.e. it is more of a 
systems language. However I do think that D /should/ also be an 
application programming language, and it should be easy. I hit 
performance bottlenecks all the time in my C# apps, and I would love to 
be able to redesign them in D. At this time though, its not worth the 
effort.

Maybe I was wrong in thinking that D was headed in this direction, and 
all Walter wanted was an easier C for writing robust OO operating 
systems and drivers and libraries, web servers and infrastructure apps. 
But if that is the case, I'm out of here, because that's not what I do.