Number literals (Was: Re: Case Range Statement ..)

[Andrei Alexandrescu]

Yigal Chripun wrote:
> Let's compare: 
> Bartosz took upon himself the task of designing concurrency for D. For that he put forth a lot of effort to explain the design and the rational behind it, he also encouraged discussion of the subject in the NG and answered questions.
> You've re-implemented phobos from scratch, but in contrast, your phobos (however great and superior in design) is a one man job. you didn't involve the community in the process like Bartosz has done even though you know that a new API in the standard library has a huge impact on the community. 

The new phobos is mostly based on ranges, which have been thoroughly 
discussed in this group before being implemented. I have often discussed 
dilemmas I've had and am still having (such choosing between value 
containers and reference containers) in this group. Ideas and code 
submitted by others have been integrated into Phobos with credit, most 
recently Shin Fujishiro's code for printing and parsing enums as symbols 
which has been added to std.conv (and Shiro has been added to the list 
of module authors). I'm still working on integrating Shin's code for 
BlackHole and WhiteHole, which is awesome but very involved.

By the way, it seems that Walter hasn't regenerated the documentation 
when releasing 2.031, for example search this page for "enum"

http://www.dsource.org/projects/phobos/browser/trunk/phobos/std/conv.d

vs. this page:

http://www.digitalmars.com/d/2.0/phobos/std_conv.html

> especially irritating to me is the fact that after all the hard work Bartosz has put into designing and explaining his idea you simply canceled it with a two line post. Have you found flaws in his design? please explain to us simple humans so that we can share that understanding and provide a better design.

This group has discussed Bartosz's design only a little. I've discussed 
a lot more with Bartosz and others. My issue with his design can be 
resumed to one point. The design helps with only one style of concurrent 
programming. That particular style is IMHO becoming obsolete and should 
not be the focus of a major language addition. Anyway, you are ascribing 
way too much influence of my opinion: Walter makes the decision, and 
Bartosz communicates with him as much as I do.

> Remember, There's a saying in business - "The customer is always right" and here the community of D programmers are those "customers" of the D language. No matter how stupid, annoying, ignorant or whatever they can be (I'm sure I fall under some of these categories) if you continue condescending them they *will* go to the competition (Python, C#, Ruby, Java, whatever).

You're right, thanks for sharing that.

> Personally, I already switched. I see no future in D as long as you keep torturing its community. 

Now, torture is illegal isn't it :o).


Andrei