DMD 2.000 alpha release

[eao197]

On Wed, 20 Jun 2007 14:19:47 +0400, Daniel Keep  
<daniel.keep.lists at gmail.com> wrote:

> eao197 wrote:
>> <old-talk>
>> Some time ago I waited for D 1.0. Now I'm awaiting D 2.0. Then, I'm
>> afraid, I would wait for next major stable version of D. There ins't any
>> sign that after almost seven years of development D will became stable
>> and mature language.
>>
>> Excuse me for repeating that.
>> </old-talk>
>
> So use D 1.0--it's stable now!  As Walter has said, it's just going to
> be maintained with bugfixes now, no more moving target.
>
> If your argument is "I can't use D 1.0 because D 2.0 exists", then I
> suspect you won't be happy until D has ceased all evolution and updates,
> effectively dying, at which point you'll probably start using a
> different, shinier language.

It is not easy to me to start use different language because I have big  
enough rather old code which I should maintain and advance. At best I will  
be able to maintain old C++ code and write new code in D.

<old-talk-again>
In my option the problem with branching D into 1.0 and 2.0 not in  
language. The problem in the libraries, tools, documetation and people  
around D.

For example, now D is attracting new people. Someone want to write new  
tools/libraries for D. Which version will they choose? Stable, but old, D  
1.0 or unstable, but very promising and attractive D 2.0? I'm affraid many  
of them choose D 2.0 (because I would do that). And because of that in  
time D 2.0 would have more tools/libraries than D 1.0.

The same thing with documentation. The same thing with teaching people to  
program in D. Even worse. Designing programs without consts/invariants is  
much different than designing programs with consts/invariants. So why to  
spend time to teaching new programmers to use old-style D instead of  
teaching them new-style D?

I like to see consts/invariants in D. But... Implantaion of D as main  
programming language (instead of C++, like in my case) in industry needs a  
lot of efforts to re-teaching people, creating some proof-of-concent  
projects, defining some corporate politics and best practices. Even in the  
best conditions it may took months, may be years. Why to make thing ever  
worse with language which is always in beta-stage?

I need to remember that in addition to early adopters (like you, Daniel)  
and technologist (like I) there are many, many rather conservative,  
pessimistic and simply lazy programmers who don't want to study something  
new every month or even a half-year. But they do much, much work. Please  
keep them in mind.

Excuse me for repeating that yet more time.
</old-talk-again>

-- 
Regards,
Yauheni Akhotnikau