Future of DMD 1.xxx

[Lutger]

Simen Haugen wrote:
> "Bill Baxter" <dnewsgroup at billbaxter.com> wrote in message 
> news:fbj50b$2dd2$1 at digitalmars.com...
>> I've been meaning to ask...
>> Am I right in thinking that basically nothing from WalterAndrei.pdf will 
>> be going into a DMD 1.xxx version?
>>
>> If so, that's a bummer because some of the upstream libraries aren't on 
>> D2.0, and the maintainers have stated that they're not going to port to 
>> 2.0 "until the 2.0 design is pretty conclusively settled", or "until it is 
>> near to release".
>>
>> --bb

But if these features would get in 1.0, the whole stable vs unstable 
branch would be undone (could just as well merge back), and D will start 
to be all moving target again. We can't live on the bleeding edge and 
have stability at the same time. I'd like to think the future of D 1.xxx 
will be the applications written in it.

> And others are dropping 1.0 even before 2.0 is stable...
> 
> I started using D when I saw the 1.0 mark, but I wished I had waited... 
> Tango vs. Phobos and the fact that 1.0 was not frozen has caused a lot of 
> headaches. Now also 1.0 vs. 2.0 is starting to be a problem. Guess it will 
> take quite some time before things will get stable.
> 

Probably true, but it also depends on what you expect from a stable 
system and want out of D. Really the fork is what I think D 1.0 should 
have been to begin with, and things have been quite stable since 
compared to before this split. Now the source of instability is not 
changing language features, but lies more in the library landscape.

I don't see 1.0 vs 2.0 as a problem, on the contrary, it is just a 
trade-off that gives you a choice: develop on the bleeding edge or 
program to a language that won't change every couple of weeks / months. 
Before, we had only bleeding edge, with sometimes new or changed 
languages features twice a month! Now you have the choice, it's a good 
thing. With some care, upgrading to 2.0 in the future will not be a big 
problem.

Something similar goes for Phobos / Tango, you have a choice now. 
Although here it also fragments other libraries, which D2.0 doesn't seem 
to do yet, and I hope that will not happen soon.