Boris Wang wrote:
> The reference return type , should be must-have feature for D 1.0.
> 
> "Tom S" <h3r3tic@remove.mat.uni.torun.pl> ??????:e7oc3h$261t$1@digitaldaemon.com...
> 
>>Boris Wang wrote:
>>
>>>In addition, reference return type, variable template parameter, etc.
>>
>>These are not must-have features for D 1.0
>>
>>
>>
>>-- 
>>Tomasz Stachowiak  /+ a.k.a. h3r3tic +/
>>
> 
> 
> 

At some point, you have to say enough is enough, and just make the 1.0 release.  Will I be upset if those features don't get into 1.0?  Not really.  After all, the language will continue to evolve after the 1.0 milestone.

    -- Daniel

Some feature go into D, may make D not compatible with old version. So these features need go into before 1.0.


"Daniel Keep" <daniel.keep.list@gmail.com> ??????:e7qm4a$ugt$1@digitaldaemon.com...
> Boris Wang wrote:
>> The reference return type , should be must-have feature for D 1.0.
>>
>> "Tom S" <h3r3tic@remove.mat.uni.torun.pl> ??????:e7oc3h$261t$1@digitaldaemon.com...
>>
>>>Boris Wang wrote:
>>>
>>>>In addition, reference return type, variable template parameter, etc.
>>>
>>>These are not must-have features for D 1.0
>>>
>>>
>>>
>>>-- 
>>>Tomasz Stachowiak  /+ a.k.a. h3r3tic +/
>>>
>>
>>
>>
>
> At some point, you have to say enough is enough, and just make the 1.0 release.  Will I be upset if those features don't get into 1.0?  Not really.  After all, the language will continue to evolve after the 1.0 milestone.
>
>     -- Daniel

Don Clugston wrote:
> Boris Wang wrote:
>> In addition, reference return type, variable template parameter, etc.
>>
>> The D is far from the 1.0 mile stone.
>>
>> It's good time for us to have a rest, may be one years, two years...,  :(
> 
> The problem is, you cannot develop a stable 1.0 library until you have a  stable 1.0 language. (Important example: template container classes can't be standardised until we know how good the implicit function template instantiation will be by 1.0).
> Secondly, features such as array literals, variable template parameters, etc, have been stated to be a post-1.0 feature.
> 
> 

Indeed so. And I think this is very relevant to D's development process because it indicates that early Phobos/library development may not be as productive as late development, because with early development, more things may have to be changed afterwards. Perhaps this is the reason why Phobos is less mature that the rest of the toolchain(the compiler)?
Note, I'm not saying early Phobos development is never worthwhile, (that would likely depend on each particular case), I'm just pointing out too this general consideration.

-- 
Bruno Medeiros - CS/E student
http://www.prowiki.org/wiki4d/wiki.cgi?BrunoMedeiros#D

Bruno Medeiros wrote:
> Don Clugston wrote:
>> Boris Wang wrote:
>>> In addition, reference return type, variable template parameter, etc.
>>>
>>> The D is far from the 1.0 mile stone.
>>>
>>> It's good time for us to have a rest, may be one years, two years...,  :(
>>
>> The problem is, you cannot develop a stable 1.0 library until you have a  stable 1.0 language. (Important example: template container classes can't be standardised until we know how good the implicit function template instantiation will be by 1.0).
>> Secondly, features such as array literals, variable template parameters, etc, have been stated to be a post-1.0 feature.
>>
>>
> 
> Indeed so. And I think this is very relevant to D's development process because it indicates that early Phobos/library development may not be as productive as late development, because with early development, more things may have to be changed afterwards. 

The new syntax introduced in 0.161 is a case in point. It makes delegate-based interfaces far more attractive. AFAIK, none of us had any idea it was coming.

Perhaps this is the reason why
> Phobos is less mature that the rest of the toolchain(the compiler)?
> Note, I'm not saying early Phobos development is never worthwhile, (that would likely depend on each particular case), I'm just pointing out too this general consideration.

It's definitely a contributing factor. Still, I think the main reason is that the entire D community combined is barely as productive as one Walter.
Large parts of the language are almost fully mature, and I've concentrated my effort on those (compile-time metaprogramming, numerics). But it doesn't make sense to work on something like DTL which is guaranteed to become horribly obsolete quickly.

I wonder if we could have a partial 1.0 spec? Eg, declare a feature freeze on certain parts of the language, and gradually lock more of it down.

Boris Wang wrote:
> Some feature go into D, may make D not compatible with old version. So these features need go into before 1.0.

What makes you think we *can't* make non backwards-compatible changes after 1.0?  Take a look at Python: they're currently in the middle of making a backwards-incompatible change to the way imports work?

Is it the end of the world?  Is there going to be mass riots from Python users?  No, because it's an *improvement*.  Does code need to be rewritten?  No: just keep using the old Python interpreter.

There are things I'd like to see in the language, but I'm quite content to wait until after 1.0.  I think it's far, FAR more important to get the compiler stable, bug-free, and have a concrete spec soon than developing an acute case of feature-creep.

	-- Daniel

> "Daniel Keep" <daniel.keep.list@gmail.com> ??????:e7qm4a$ugt$1@digitaldaemon.com...
> 
>>Boris Wang wrote:
>>
>>>The reference return type , should be must-have feature for D 1.0.
>>>
>>>"Tom S" <h3r3tic@remove.mat.uni.torun.pl> ??????:e7oc3h$261t$1@digitaldaemon.com...
>>>
>>>
>>>>Boris Wang wrote:
>>>>
>>>>
>>>>>In addition, reference return type, variable template parameter, etc.
>>>>
>>>>These are not must-have features for D 1.0
>>>>
>>>>
>>>>
>>>>-- 
>>>>Tomasz Stachowiak  /+ a.k.a. h3r3tic +/
>>>>
>>>
>>>
>>>
>>At some point, you have to say enough is enough, and just make the 1.0 release.  Will I be upset if those features don't get into 1.0?  Not really.  After all, the language will continue to evolve after the 1.0 milestone.
>>
>>    -- Daniel 
> 
> 
>