All this talk about 1.0 makes me worried.

[John Demme]

kris wrote:

> Charles wrote:
>>  > I suppose many people like the idea of placing a stake in the ground ~
>>  > as means of marking/stamping progress?
>> 
>> I know of alot of people ( myself partially included ) that are waiting
>> for a D 1.0 untill they really commit to using D.  Untill it reaches a
>> 1.0 ( which feels like never -- in which case if it is going to take
>> another year or greater we should be talking about how to handle and
>> work with a perpetually changing language - id be curious to know how
>> often you have to update mango in response new DMD releases )
> 
> There was a bit of turmoil back when it started (~March 2004), but after
> that? Hmmm ... discovering that AA's were broken caused some
> re-engineering; the "length" pseudo-keyword issue caused some more. Then
> there was quite a bit of effort to cleanup using -w. The char, wchar,
> dchar support could not really have been done until templates came
> along, so that doesn't really count, I suppose? Mango has always used
> 'bool' instead of 'bit', so that wasn't an issue. To be honest, I think
> internal redesign has caused more changes than language evolution.
> 
> 
>> i doubt
>> 'corporations' would even consider using it.  Even post D 1.0 is going
>> to take a big effort from the community to get D in the mainstream.
> 
> Very true. However, those same corporations likely won't consider D
> until library support is notably better? I've always felt that would be
> the sticking point, and that Phobos was a bit thin for that purpose.
> This leads me to conclude thusly:
> 
> D will not be ready for commercial usage until the availability of
> libraries reaches some critical mass. Who is going to write those
> libraries? My guess would be the the early adopters? So, why is it that
> there's perhaps only a few handfuls of people who are prepared to make
> that happen? Don't wish to be critical of anyone, or start wagging any
> fingers, but surely those calling for a v1.0 should be equally
> determined to construct the libraries?
> 
> - Kris


I agree with both of you.  Here is what I would like to see soon: DMD
1.0RC-1 with a guarantee that all of the language features and syntax are
tacked down, and there will not be more changes that will break code.  From
there, Walter handles only fixes and feaures like debugging output.  We can
also spend time getting libraries ready to go (read: libraries to be
included with DMD)... I don't want DMD 1.0 to be released until there's a
decent library (read: Mango) to be released with it.

~John Demme