If D becomes a failure, what's the key reason, do you think?

[Dave]

BCS wrote:
> In article <e8olc3$12i4$2 at digitaldaemon.com>, Sean Kelly says...
>> Don Clugston wrote:
>>> I think it's a two-step process. My opinion:
>>> Announce a 1.0 feature freeze immediately. Don't actually announce DMD 
>>> 1.0 until the regressions related to protection are fixed. A 1.0 
>>> announcement is a major public relations opportunity (one of the biggest 
>>> the language will get), it's important to avoid a PR disaster. But 
>>> internally, we can view 0.162 as DMD 1.0 RC1.
>> Agreed.  As soon as the feature list for 1.0 is decided and any 
>> contentious issues have been resolved (ie. the visibility issue being 
>> discussed now) I think an announced feature freeze would be the proper 
>> first step towards a 1.0 release.  Compiler bugs could continue to be 
>> worked on, and perhaps a joint effort could be made to fill out the spec 
>> in places it's either confusing or sparse.
>>
>>
>> Sean
> 
> How about feature-freeze NOW at what is in DMD already and things that are under
> major discussion?
> 
> Reference:
> http://www.digitalmars.com/drn-bin/wwwnews?digitalmars.D/39770
> 
> Only one of the sighted is a feature (const) the other is semantic adjustment of
> a pre existing feature.
> 

Hear, Hear to all of you. I'd be satisfied to even let the const issue 
slide (assuming 'const by default' is out of the question) because it 
looks to me what Walter has in mind (*) could even be a v1.1 solution 
and not break much if any code.

Nailing down the the visibility and accessibility rules is obviously a 
must for v1.0 though.

* My earlier post actually looks very doable, is consistent with the 
language and reference implementation and actually meets Walter's 
criteria of being meaningful to the compiler:

http://www.digitalmars.com/drn-bin/wwwnews?digitalmars.D/39757

Don't know what to do about a 'const' return modifier though.