backward compatibility

[Axel Simon]

On 21.01.2011, at 03:12, Ian Lynagh wrote:

> On Thu, Jan 20, 2011 at 09:22:37PM +0100, Axel Simon wrote:
>>
>> In the case of the layout "bug", I think it might be worth  
>> considering
>> going the other way: adjusting the standard with what ghc has always
>> done.
>
> Anyone can propose language changes - the process is described here:
>    http://hackage.haskell.org/trac/haskell-prime/wiki/Process
>
>> I therefore think that keeping the number of extensions
>> to a minimum should be a high priority. It seems that the ghc team is
>> going overboard with the amount of extensions and their granularity  
>> that
>> I do not believe that there will ever be another compiler since
>> implementing all these extensions is a nightmare. The road of may
>> extensions is leading down the road that the Haskell standards  
>> aimed to
>> avoid: having a single implementation defining what a Haskell  
>> program can
>> be.
>
> I'm not sure if you're saying there should be fewer new language
> features implemented, less fine-grained control over which are  
> enabled,
> or something else?
>

I agree that new language features is required to make Haskell a  
research platform. So it would be awkward to try to stop people from  
adding new language features. I don't mind if highly experimental  
language features come in many varieties (i.e. with fine-grained  
control about what is allowed).

I'm more concerned about standard extensions that many (even most)  
people use somewhere in their projects. I get the feeling that these  
get split up into too many individual language features which will  
make it difficult for other compilers to implement the extension and  
the switched-off extension correctly. Haskell 2010 is going into the  
right direction in making some of these features mandatory which means  
that they become language features that you can't switch them off  
anymore. This makes it more likely that future compilers can implement  
a comprehensive Haskell language. So yes, less fine-grained control  
for features people use often.

The layout rule seems to be one extension that can go into the report  
(both 98 and 2010) without breaking any program. If this is really so,  
then I think it is easiest for compiler writer and users if such a  
change is made to the report. I think only very few language features  
qualify for going directly into the reports since most of them can  
break programs.

Hope this clarifies my opinion,

Cheers,
Axel

>
> Thanks
> Ian
>
>
> _______________________________________________
> Glasgow-haskell-users mailing list
> Glasgow-haskell-users at haskell.org
> http://www.haskell.org/mailman/listinfo/glasgow-haskell-users