Appropriate location for list of language suggestions
This email is in response to the (ongoing) Haskell-cafe discussion about type declaration syntax. There are a number of hypothetical changes to Haskell that could be aesthetically pleasing or convenient if implemented, but not to such a degree that the added convenience outweighs the added complexity of a pragma-gated feature or the cost of breaking existing code. (Case in point: renaming type/newtype/data to be less confusing to first-time users) However, there may be better opportunities to implement such changes in the future, and it would be a shame if every good (but perhaps somewhat trivial) idea was lost to the annals of the mailing list. Therefore, I proposed that we might maintain a list of nitpicks, which could be used to inform future decisions about breaking changes to the Haskell specification. I'm not sure what the most appropriate forum is for such a list. The mailing list is sub-optimal, because we might end up with a huge email chain littered with bikeshedding reply-alls. I would be willing to host and maintain such a list on Github, but that seems somewhat antithetical to the Haskell community's generally more distributed style of discussion. The wiki might be appropriate, but I don't know if it's conducive to proper moderation; we may have to filter out suggestions like "add for loops". --Will
I think that a wiki (either the one at wiki.haskell.org or the one at https://ghc.haskell.org/trac/ghc/, but probably the first is more general) is a great place for such a list. Any such list will be treated critically by any consumers (that is, future implementors/language designers), so I'm not worried about silly suggestions like for loops. Additionally, if it's started with suggestions containing links to mailing-list discussions advocating for a design, then a consumer can see which suggestions have (at least some) buy-in from the community.
Richard
On Aug 9, 2015, at 7:49 PM, Will Yager
This email is in response to the (ongoing) Haskell-cafe discussion about type declaration syntax.
There are a number of hypothetical changes to Haskell that could be aesthetically pleasing or convenient if implemented, but not to such a degree that the added convenience outweighs the added complexity of a pragma-gated feature or the cost of breaking existing code.
(Case in point: renaming type/newtype/data to be less confusing to first-time users)
However, there may be better opportunities to implement such changes in the future, and it would be a shame if every good (but perhaps somewhat trivial) idea was lost to the annals of the mailing list.
Therefore, I proposed that we might maintain a list of nitpicks, which could be used to inform future decisions about breaking changes to the Haskell specification.
I'm not sure what the most appropriate forum is for such a list. The mailing list is sub-optimal, because we might end up with a huge email chain littered with bikeshedding reply-alls. I would be willing to host and maintain such a list on Github, but that seems somewhat antithetical to the Haskell community's generally more distributed style of discussion.
The wiki might be appropriate, but I don't know if it's conducive to proper moderation; we may have to filter out suggestions like "add for loops".
--Will
_______________________________________________ Haskell-Cafe mailing list Haskell-Cafe@haskell.org http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
As i have said in the other thread, i like the idea of keeping a list of hypothetical improvements to Haskell syntax for whatever purpose it can serve. I want just to add that IMO dissatisfaction with a programming language's syntax is not a negligible motivation in creating new languages: see [CoffeeScript](http://coffeescript.org) and [MoonScript](http://moonscript.org) for example. Alexey. On Monday, August 10, 2015 at 2:49:28 AM UTC+3, Will Yager wrote:
This email is in response to the (ongoing) Haskell-cafe discussion about type declaration syntax.
There are a number of hypothetical changes to Haskell that could be aesthetically pleasing or convenient if implemented, but not to such a degree that the added convenience outweighs the added complexity of a pragma-gated feature or the cost of breaking existing code.
(Case in point: renaming type/newtype/data to be less confusing to first-time users)
However, there may be better opportunities to implement such changes in the future, and it would be a shame if every good (but perhaps somewhat trivial) idea was lost to the annals of the mailing list.
Therefore, I proposed that we might maintain a list of nitpicks, which could be used to inform future decisions about breaking changes to the Haskell specification.
I'm not sure what the most appropriate forum is for such a list. The mailing list is sub-optimal, because we might end up with a huge email chain littered with bikeshedding reply-alls. I would be willing to host and maintain such a list on Github, but that seems somewhat antithetical to the Haskell community's generally more distributed style of discussion.
The wiki might be appropriate, but I don't know if it's conducive to proper moderation; we may have to filter out suggestions like "add for loops".
--Will
_______________________________________________ Haskell-Cafe mailing list Haskel...@haskell.org javascript: http://mail.haskell.org/cgi-bin/mailman/listinfo/haskell-cafe
participants (3)
-
Alexey Muranov -
Richard Eisenberg -
Will Yager