My preference for feature removal starts with making the feature optional.

Once you can turn it off, it's relatively easy to test for what that breaks.

Once most of the breakage is fixed, you can release with it turned off by
default.

(The migration from x./y. to x/y for verb parameters skipped directly to
"turned off by default" which I think was a mistake. But that's old news
now.)

That said - do you happen to know where the dictionary documents that 2
:'u,v' and 2 :'x,y' behave the same?

Thanks,

-- 
Raul

On Wed, Oct 8, 2014 at 7:47 PM, Henry Rich <[email protected]> wrote:

> It's just backwards compatibility, but a lot of the standard scripts rely
> on it, and someone would have to do a bit of work to allow the feature to
> be removed.
>
> Henry Rich
>
> On 10/8/2014 6:20 PM, Raul Miller wrote:
>
>> I think you want http://www.jsoftware.com/help/dictionary/d310n.htm
>>
>> Especially points 4 and 5.
>>
>> But I actually haven't found anything that suggests that x and y are
>> analogs for u and v when u and v are not used. I think that might be pure
>> backwards compatibility (which suggests it should be treated in am manner
>> analogous to 9!:48 and 9!:49).
>>
>> Thanks,
>>
>>  ----------------------------------------------------------------------
> For information about J forums see http://www.jsoftware.com/forums.htm
>
----------------------------------------------------------------------
For information about J forums see http://www.jsoftware.com/forums.htm

Reply via email to