m`n is fully defined and cannot be changed without breaking old code.
Henry Rich On 10/15/2021 1:14 PM, 'Pascal Jasmin' via Programming wrote:
technically, 1`2 1 2 exists as a non-error though silly use unless trying to trigger error if an boxed paramerter were provided, which still seems like a silly use, considering both that such careful guarding would guard against so much more parameters, and , will also fail with boxed paired with unboxed parameter, and , provides the natural :: "deal with error" function, that can't as easily be done with ` errors. , behaves identically to ` with nouns. There's no real reason to use ` with nouns with its current definition. I'd suspect that any historical use would be for purely esoteric obfuscation motives. so if 9.03 is already committed to breaking things. I think reworking ` to produce ar s all the time (which is what "ti" is defined to do, as well as "atomic representation of unboxed nouns in place of those nouns." you refer), I think would provide a welcome specialization of ` for use in gerund forming. This would solve the following common errors as well f`0`]} to transform head (or other specific) index of an array. This form is allowed in } as long as nouns are properly gerundified. On Friday, October 15, 2021, 09:43:56 a.m. EDT, Raul Miller <[email protected]> wrote: Perhaps also worth noting that ` could have its definition updated to use the atomic representation of unboxed nouns in place of those nouns. That's not exactly the definition of your ti conjunction, but it also would not change the behavior of any current non-error case, and would address the common cases. (Also, boxing in general is going to require a little extra coding work because the abstraction is all about combining data structures which otherwise would not be combined.) Thanks,
-- This email has been checked for viruses by AVG. https://www.avg.com ---------------------------------------------------------------------- For information about J forums see http://www.jsoftware.com/forums.htm
