On 2/21/22 4:44 PM, Bradley Lucier wrote:
> On 2/20/22 9:31 PM, Arthur A. Gleckler wrote:
>> On Sun, Feb 20, 2022 at 12:55 PM Bradley Lucier
>> <xxxxxx@math.purdue.edu <mailto:xxxxxx@math.purdue.edu>> wrote:
>>
>>     The first two arguments are required, but I'm beginning to think that
>>     the short required argument, the dimension, should come before the
>> data
>>     itself, which can be quite large.
>>
>>
>> [Editor's hat off.]  That sounds smart to me.  Most of the time, the
>> data won't be literal, but sometimes it is, and I don't see any reason
>> to choose the other order.
>
> Thinking about it again, because a list/vector/list*/vector* is being
> converted to an array, with some auxiliary information, maybe it is a
> good idea to have an argument-list convention of
>
> (define (list->array    list    domain-interval ...))
> (define (list*->array   list*   dimension ...))
> (define (vector->array  vector  domain-interval ...))
> (define (vector*->array vector* dimension ...))
>
> Brad

Based on your comments I made the argument order

(define (list->array    interval  list    ...))
(define (list*->array   dimension list*   ...))
(define (vector->array  interval  vector  ...))
(define (vector*->array dimension vector* ...))

Of these routines, only list->array appears in SRFI 179, so this is an
incompatible change for that procedure.

Brad