I'm kind of trying to figure out when I'd want that operation.

groupsWithSeparatorsWhereSeparatatorsMatch: feels like it says what it does

On Dec 12, 2017, at 10:19 AM, Ben Coman <btc@openinworld.com> wrote:



On 12 December 2017 at 20:03, Nicolas Cellier <nicolas.cellier.aka.nice@gmail.com> wrote:


2017-12-12 13:01 GMT+01:00 Henrik Sperre Johansen <henrik.s.johansen@veloxit.no>:
Ben Coman wrote
>
> But after pondering a while for a better name, I wonder what is wrong with
> the existing?
> Googling "define aggregate" provides...
>     aggregate (noun) = a whole formed by combining several separate
> elements.
>     aggregate (verb) = form or group into a class or cluster.
>
> The separate elements are runs defined by the block.
> The existing name seems quite precise to me.
>
> cheers -ben

When I read aggregateRuns:, I think the opposite; something combining
different runs into a single entity, not something that splits a single
collection into constituent runs...

okay. I can understand that perspective. 
 

Something like (split/collect)runsAccordingTo: sounds more descriptive to my
ears, at least.

Cheers,
Henry



I like groupBy: but in Squeak groupBy: produces a dictionary with unique keys ignoring the multiple sequences...
split already carries the meaning of preserving the sequences, but we have to tell to split at a change of value....
Run is also quite explicit...

groupByRuns: / splitRuns:

+1 
groupByRuns:   is good.

cheers -ben