Article:
 |
|
Higher-Order Messages in Cocoa
|
| Subject: |
|
A couple of notes |
| Date: |
|
2004-07-19 11:26:58 |
| From: |
|
mweiher
|
|
|
|
Great to see this idea being promoted without having to do it myself! :-))
However, I should point out that Higher Order Messaging and blocks (or ActionExpressions) are distinct mechanisms. Blocks are not an implementation of Higher Order Messaging.
In fact, HOM was specifically developed as an alternative to blocks/closures/higher-order-functions in a language where blocks aren't available, and the name was chosen to make clear that this is a higher-order mechanism based on messaging instead of (anonymous) functions.
A presentation held at MacHack a while back gives some more background info, though it is admittedly a bit terse: http://www.metaobject.com/papers/HOM-Presentation.pdf.
Furthermore, it does seem a bit odd that the original and definitive implementation of HOM, MPWFoundation, isn't mentioned.
Marcel
|
Showing messages 1 through 1 of 1.
-
A couple of notes
2004-07-21 18:53:02
iapole
[View]
Holy _cow_ am I ever sorry. To make it clear, I used MPWFoundation as a reference for my own trampolines implementation back in the day, and it was in fact your work that inspired me to come up with this. And I remembered that and planned to write about it but it slipped my mind. My deepest apologies Marcel, you deserve full credit for this.
As to Blocks/HOM, it may be my more recent experience with languages such as Io which blurred the line for me. You're perfectly right, but I've always seen Blocks as being inline methods, so it made sense to me.
Thank you very much for the correction and your patience! Your work on MPWFoundation opened a lot of conceptual doors for me.