Comments for sykosomatic

Comment on Pivot tables in PostgreSQL by sykopomp

sykopomp — Wed, 04 Jan 2012 23:23:26 +0000

Yes, I already mentioned crosstab() in the article. Second to last paragraph. :)

Comment on Pivot tables in PostgreSQL by Henry

Henry — Wed, 04 Jan 2012 22:05:50 +0000

see crosstab on http://www.postgresql.org/docs/current/static/tablefunc.html

Comment on Chillax and Protocols by Michael J. Forster

Michael J. Forster — Sat, 06 Nov 2010 19:34:02 +0000

As Alan Kay said, it’s really about the messages, not the objects.

http://www.purl.org/stefan_ram/pub/doc_kay_oop_en

Comment on Chillax and Protocols by Vladimir Sedach

Vladimir Sedach — Fri, 05 Nov 2010 04:34:34 +0000

I actually just got around to reading Keene’s book last week. One thing she recommends that’s not immediately obvious (at least it wasn’t to me) is that “object” construction should be done with explicitly defined constructor functions instead of make-instance (so instead of (make-instance ‘standard-server) you’d have a make-server (generic) function). Then you have object creation as part of your protocol, and can swap out implementations (for example make the hypothetical “make-server” return those hash-tables instead of an actual CLOS “object”).

Comment on Chillax and Protocols by Christophe Rhodes

Christophe Rhodes — Thu, 04 Nov 2010 21:50:45 +0000

I have nothing much to add to the post (other than I agree with you!), but I have to say I smile widely at the treatment that your blog engine gives to “:DEFAULT-INITARGS” :-)

Comment on Chillax and Protocols by Andreas Fuchs

Andreas Fuchs — Thu, 04 Nov 2010 04:24:15 +0000

Excellent post. This is pretty much how I tried to structure autobench, and it has made my life so much easier.

Comment on Chillax and Protocols by Thomas M. Hermann

Thomas M. Hermann — Thu, 04 Nov 2010 01:47:21 +0000

Small critique, don’t mix :INITARG with :INITFORM. If you want the user to be able to be initialize the slot with an :INITARG but default to some value, use :DEFAULT-INITARGS. Otherwise, use :INITFORM if you want the slot to have an initial value. Most of the time, this is not an issue, but becomes a problem when you start getting into situations where you are directly using INITIALIZE-INSTANCE and REINITIALIZE-INSTANCE. See Chapter 9 of Keene for a detailed discussion of why these slot options conflict.

Comment on Chillax and Protocols by Brian T. Rice

Brian T. Rice — Wed, 03 Nov 2010 22:38:59 +0000

Echoing sykopomp, I agree that the protocol focus over class structure up front is a big feature of Smalltalk style design. In Slate, my dialect with multi-methods, code definitely looks like this, without generics or prefixing typenames to every method name.

Comment on Chillax and Protocols by sykopomp

sykopomp — Wed, 03 Nov 2010 21:13:53 +0000

Besides the multiple dispatch? I’m not sure it’s very different as long as you keep in mind that the protocol (the messages you expect will be there) is what’s important, not what the objects ‘are’.

Comment on Chillax and Protocols by Duncan Mak

Duncan Mak — Wed, 03 Nov 2010 21:10:42 +0000

I think this is a great way to structure code too, but I have to ask:

How is this different from Smalltalk?