I think the message is to convey "relatedness" with multiple location SETFs. That may be over stating it a bit but the more important message is "go for it but indent correctly (I was just being lazy)". I am implementing code written in pseudo language so it is natural to enter multiple setf's since I am translating. But studying the code I found I didn't like so many SETFs cluttering up my reading process. But I was concerned that if the code had to be rewritten, then it might be easier to keep assignments isolated. I think the measure I will use is relatedness, however arbitrarily I define such a thing.
I am trying to maintain linkage with CLRS so I will stick to their formulas and not rename variables. Thanks for the feedback! I am going to take liberty with CLRS and cut back the SETFs but keep their overall formula style. ----- Original Message ----- From: "Rob MacLachlan" <[EMAIL PROTECTED]> To: "Pat Monardo" <[EMAIL PROTECTED]> Cc: <[EMAIL PROTECTED]> Sent: Thursday, December 12, 2002 6:48 AM Subject: Re: SETF style > > Well, I've perpetrated multi-location setf in numerous places in CMU CL, > and don't have any problem with it. What sort of maintenance problem were > you thinking of? > > Rob > >
