Flatted/inline slots

[aartaka]

So we mostly stick to the overall structure of defclass with our define-class, which is cool. but then there's an extra layer of parentheses to slot definitions we have:

(define-class name ()
  ((slot-a 5)
   (slot-b 8)))

While this is consistent with historic form, it would be cool to drop this extra layer. Several problems with it:

• It becomes much harder to distinguish (slot-a 5) and (slot-a slot-b slot-c). Maybe even impossible.
• We lose an additional layer of indentation separating slots from options, which was visually useful.

I'm extremely uncertain about this, but if we make it deterministic enough, then why not have it?

[Ambrevar]

Another point:

• How to you distinguish options from slots? Could check if first element is a keyword or a symbol, but we should see what the standard allows here.

[aartaka]

• How to you distinguish options from slots? Could check if first element is a keyword or a symbol, but we should see what the standard allows here.

Keyword checking is safe, I guess. defclass doc says:

Slot-name--a symbol. The slot-name argument is a symbol that is syntactically valid for use as a variable name. 

And then, keywords cannot be variables, at least. Spinneret uses keywords as macro names, and you can defun a keyword function, which is not cool, but slot name... I'm not sure—keywords are not valid boundp variable names, at least (҂⌣̀_⌣́)

[aartaka]

Uh oh, that's bad. CCL and ECL both allow keywords as slot names.

[aartaka]

Closing as not-implementable-reliably then.