[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: structure type specifier
- To: Larry_Masinter.PARC@Xerox.COM
- Subject: Re: structure type specifier
- From: Robert W. Kerns <RWK@YUKON.SCRC.Symbolics.COM>
- Date: Fri, 11 Dec 87 15:56 EST
- Cc: cfry@OZ.AI.MIT.EDU, Common-Lisp@sail.stanford.EDU
- In-reply-to: <871208-162102-5563@Xerox>
Date: 8 Dec 87 16:20:16 PST (Tuesday)
Its legit and even good design practice to treat every data type as if it might
be a structure. If Xerox Common Lisp had STRUCTUREP, it would be true of every
object. (Of course, some objects, like fixnums, don't have slots and are
STRUCTURE (and STRUCTUREP) would just indicate whether they follow
the protocol of things you can to do **named structures**. (That is,
Inquiring about slots, acessing the slots found, and basically getting
your hands on the information in the original DEFSTRUCT). Since
fixna are neither defined nor definable as named structures, for them
to be STRUCTURE's would be a travesty.
I don't know just what YOU think the point of having such a type would
be if you're going to make Xerox Common Lisp have everything be a STRUCTURE.
Perhaps your point was really that some subtypes of COMMON may be implemented
as DEFSTRUCT's (for example, RANDOM-STATE is a good candidate).