Cart
[Home] [By Thread] [By Date] [Recent Entries]
> > Right, but if you view attributes at "attributes of a type", and > > content as "attributes of a type" (syntactic sugar), we get into > > a funky world where one asks why you can't specify the ordering > > of attributes as you can content. This is kind of where the > > SML folk were coming from. > > The question of "order" is separate from that of a "type". In relational > databases, each returned row can be seen as an instance of a type (a la C. > J. Date), but the components of that type form a set, not a list. I think ordering is just one more type constraint. I'm using a non-standard definition of "type" though... where type is proven by a set of assertions... kind of like Is it red? Yes. Does it have seeds? Yes. Is it sweet? Yes. Is it crispy when bitten into? Yes. ... I declare this to be an apple!
|
