Re: Naming inconsistency

> James Clark <> writes:

> > Schemas part 2 provides three elements for deriving new types:
> > - list
> > - union
> > - restriction
> > Each of these elements has an attribute specifying the type or types
> > from which the new type is derived:
> > - "list" has a "itemType" attribute
> > - "union" has a "memberTypes" attribute
> > - "restriction" has a "base" attribute
> > Surely it would be more consistent for the attribute on "restriction" to
> > be "baseType" rather than "base".

> Yes in principle, but in practice we are used to it by now, and didn't 
> think anything would be gained by a change at this late date.

> (personal opinion, the WG hasn't discussed this)

> ht

"Used to it"?  Given that a lot more people will ultimately be using XML
Schema than are now "used to it" (and a lot more applications will be
written...), I would think it best to "improve the user interface" where
possible by making small changes like this.  The point has been made
several times that a more readily learnable specification IS something
worth striving for, particularly when no penalty in expressive power is
incurred.  Anyway, this change would certainly be less disturbing to
current practitioners than, for instance, the large number of
alterations, shiftings, and additions/deletions in the list of built-in
datatypes that occurred between CR and PR.


Received on Monday, 19 March 2001 20:45:10 UTC