- From: Joe English <jenglish@crl.com>
- Date: Mon, 19 May 1997 18:23:22 -0700
- To: w3c-sgml-wg@w3.org
Andrew Layman <andrewl@microsoft.com> wrote: > Several writers have recently suggested that we can do without > structured attributes by allowing each application to hide some > structure inside what looks to the XML parser like an unstructured > attribute, perhaps using a LISP-like syntax. Others have also suggested -- and I agree -- that we can do without structured attributes because XML already supports structured elements. If you need to represent some thing in XML, and that thing has parts, then it is best encoded as an element. If any of the thing's parts have parts themselves, they can be encoded as sub-elements. Attributes can be used to encode the parts that don't have any (important) sub-parts. > Others have suggested that > we can greatly reduce the size of documents by using > comma-separated-values in element contents, rather than explicit tags > within. This is foolishness. I agree. Terseness is explicitly *not* one of XML's design goals. > The whole point of XML is to have an > infinitely-extensible, self-describing syntax that allows the structure > of a document to be determined by a very simple, regular parser. > Pointing out that structural limitations in public XML can be solved by > private hacks just gets us farther away from this goal. --Joe English jenglish@crl.com
Received on Monday, 19 May 1997 21:26:06 UTC