The discussion over proposals for a simplified markup language -- essentially a pared-down XML -- is proving far from simple itself. The latest issues discussed include internationalization, attributes and a challenge over use-cases.

A row over English-only element names was sparked by Don Park's publication of a proposed EBNF grammar for SML. As noted by David Carlisle, the EBNF only allowed for ASCII alphanumeric characters as tag names. Park responded that the limitation was a suggestion to test opinion.

In another thread of the discussion Rick Jelliffe, a staunch defender of XML's sufficiency, cast down the gauntlet to find use-cases where XML failed and SML was required. Seeing as a lot of discussion is revolving around small-memory devices such as cell phones, and nobody in the discussion seems to know how much memory such a device actually possesses, it will be interesting to see if any failure cases of XML are proposed.

Following the publication in XML.com of an article about SML, renewed debate ensued over the exclusion of attributes from SML. Don Park opened the issue for debate, and Rick Jelliffe responded by saying that attributes make programming convenient:

It seems that most XML programs are written so that the elements "push" programs but attributes values are "pulled" in. So the programmer does not need to concern themselves with attributes that are redundant to their task. To make everything an element would reduce the choices available to a programmer.

Paul Tchistopolskii disagreed with Jelliffe, claiming that the convenience is there for writing the document, but complicates the processing of a document. In another post, as an aside, he urges that S-XSLT (simplified XSLT) should be written. One can only hope that debate is saved for later!

• Don Park has posted a second cut of the EBNF for SML, taking into account comments, and as a provocation for further comment.