After suggesting that UIDs should be attached to elements to leave the choice of their names and language up to application designers, the debate on ebXML mailing lists has expanded to cover many points often discussed on XML forums -- including the need for universal "names", the readability of XML documents, the need to translate element and attribute names, and the usage of namespaces URIs.

The language, character encoding and meaning of XML tag names is an issue that is often debated and is especially important for ebXML where the charter and affiliation to the United Nations gives an obligation to be universal and understandable by everyone on the planet.

To work around the problem of the tag names, one proposal is to add an attribute referring to a Universal IDentifier (UID) to each of the elements -- and eventually to define this attribute as FIXED in the DTD for the document.

This solution would address the issue by defining both a user readable name and an identifier, but the debate expanded after William Kammerer challenged the suggestion of reducing tag names to opaque UIDs, and questioned the value of multilingual tag names:

Instead of fantasy examples (as if the Thai merchant really wants to jack around with XML - he or she will probably want to use affordable shrink-wrapped software), perhaps Duane and David can demonstrate how these Bizcodes or UIDs could be used to make RosettaNet (or OTA or S.W.I.F.T. XML) messages understandable within the context of ebXML Core Components.

Hence, we don't yet have to rule out basing tags on a limited vocabulary like that held in the Basic Semantics Register. This avoids the inconvenience of the indirection required by the unintelligent UIDs and BizCodes. By unintelligent, I don't mean Duane's proposal is stupid, but merely that the UID by itself gives me no hints as to meaning.

The lack of meaning carried by these UIDs can also be seen as an advantage, says Irvin Chmielewski:

One of the main problems we found in the past with Natural Language is that word can have multiple meanings. Take the word STOCK. - Something you buy on the NYSE or something you can buy at a cattle ranch or something you can buy at a gun store or something that you can buy to make soup.

And Duane Nickull explains how namespaces are different from UIDs:

An XML Namespace does not tell a parser anything useful either.

An XML Namespace does not provide any semantics about an element.

The only thing an XML namespace value does is allow a code writer to build a handler routine which can tell an application that the domain of an element is different from another element based on the fact that the namespace attribute is unique (ie - has the domain url in it). There are no rules saying that urls have to point at anything which provides semantic rules.

UID's, as contemplated in ebXML, will have that mandate.

At the end of the day, does it mean that the only universal language would be a string of digits?