What Is XML, or MicroXML?

On Sun, Nov 17, 2013 at 9:35 AM, Simon St.Laurent <simonstl@simonstl.com> wrote:

On 11/17/13 10:38 AM, David Lee wrote:

This is a good idea. �I've wondered for a while when the next phase transition would come along - first SGML -> XML, then XML -> ????.

I see some good hints in MicroXML and FtanML in particular, but it's probably time or past it. �I don't think there will be quite the same uptake. XML benefited simply from saying "these things are actually possible" at a time when that was still a surprise.

Simon, you won't be surprised to hear it, but for the benefit of others I do want to clarify that I *hope* MicroXML never has anything near the same uptake as XML.

XML was way too popular for its own good. I myself came to XML after a brief sojourn in SGML because i had recognized that the traditional data formats I was dealing with were hopeless in representing the semi-structured data that's all over the customer and consumer relations area (my domain specialization in the mid-90s). I immediately warmed to the document-centric approach. I heard a lot of folks saying "hey this stuff is great for data as well" and I didn't know enough to demur, so I shrugged and went along with that. It wasn't very long *at all* before I realized using XML for "pure data" was a bad, bad idea. It wasn't just the mess made in the likes of WDDX and XML-RPC, or the XML-as-CSV I was seeing crop up as the latest form of bloatware everywhere, but it was the fundamental realization that XML has been essentially optimized around natural language (character data) with a modest layer of annotations suitable for natural language, which is collectively text. Almost anything else was using a hammer to fasten in screws. Folks here on XML-DEV were very helpful in making clear details of this realization.

When Simon and co went off to create Common Core and SML, when Simon put a lot of work into seeing how ASN.1 could be better appreciated by the hammer-wielders, when PaulT painstakingly catalogued the alternatives to XML, especially for programming and DBMS, when XAML started to make waves, and when JSON eventually took off, many of us were cheering because we saw it as a way XML could be saved from the personality split of doc-heads versus data-heads. Instead we got Namespaces, PSVI, XDM etc. In a way data appeared to have colonized XML, but the planet was never really right for data, no matter how assiduous the terraforming.

MicroXML is pretty much just back to minding the XML 1.0 knitting, but with even fewer temptations for the hammer-wielding colonizers (e.g. we considered just ignoring XML Namespaces structured and decided in the end it was better to make it impossible to even write anything that looks like XML Namespaces). That means that we'll never have the sorts of participants who are so adept at pushing a technology up the Gartner Hype Cycle, and that's quite fine by me.

If some of the folks in recent threads instead want to have some XML-like all-singing, all-dancing axioms-all-the-way-down structured data representation system, I have no interest whatsoever in stopping them. I just won't accept their calling that system XML, because it's not. MicroXML is much, much closer to what XML really is, and yet we know better than to call it XML.

Uche Ogbuji � � � � � � � � � � � � � � � � � � � http://uche.ogbuji.net
Founding Partner, Zepheira � � � � � � � � �http://zepheira.com

Author, Ndewo, Colorado � � � � � � � � � ��http://uche.ogbuji.net/ndewo/
Founding editor, Kin Poetry Journal � � �http://wearekin.org
Editor & Contributor, TNB � ��http://www.thenervousbreakdown.com/author/uogbuji/
http://copia.ogbuji.net� ��http://www.linkedin.com/in/ucheogbuji� ��http://twitter.com/uogbuji