> It just makes me wonder what the hell they thought they were doing all that time... The YAML project was a convergence of several different efforts at inform

> It just makes me wonder what the hell they thought they were doing all that time... The YAML project was a convergence of several different efforts at information representation including people from Perl, Python and Ruby, each with our own ideas. I happened to be involved in the outer ring of the XML community, in particular a group SML-DEV where we were looking for a better information model more suitable to data serialization that would use an XML compatible syntax. At that time, especially since serializing data with XML was all the rage, "ML" or "Markup Language" was commonly associated with data serialization. In fact XML is very inconvenient for actual markup, even though it derives from SGML (of which HTML is an example). The "YA" part did deliberately come from YACC, the reason why is that XML (and we hoped YAML) would be the basis of domain specific serialization languages. Hence, you can think of it as a meta-language for building sub-languages, like the application I was working on, a serialization for accounting data. Hence, that's the origin of YAML acronym, which happened to have a domain name open. In fact, you can look at archive.org starting in 2001 and you'll see the 1st public pass of YAML more closely followed XML like bracket syntax. I disliked it, but, it was what came out of the SML-DEV collaborations. Even so, the important thing was the information model, which was a simple typed graph and not an element tree. The syntax evolved with many revisions after the model and goal were set, with lots of feedback to address concerns and usability tests with domain experts. The syntax got more and more lightweight, inspired from RFC0822 (email) while adding dashes for list items. Testing the syntax with domain experts, e.g. accountants and the like, was exceptionally important part of the process. Users liked this serialization syntax since it made their data "pop". So. A year or so passes while we focus on getting things to work and helping people. Then, the product differentiation question comes up. Since XML had the dominant position in the data serialization mind share, how does YAML compare? Well, the YAML model was a typed graph, while XML was an element tree. One required no special libraries to manipulate, the other require a DOM to translate. But perhaps more importantly, it's because XML borrowed its model and tags from SGML and SGML was a true "markup language". So, XML was a "markup language" where it was impractical to do actual markup. Then, it dawned on us, well, of course a data serialization language isn't markup problem. I'm not sure any of this was obvious at the time. Anyway, in a very fun chat, Oren pointed this out and then Ingy said: well then, YAML Ain't Markup Language! So, the new name actually represented what we had set out to do in the first place. Further, I would suggest that the industry understanding of how serialization languages are poorly supported by markup approaches (XML) is at least somewhat due to our name change and fun filled articulation at conferences.