Nowadays the wording "Semantic Web" is rapidly becoming as common and widespread as the Web itself, and lot people are working hard to make a success of it. Back in 1998, however, the situation was much different. The Semantic Web was just a cool wording for few people, mainly fostered by some discussion on the blackboard (in fact, a whiteboard) in Tim Berners-Lee's office at MIT. The Resource Description Framework (RDF), the first basic brick, was still under development, and yet to become a standard (it become such on February 1999). However, Tim's vision was a grand one, and as you can imagine, he quickly managed to contagiously spread enthusiasm about his idea. Especially, oh boy, with those sketches on a whiteboard.
RDF was just meant to become the basic brick of the Semantic Web architecture, and therefore, there was the need to showcase the "upper layers", where the semantics could really enter in the game and make useful things: reasoning, query, logic. Therefore, it's been in those early times that the Metalog project started, as the first demonstration that the Semantic Web architecture was a feasible, and promising, one. Metalog was supposed to fill in the "logic" slot on top of RDF, by introducing reasoning and query capabilities. Moreover, it was also supposed to demonstrate how the Semantic Web did not need to necessarily be a complex thing (in fact, showcasing a vision usually is much more successful if normal people, and not just geeks, can understand it...): from this, the development of the PNL interface took place.
After some internal development, Metalog was publicly released to the world on July 16, 1998, together with an initial partial prototype.
Later, other public releases, refining the system, have occurred on August 28, 1998, on October 7, 1998 and on November 24, 1998.
On December 3, 1998, Metalog was presented with the paper Query + Metadata + Logic = Metalog, at the W3C Query Languages 98 (QL'98) event, held in Cambridge, USA (incidentally, if you are reading this, please have a look at the whole series of papers presented at QL'98: that has been a seminal event, and still at these times, that collection constitutes a unique blend of visions, perspectives, works and ideas). After the workshop, some serious development about the future of query systems within W3C was in order, and the choice had to be done. As RDF was going to be completed, and opening an official effort on the "upper layers" of the Semantic Web seemed too premature at the time (remember even the same wording "Semantic Web" was almost unknown at the time...), work on reasoning and query for the Semantic Web was postponed. Instead, another related major work started its activity, always fostered by the huge discussion at QL'98: the XML-Query project, deemed to provide a standard for querying (and doing more elaborate transformations and manipulation) XML. Some stabilization of Metalog appeared with the subsequent papers Let's Reason on the Web: Metalog (February 1999), and finally Towards the Semantic Web: Metalog (May 1999). By that time, the XML-Query effort had taken momentum, while work on the reasoning/query/logical aspects of RDF still seemed premature, and so the Metalog effort stopped.
Nowadays, needless to say, the Semantic Web has come back to a new life, and what were only jargon, animated chats and sketches on a whiteboard now seem much more than a vision. New intermediate layers of the Semantic Web have taken place: RDF Schema, and OWL, and the same roots of RDF have undertaken a major revision and "clean up". Therefore, the times are now ready again to start talking about the next upper layers, the reasoning/query/logic. Consequently, the Metalog project has resumed on 2003, and shipped a new complete implementation, v2.0b (check the latest version). The "upper layers" game is starting again, and this time it isn't going to stop... isn't it an exciting World (Wide Web)?