X-Logic was first devised as an alternative to the W3C approach to realising
the idea of "The Semantic Web", a way of reconstruing data and computation
distributed over the internet as propositions and inference.

It was manifest at that time (2000) as a small formal model which exhibited
the following characteristics:

\begin{enumerate}
\item  The premises and conclusions of inferences viewed at this level
   were to be publicly accessible structures on an open network,
   rather than in a theory heirarchy private to the use of one
   proof tool by one user.

\item  The notion of proof was liberalised to encompass by construing the
   evaluation of a program meeting some specification as conducting
   an inference determined by the specification.

\item  A system of trust or assurance marking was envisaged, which would
   protect users requiring high assurance of their results from contamination
   by the involvement of decisision procedures or other computational
   inferences which may not be sufficiently trustworthy.
\end{enumerate}

This system lay fallow for a decade, and then acquired a new significance
for me which brings it once again central to my concerns, and has caused me
to begin work on a new version.

The first stage in this work is a slight augmentation of the trust/assurance
marking system, and a fuller separation of this from other aspects of the system.
The hard core of the talk will be a presentation of this system.

The second, much more substantial development will also be trust related,
and appears through the concept of "epistemic retreat".
The whole is intended to supply a very general methodological framework intended
as a successor to the efforts of Rudolf Carnap on formal analytic methods for
philosophy and science (and perhaps more importantly, engineering).