{Didier Verna}
{A large scale format compliance checker for \TeX\ font metrics}
{As part of \acro{ETAP}, our experimental typesetting algorithms
platform, we have developed a parser for TeX Font Metrics files, simply
called \acro{TFM}. A robust parser for an offical data format must
handle all sorts of compliance problems of various degrees of
seriousness, ranging from simple warnings to non-recoverable errors.
\acro{TFM} not only provides a rich (hopefully exhaustive) ontology of
errors, but also a powerful recovery mechanism, allowing to go as far as
possible in the parsing, for example by fixing errors on the fly or
discarding problematic input.

As a side-effect, it is possible to use \acro{TFM} as a validation tool
rather than for actually loading font information. Indeed, the
\acro{TFM} exception handler reifies the problematic situations into
objects (in the ``object-Oriented'' sense) which can be silently
collected until the parsing is over or needs to be terminated
prematurely. These objects can in turn be used to produce a full
compliance report for the analyzed file. We have automated this process
for the whole \TeX\ Live distribution, resulting in the (in)validation of
almost 80,000 fonts.}