Added tests for schema - and all pass.

docs/codox/AgainstTruth.html

@@ -9,7 +9,6 @@
 <h2><a href="#introduction" name="introduction"></a>Introduction</h2>
 <p>This document is in two parts: a statement of a problem, and an account of an attempt to address it. The problem is stated briefly in the first chapter, and fleshed out in the following two with a history of attempts which have been made in the past to address it, and an analysis of what would be needed to solve it.</p>
 <p>The second part starts with an account of a system built by the author in collaboration with Peter Mott, describing particularly how the problem was addressed by this system; subsequent chapters will describe the development of a further system, in which the analysis developed in the first section will be applied.</p>
-<p>This document deals only with explanation. Issues relating to inference and especially to truth maintenance will undoubtedly be raised as it progresses, but such hares will resolutely not be followed.</p>
 <h2><a href="#note-on-the-quality-of-the-text" name="note-on-the-quality-of-the-text"></a>Note on the quality of the text</h2>
 <p>Much of this text was written between 1986 and 1988 on Xerox 1108 and 1186 workstations, in their native WYSIWYG document system, and printed as hard copy; and some was written on the very first generation of Apple Macintosh computer, and again printed as hard copy. The text here is the consequence of scanning the hard copy and running optical character recognition on the scans. It isn’t perfect. I am proof reading as I go and I hope that it will improve.</p>
 <h2><a href="#contents" name="contents"></a>Contents</h2>

docs/codox/Analysis.html

@@ -229,9 +229,7 @@ a is x
 <p>The use of ‘correct’ here is the problem. It appears to imply that one knows that the answer which one knows maps in some unproblematic way onto something real in the external world. Thus Achinstein, too, falls into the trap sprung by the radicals. In the absence of some definition of what is to be understood by ‘correct’, these definitions are simply meaningless.</p>
 <p>This account does not address many of the facets of common sense explanation.</p>
 <p>It has nothing to say about the amount of detail contained in an explanation. It has nothing to say about the need to express an explanation in terms of an account of the world which is accessible to the auditor. lt fails to account for the possibility of explanation by analogy, or of unintentional explanation.</p>
-<p>It appears that Achinstein’s motivation in producing a new account has less to do with addressing these real world problems than with overcoming such philosophical puzzles as the Paradox of the Ravens; so his account takes us no nearer to providing a model which will support the construction of better common sense explanations. 5</p>
-<p>A [ mh L/Go/w%` \</p>
-<p>` Eegozi</p>
+<p>It appears that Achinstein’s motivation in producing a new account has less to do with addressing these real world problems than with overcoming such philosophical puzzles as the Paradox of the Ravens; so his account takes us no nearer to providing a model which will support the construction of better common sense explanations.</p>
 <h2><a href="#toulmin" name="toulmin"></a>Toulmin</h2>
 <p>Turning aside for the moment from those philosophers who have made a study of explanation per se, there is another whose work is currently attracting fashionable attention in Expert Systems (and especially mechanised explanation) circles; one must make passing reference to Toulmin, and, unless one chooses to use his argument schema, one must produce a very good argument for not doing so. Therfore it is important to know what he actually argued.</p>
 <p>A Toulmin’s programme was to replace the syllogistic with a new logical schema of his own. He did not address symbolic logic except by dismissing it as irrelevant to the study of real world arguments:</p>