4-24-03 New Thinking in Patterns, HTML conversion

The changes are quite significant, since the book has been completely restructured and most of the examples have been refactored significantly. There have been numerous new examples and sections, and new exercises. I've also added the "changes" section at the very beginning of the document so you can see in more detail what's changed.

In the next seminar (in Prague; there are still spaces available), I will use all the new and refactored material, plus Bill will be researching new, non-GoF patterns and finding ones that are particularly interesting. He will present these throughout the seminar in his presentation/discussion format, to add additional mental stimulus to the course. If you have particular favorite non-GoF patterns, please tell Bill about them.

One thing -- the prose has not been changed much, nor rewritten as it needs to be. But I felt it was worthwhile to get the code up.

Also, the code is still built using makefiles. We need to modify the tool used to create build.xml files in order to convert to Ant.

Years ago (that is, several versions ago) I tried to use Microsoft Word's "export to HTML" feature, and found it woefully lacking. Today I was thrashing around trying to bring my build files up to date for the patterns book and not having a very good time of it. In desperation, I tried Word XP's "save to html" and discovered that it had changed dramatically. For one thing, many of the features that I remember, concerned with fancying up the output, had been removed. Instead they seemed to refocus on just getting decent HTML out.

There were two flavors of export, including the basic html version which was very fat and apparently maintained extra information that related to Office (possibly so that the document could be recovered into Word). This looked better in Mozilla 1.3a than in IE6. The second version was "filtered" which apparently removes the extra Office information and dramatically slims down the results. Still looks better in Mozilla than IE6 (but tolerable in IE). It automatically converts the diagrams to GIF files so they show up in the document looking quite nice. In general, I feel like this will provide a decent quick export (I never completely got the Logictran build system going for TIJ3 -- almost, but I got stuck on the Cascading Style Sheet. The results of the Logictran approach are much fancier and nicer, but it's good to be able to get something out with minimal effort).

There's also something called a "web archive" with a ".mht" extension that Word is willing to save into. This is a single monolithic file, which is significantly larger than the html file, but Mozilla will also read it and do a decent job of formatting; better than IE which still messes up any lines that contain footnote references. Mozilla 1.3a doesn't get the footnote links correct and gives an error message if you click on them, but at least the lines are all readable.

I tried OpenOffice Writer 1.0.2 but its save-to-html was miserable (I'll hope for better with the upcoming 1.1). So oddly, the best results are achieved by writing from Word XP to "filtered" html, and reading the document with Mozilla.