I have decided to use a different SGML parsing package than the Amsterdam Parser (ASP). Instead, I will be using the sp system created by James Clark. SP is an updated version of Clark's widely used sgmls package. The version of sp that I have downloaded is from 1996.
With this new software I have been successful in parsing two documents based on two different SGML DTDs (an HTML 2.0 document and a document based on the ISO 'article' DTD.) For example, I parsed the HTML document you are now reading:
The documentation for sp is very cryptic. I will be spending much more time in creating good documentation than I would have if I had stayed with ASP. I think, however, that once useful help pages are available sp will be as easy to use as ASP. One very nice thing about sp is that it does not need to generate a parser instance - everything is done at runtime. Users will therefore not have to worry about whether they have adequate disk quota to be able to use the parser.
I have compiled sp under Digital Unix 3.2C. Thus far, I have been unsuccessful
in compiling it under Ultrix 4.4. The three executables generated by building sp -
nsgmls, spam, and sgmlnorm are installed on the machines
raven and penguin.
There are also several DTDs I have found and installed in /usr/local/lib/sgml/
on raven and penguin. I have not
tried using each DTD yet, but plan on doing so to make sure they work correctly under sp.