I don't know if anyone here has noticed or cares... Our LSP project (at
lsp.everest.com) has been rather quiet of late. The primary reason is
that it was more important for me to get our product wrapped up. This
time away from LSP has given us the opportunity to re-evaluate where we
stand with the project, and have chosen a different (and much more
promising) long term direction for the project.
To recap, LSP (Lisp Server Pages) is a web page scripting language,
loosely in the spirit of JSP/PHP/what have you for generating lisp code
that can be executed in an HTTP server and generate dynamic HTML. Two of
the advantages that LSP offers is that it is structurally aware, and it
brings over some lisp semantics to the scripted HTML page.
We have come to the realization that LSP is in its current form rather
limited. In particular, although LSP obeys SGML semantics, it does not
give sufficient distinction between the document and the embedded lisp
code. Fixing this situation has prompted us to consider substantial
changes to the LSP infrastructure.
Specifically, we are considering moving LSP from an HTML base to an XML
base. Lisp is ideally suited for handling dynamic content, and XML
provides much of what we need to easily separate content from code. We
have a plan to take us from the current monolithic implementation of LSP
to a much more modular system where all reformulations of the source
document are expressed as external functions that operate on the parsed
document. So, for example, the code for manipulating forms shall be
turned into an external module that acts through a core library of
functions that modify the document structure. As a result, it should be
substantially easier to modify any one module, or to add a new module
when needed.
I hope to have the new version of LSP available in some form over the
next couple of weeks. And hopefully by then I will also have a dedicated
mailing list going.
Sunil