Best viewed with VIEW.EXE ;-)

HTML to IPF conversion tool

,HTML -> .IPF conversion utility

In the middle of Mar, 1997 I downloaded a handy OpenGL API reference from Digital corp. WWW site. Alas, I hate documentation in HTML format since it takes too much time until browser gets loaded, beside this each such documentation is a big set of HTML files, which you should keep in a separate directory, otherwise you`re completely messed up. I don`t like this since I have really lots (around 40 megabytes) of documentation on my HD in the INF format, and everything fits in a single directory without worry. So I thought of a way to make an 'regular' OS/2 INF book from a bunch of HTML and GIF files.

So I took a look inside all those HTML files (believe me or not, I didn`t know the HTML language until that). Then I looked into OS/2 IPF (Information Presentation Facility) reference and - oh God! - I've seen almost the same tags, except that IPF tags are written using ':' in the beginning and '.' at the and, and HTML tags are written using '<' and '>' respectively. (Seems that OS/2 IPF language is a slightly enhanced and modified version of HTML 1.0, although maybe I`m wrong - it could be the Unix MAN format, I`m not familiar with it).

It took me about two weeks to complete the task (I have a job beside hobbies, you know :-). I first wrote a 'first-version' in one night (it was really SLOW), then I invented a neat way to parse HTML tags in REXX (if you`re interested - take a look at the tech info section or even directly to the REXX script), and then about a week I`ve added some bells & whistles (you know, they`re taking most of the time). And here is it, the first version of HTML2IPF.

This REXX script will convert a set of .HTML files into one big .IPF file. The .IPF file can be used as a input to IBM's IPFC (Information Presentation Facility Compiler) to create regular OS/2 INF files. One more usage for this is that you can use any of lots of available today HTML editors to create OS/2 books.

What's the pro and cons for doing this?

Pro: Basicaly .INF files are much smaller than .HTML files. Usualy an INF file generated from a set of .HTML files occupies half of space occupied by HTML files (a exception from the rule are files with many GIF & JPEG pictures since IPFC picture compression is much worse than that of GIF and JPG). Another gain is that you have a single .INF file which you can easily manage.
Beside this, many software makers publishes complex HTML documents on their Web site, which many of us want to have at hand. This is especially crucial to miscelaneous white papers (say, the OpenGL reference manual -- the main cause why I started to write HTML2IPF), since INF file are fully (and fast) searchable.
Con: You lose portability. INF format is not so widespread as HTML, so you basicaly should convert documents for internal usage only (or documents which presents interest to OS/2 (and maybe DOS) community only). Beside this, IPF tag language is much less flexible than HTML, so complex documents (especially which makes heavy use of HTML 2.0 and 3.0 tags) will be converted into a unusable book (although in most cases you can manually fix this later, if the task is worth it).

This document contains several topics: