JTM - webbification

Ancient history

John Larmouth, as the original editor of JTM, had used the early SGML publication system at ISO, and had the texts of the documents in SGML  (that is as a text file tagged according to the SGML DTD for ISO standards). However, the turnaround in processing between Salford and Geneva had led him to develop a Basic program that converted the SGML-tagged file into a file tagged for the Xerox Ventura desktop publishing system. (This was before word processors were wysiwyg or had the power to match the ISO house-style).

John passed the editorship to me in 1988, along with the documents and the Basic program. I didn't have access to a machine with intelligent Basic, so re-wrote the program in SNOBOL (I'm just telling you what happened here, not claiming that it all makes sense - each decision step was probably reasonable at the time :-). The SNOBOL program then gradually got more sophisticated - it handled the numbering of the clauses and sub-clauses (always a nightmare with the ISO house-rules, though JTM follows a rule that was changed soon afterwards) and conversion of the cross-references from a symbolic form to numbers (so I could write things like (See <REF REFID=HOLDPROC>), but it printed as (See 2.3.2). After the first edtion had been sent for publication (in 1989), the remaining work was the amendment to the protocol standard to cover the Full Protocol. This had to be ballotted as a document of editing instructions but to be comprehensible, it also needed to be circulated as a change-marked text. I mangled the SGML (which had no direct relation to the ISO standards DTD by now), to handle this, and with the SNOBOL program and quite a lot of fiddling with Ventura settings, was able to produce the old text, the editing instruction form, the change-marked text or a clean merged text all from the same source file. (It got even more complicated when there were corrigenda involved as well as the amendments)

This arrangement was used for the publication of the second edition of JTM in 1992. (I was also using the SGML/SNOBOL/Ventura setup for the CCR standard but eventually gave it up and converted to Word)

And now

JTM then lay on the shelves for five years. After it had been withdrawn, and ISO said I could put it on the web the obvious thing was to convert it to HTML. (I have never checked to see if my change-marking variant of SGML could be handled by an appropriate DTD, or if it could be changed to be legitimate XML. I rather doubt it, since the change-marking is liable to overlap with the other, regular, tags). In any case, all that was needed was to clear out everything that wasn't in the final version. Rather than resuscitate my knowledge of SNOBOL, I wrote a set of Perl scripts that did this, and others that converted the SGML tags to the appropriate HTML one (including some further oddities that I had introduced to twist Ventura into doing what I wanted). All the cross-references could be made into hyper-links, tables converted (Ventura has its tables in separate files, but using a standard, and thus parsable, syntax) etc.

The diagrams in the original standard were GEM files (I've forgotten what the application was called that wrote these - Gem, I think). Fortunately, CorelDraw can read these, and the only thing it loses are the arrow-heads. Although they are all line-drawings, I have written them out as gifs, to maximise the chance of readability. Some of them are rather large on the screen, but since they are all just black-and-white, they don't come out as very large files.

Then just a few extra things - linking buttons to chain through the sections, tables of contents, and the explanation of what JTM was, and here you are.

I wondered whether to make each document all one file, or split them up more. As it is, they are just the files as they were 10 years ago - split mostly for size, I think (remember, this was started on an IBM PC/XT - the SNOBOL program was developed on a 286).

So you can now go

  • back to the general explanation
  • ISO 8831 - JTM concepts and service
  • ISO 8832 - JTM protocol



  • Peter Furniss
    peter@furniss.co.uk
    8 February 2000