[ptsefton.com] | [CV & Bio]

More about ICE and why XHTML is important


Peter Albion of USQ has responded to my last post on XHTML export from word processors.

There are a few points I think that need to be cleared up because I seem to sending the wrong message about ICE. (Peter also has a theory on my motivations for obsessing on HTML export but I'll deal with that separately with more detail about why it is important for usability and for ensuring the long term-sustainability of our content).

Peter says:

Unless I’m missing the point, both ICE and GOOD are being developed primarily to enable academics to prepare well presented course materials that can be published equally well on paper or electronically on the web or CD-ROM. I think, though not having prepared a course that way I’m not sure, that each of the electronic formats includes a PDF version of the material to facilitate printing by students.

Yep. That's the idea. This will involve PDF for printing and HTML for browsing, with all the materials packed up in IMS package format; that's the standard for interchanging webs of course content between content management (CMS) and learning Management Systems (LMSs).

This is important. ICE lets academics work in a word processor (OpenOffice.org Writer this week, and Microsoft Word within a few more weeks) and will automatically create both HTML and PDF. The PDF is currently manually generated if you want it – a painless process.

ICE is not going to be painful to use, not as painful as trying to create the same quality of output without an application like ICE to help, anyway. There is no need to export HTML and clean up the results. ICE does that. There is no need to worry about building navigation or packaging as an IMS package. ICE does that for you. There will be no need to export each document as PDF. ICE will do that for you. There will be no need to back up your files. ICE will do that for you. No problem working from home. ICE won't do that for you though!

Dr Albion continues:

If that is the case, then why bother with all the conversion from Word (or the Open Office equivalent) to XML and thence via XSLT to XHTML and/or PDF? Word, especially if used on a Mac where PDF capability is built into MacOS X, can easily enough generate PDF directly.

Why indeed. That's what my article was about but I obviously didn't make myself clear. We need to work on systems to export HTML because it is impossible to make a system export any given word processing document to HTML, that's because HTML is more constrained than Word or the OpenDocument format used by OpenOffice.org. And the vendors don't ship standard templates or stylesheets that are any use, preferring to focus on other areas of functionality.

Ultimately each course in ICE will have:

NOTE: Not all the functionality described above exists yet but we have made a good start.

Peter suggests:

Start with a good set of templates and you could get near enough to the same outcome for a fraction of the effort. The technology might not be so neat but it does work and would do the job.


Um, yes – that's what we're doing, starting with good set of templates, using whatever software we can and tying it together with the smallest possible application to make it easy to use. Templates and styles are worth doing even when there's no web publishing involved, and if that's all that ICE achieves then we're still ahead.

While we we're expending effort we will be saving many times that effort for others and, we hope, improving the technical side of courseware delivery by enabling more people than ever to create seamlessly integrated web and print courseware without lots of mucking around.

My point in the previous postis that if you start with a good set of templates you still need to a fair bit of work to customize the HTML output of any word processor to produce reasonable HTML. (Hands up who hasn't seen the awful results of HTML export from Word.) I don't know what Peter thinks is near-enough but what comes out of my version of Word or OpenOffice.org when I export stuff that I've made using my template is not near enough.

Yes there has been some effort put into HTML export in the ICE project, but not that much, we used prior work both from outsiders and from my previous efforts elsewhere (basically I arrived at USQ a year ago with working code that forms the basis of the conversion code in ICE, under the free GPL license, and that code has now passed into USQ ownership, and the template design we're using has been in use for ten years across a number of organizations).

A lot of the effort in ICE is going into building infrastructure for distributed authoring with version control and making the IMS packaging process seamless.

The ICE team is currently working on easy creation of interactive quizzing and easy embedding for multimedia, and doing so in a way that will integrate with other USQ systems and provide the best possible compromises for print and web delivery (such as allowing for transcripts of audio for the print version, and automatically creating answers up the back of the book for formative quizzes in the PDF version, while the HTML version will be interactive).

So, I think while he understands the essence of what ICE is setting out to do, maybe Peter Albion is missing part of the point. We are investing some programmer time in this, but we have active pilot users who we hope will help us make software that solves their problems, and allow USQ (and others) to generate top quality courseware in HTML and PDF with minimum effort, not just “near enough”. If we can save 100 people a few hours or days per semester for the next few years by automating previously manual conversion and re-formatting tasks that's worth doing isn't it? If we can improve the flexibility of USQ's hybrid-delivery by allowing Print, web and CD delivery with a format for every occasion and taste from the same source, that's worth doing as well, isn't it?

And finally – Peter, and anyone else at USQ if you'd like a demo, email sefton@usq.edu.au or call me on +61 (07) 4631 1640.