« 23.02.2006 | Calendar
Tuesday, 28 February 2006
This marks the end of the original programme of work and presents a summary of what the project has achieved and what remains to be done. The project has met or exceeded all its original design objectives. People who have tried it have sent a number of positive and unsolicited Testimonials.
1. Original Aims
The goal was to develop a software service to support print-on-demand:
- let an author work with others to create a one page letter, a 10 page paper, 100 page report, or 1000 page book, using only a web browser
- let authors generate either print (PDF) or web versions of the document with one mouse click, applying suitable formatting for each medium
- produce printed output of desktop-publishing quality, without authors having to become publishing experts
2. Where are we now
There are few websites that offer any kind of print-on-demand service that we can find. The typesetting server is already more advanced than these in the breadth of capabilities it offers, including:
- assembly of multiple web pages into a single print document
- floating and resizing images, with the ability to substitute high resolution images for printed output
- automatic calculation of table column widths and support for long tables with running headers
- equation markup for dynamic mathematical typesetting to web and print
- advanced typesetting features, such as drop capitals, marginal notes and hanging punctuation
- generation of print constructs such as table of contents, footnotes and page number references
- use of web forms to change the appearance of the print output with a few mouse clicks
The software still has the weaknesses one expects in a prototype:
- it usually works perfectly but it may die; any input errors in the XML mean no PDF is returned
- it takes time (over a minute) to compose and return a long document
- some features are missing (support for citations) or incomplete (support for style attributes)
The software is available under open source licence — a wikipublisher library for PmWiki and a wikibook PDF server. Some 40 sites are registered wikipublisher users.
3. Where do we want to be
The technical and commercial future are intertwined. Eventually, we would like to see:
- a wikipublisher plug-in for all the major web content management systems (such as Drupal, Mambo and Plone) and wiki engines (especially mediawiki), teaching them to produce output in Wikibook XML
- a template system that lets people pass a “skin” to the typesetting server and control the look of the PDF, in much the same way as PmWiki skins change the look of web pages
- demand-based feature enhancements, such as citations, document indexes, letters and envelopes, or cover-page images
- at least 100 wikipublisher servers and 1,000 registered sites by the end of 2007
- a viable and sustainable business based on wikipublisher-related services and fees
This direction takes advantage of Wikipublisher’s key design innovation — separating generation of print output from the web content management system. This means that, in principle, any web site can be taught to produce PDF and the PDF server does not need to know anything about the site requesting the PDF. Some web sites will be easier to teach than others.
The biggest immediate barrier is the lack of a suitable business model to commercialise the invention. To create a job for one person full-time means raising about $150,000 per year in revenue. The biggest barrier to wider adoption is the currently immature state of browser-based WYSIWYG editors for web content management systems. It doesn’t matter that a markup-based approach, such as using wiki markup, is better than WYSIWYG in all sorts of ways; too many people are unwilling to try a different way of working. Many people perceive learning wiki markup as hard and WYSIWYG as easy.
4. How will we get there
We need to develop a business model that will “monetise” the service. How are we going to do this? On the CMS end, seek opportunities for commissions to develop plug-ins for other wikis and CMS that generate wikibook XML from the site content; this may mean investigating a XHTML to wikibook translator. On the typesetting server end, there appear to be more options:
- charge an annual support fee to server installations, including a set-up service
- display advertising on the typesetter’s Splash Screen while people wait for a PDF
- charge an annual subscription for those who want to use the shared typesetting server
- offer training and documentation services
- seek commissions for functional enhancements, such as custom layouts, citations or document indexes
We need to start gathering statistics — who is requesting PDF generation and when. We need to know how much people are actually using the service.
The next phase of development needs more of a marketing focus and less of a technical focus.