In the archived files of the Pattern Recognition Association of South Africa (PRASA) 2004 symposium website, a single HTML page sits at the path docs/authorskit/prasa2004/authorskit.html. This page, preserved in the Internet Archive, contains the complete author instructions for the 16th Annual Symposium of PRASA, held in 2004. It specifies margins of 2.5 cm on all sides, a 12-point Times New Roman font, a maximum abstract length of 250 words, and a strict requirement that references follow the IEEE style. The page also links to a Word template (prasa2004.doc) and a LaTeX style file (prasa2004.sty). For a researcher preparing a paper in 2004, this HTML file was the gateway to acceptance.
Context: PRASA 2004 and the Pattern Recognition Community
PRASA was founded in 1995 to foster pattern recognition research across South Africa and neighbouring countries. By the mid-2000s, its annual symposium had become a meeting point for work in biometrics, document processing, speech recognition, and computer vision – all areas covered by this blog. The 2004 symposium, held at the University of Pretoria, featured sessions on face recognition, fingerprint classification, offline handwriting recognition, and early deep learning approaches (then called "multilayer perceptrons"). The author kit for that year reflects the state of academic publishing before the widespread adoption of online submission systems like EasyChair (which launched in 2002 but took years to dominate).
Anatomy of an Author Kit
A typical mid-2000s author kit contained several components, each designed to enforce consistency across submissions:
- Template files – Word
.docor LaTeX.styfiles that pre-set margins, fonts, and headers. - Example paper – a PDF or Word file showing the correct layout, often with placeholder text.
- Instructions page – the HTML file that explained every formatting rule and submission step.
- Copyright form – a PDF or scanned image to be signed and mailed.
The PRASA 2004 kit followed this pattern. Its HTML instructions were particularly detailed: they specified that figures must be in black-and-white or grayscale, that the bibliography must be numbered, and that the paper should not exceed six pages. The page even included a note that electronic submissions should be sent as a single PDF file attached to an email – a common workflow before conference management portals became standard.
The HTML Instructions Page as a Historical Artifact
Why HTML? In 2004, HTML was the most universal format for distributing structured information. It could be viewed on any operating system without special software, and it allowed hyperlinks to template files. The PRASA 2004 page used a simple table for layout, with a blue header and bullet points for each rule. It did not use CSS beyond basic font tags. This simplicity is itself a document of the era: web standards were still evolving, and conference organisers often hand-crafted such pages in a text editor.
The page also included a section on "Language and Style" that encouraged authors to write in clear English and to avoid jargon without definition – a reminder that the audience included engineers from multiple disciplines. Another section warned against using scanned images of text, as they would degrade in print. These instructions reveal the practical challenges of producing a printed proceedings volume, which was still the norm for PRASA symposia.
Impact on Submission Quality and Consistency
Author kits like PRASA 2004's significantly reduced the number of formatting rejections. Before their widespread use, conference chairs often received papers with wildly different margins, font sizes, and reference styles, causing delays in the review process. By providing a single HTML page with explicit rules and downloadable templates, the organisers shifted the burden of compliance to the author. The result was a proceedings volume that looked uniform and professional.
The kit also introduced authors to LaTeX, which was gaining traction in the pattern recognition community. The prasa2004.sty file defined commands for section headings, the abstract environment, and the bibliography style. For many South African researchers, this was their first encounter with a markup-based document system – a precursor to the XML-based workflows that would later become standard in academic publishing. The transition from author kits to online submission systems eventually led to the use of Atom XML feeds for syndicating conference metadata, a topic explored in an earlier post on this blog: How Atom XML Became the Academic Syndication Standard for Pattern Recognition Research.
Legacy and Evolution
By 2008, PRASA had moved to an online submission portal (first a custom PHP system, later EasyChair). The author kit HTML page was no longer the primary entry point; instead, authors uploaded papers directly through a web form. However, the 2004 kit remains a valuable snapshot of the pre-automation era. It shows how much labour was involved in formatting a single paper – adjusting margins, embedding fonts, converting images to grayscale – and how much trust organisers placed in authors to follow instructions manually.
Today, the PRASA 2004 author kit can be accessed via the Wayback Machine at its original URL. The HTML is still rendered correctly in modern browsers, though the linked template files are no longer hosted. Researchers interested in the history of scientific conferences can study such kits to understand the material culture of academic publishing. The meticulous attention to detail – down to the exact spacing of the header – reflects a time when the printed page was the final product, not just a PDF on a screen.
That same HTML file still loads in a browser from the Wayback Machine, its blue header and bullet points unchanged. The template files are long gone — the Word .doc and LaTeX .sty no longer hosted — but the instructions remain, a quiet artifact of the days when conference organisers typed margin values into a text editor and hoped every author would read them carefully.
