The PRASA 2003 programme committee sent out acceptance notifications on 12 October 2003. The accompanying email contained a single line that would dominate the next four weeks for every author: "Camera-ready papers must be submitted by 15 November 2003, 23:59 GMT." That date—the first and only camera-ready deadline for that symposium—was non-negotiable. No extensions were granted, and the proceedings were scheduled to go to the printer on 20 November. For researchers accustomed to the fluid timelines of journal revisions, this hard stop was a shock to the system.
In the mid-2000s, the camera-ready submission was the final gate before publication. Unlike today’s continuous online proceedings, most conferences printed physical books. The printer needed a complete, formatted PDF for every accepted paper—no last-minute corrections, no reordering of pages. The deadline therefore carried a weight that is hard to appreciate in the era of digital-first publishing. Missing it meant your paper would not appear in the proceedings at all, even if it had been accepted months earlier.
The Anatomy of a Camera-Ready Deadline
A typical conference timeline in the mid-2000s had three or four major dates: abstract submission, full paper submission, notification of acceptance, and camera-ready submission. The camera-ready deadline was almost always the last one, and it was usually set six to eight weeks after notification. That window was deliberately tight—organisers assumed that authors had already prepared their final text during the review process and only needed to incorporate reviewer comments and format the document according to the conference style.
The phrase "camera-ready" itself came from the days when authors submitted physical sheets that were photographed for offset printing. By the mid-2000s, the term referred to a digital file that met a precise set of specifications: correct page size (often US Letter or A4), embedded fonts, no hyperlinks except for the DOI, and a copyright block at the bottom of the first page. Conferences like PRASA and Contacts provided LaTeX style files or Word templates, but the responsibility for producing a compliant PDF rested entirely on the author.
The First Submission of the Camera-Ready: Why "1" Matters
Some conferences used a multi-stage camera-ready process—a first submission for review, then a corrected final version. But for most mid-2000s events, there was only one camera-ready submission. The "1" in the phrase "important dates details 1 submission of camera ready papers" refers to that single, all-important deadline. There was no second chance. If the PDF failed the automatic checks for font embedding or page count, the paper would be rejected or, at best, published with a disclaimer.
This one-shot approach created a specific kind of pressure. Authors had to ensure that their LaTeX code compiled without errors, that all figures were vector graphics or high-resolution bitmaps, and that the bibliography was complete. Many researchers spent the final weekend before the deadline wrestling with obscure TeX errors—missing packages, incompatible class files, or stubborn overfull boxes that pushed the last line of text beyond the page limit.
The organisers, in turn, ran automated scripts to validate each PDF. Common checks included verifying that the PDF version was 1.4 or lower (to ensure compatibility with older printers), that all fonts were embedded as subsets, and that the file size did not exceed a few megabytes (since email attachments were still the primary delivery method).
Papers that failed these checks were returned to the authors with a request to fix and resubmit—but only if there was time before the print deadline. In practice, this meant that the first submission had to be nearly perfect.
Technical Challenges at the Camera-Ready Stage
The mid-2000s were a transitional period for document preparation. LaTeX was dominant in computer science and engineering, but many speech and pattern recognition researchers used Microsoft Word, especially for shorter papers. Word’s PDF export was unreliable—fonts often defaulted to non-embedded TrueType, and page margins could shift between versions. The recommended tool was Adobe Acrobat Distiller, which required authors to print their document to a PostScript file and then distill it into a PDF with controlled settings.
For LaTeX users, the standard workflow was to compile with pdflatex and then run a script that checked for font embedding.
A common error was the use of Type 3 fonts (bitmap outlines) instead of Type 1 (vector outlines). Type 3 fonts looked blurry when printed and were often rejected by conference PDF checkers. Authors had to ensure that their LaTeX distribution used the cm-super package or similar to get Type 1 versions of Computer Modern fonts.
Another recurring issue was the inclusion of hyperlinks. Many authors used hyperref to create clickable cross-references in their PDF, but some conferences forbade any hyperlinks except for the DOI. The solution was to compile with hyperref disabled or to strip hyperlinks using a post-processing tool. The PRASA 2003 guidelines explicitly stated: "No embedded hyperlinks are permitted. The PDF must be suitable for black-and-white printing."
The Human Element: Committees and Correspondence
Behind every camera-ready deadline was a small team of organisers who handled the submissions manually. The Karin van Wyk and the Contacts 3 Organising Committee exemplifies the kind of dedicated individuals who spent hours verifying PDFs, answering frantic emails, and occasionally granting a short grace period to authors whose files had been corrupted by a faulty FTP upload.
These committees operated without automated submission systems—most used email attachments or simple web forms that accepted a single file per paper. The camera-ready deadline was as much a test of the organisers' patience as it was of the authors' technical skill.
One memorable incident from a 2005 speech technology conference involved an author who submitted a PDF that appeared correct but had a hidden watermark from a trial version of a PDF editor. The committee chair noticed it during a manual spot-check and contacted the author, who had to re-create the entire PDF from scratch the night before the print deadline. Stories like these circulated in lab meetings and hallway conversations, reinforcing the importance of testing the camera-ready file on a different computer before submission.
The broader context of how these deadlines shaped the research agenda is explored in the article How Calls for Papers Shaped Research in the Mid-2000s. The camera-ready deadline was the final step in a process that began with a call for papers months earlier, and the tight schedule meant that authors often had to choose between polishing their experiments or fixing formatting issues. Many chose the latter, because a perfectly formatted paper with average results was better than an unformatted paper that could not be published.
Deadline Day: A Ritual of Its Own
The hours before a camera-ready deadline had a ritual quality. Authors would gather in computer labs or cluster around a single machine, checking and re-checking the PDF. The final submission often involved uploading the file to a conference website that ran on a simple PHP script—no version control, no confirmation email beyond a thank-you page. A common practice was to include a checksum (MD5) in the submission email so that the organiser could verify the file had not been corrupted during transmission.
For conferences that accepted camera-ready papers by email, the subject line had to follow a strict format: "[CONFERENCE-ID] Paper #1234 Camera-Ready." Any deviation could cause the submission to be overlooked. Authors would send the email, then wait for a manual reply from the committee confirming receipt. If no reply came within an hour, a second email was sent—politely worded, but with a hint of desperation.
Lessons from the Archive
Looking back at preserved proceedings from that era—such as those in the Inside the PRASA 2003 Archive: A Snapshot of Mid-2000s Pattern Recognition—one can see the results of those frantic camera-ready weeks. Some papers have subtle formatting inconsistencies: a figure that slightly overflows the margin, a reference list in a slightly different font size. These imperfections are not signs of carelessness but rather evidence of the constraints under which the final versions were produced. The camera-ready deadline was a forcing function that prioritised completeness over perfection.
The single-submission model also meant that the version submitted on the deadline day was the version printed.
There was no opportunity to notice a typo after the fact and submit a revised PDF. This is why many authors still keep a printed copy of their own camera-ready paper—it is the definitive record, warts and all.
For the historian of technology, the camera-ready deadline reveals the logistical backbone of mid-2000s academic publishing. It was a moment when the abstract world of research collided with the concrete realities of printing presses, font licenses, and file formats. The next time you browse a conference proceedings from 2004, look at the copyright date and the page count. That paper made it through the camera-ready deadline on the first and only try—or it did not appear at all.