PowerPoint presentations are the central spectacle at any scientific conference. In cool, darkened rooms, slide after slide marches from background to methods to results with inexorable logic. But PowerPoint has come under criticism lately for a tendency of the form to abbreviate truth, gloss over details, create an impression of analysis where there is none, and facilitate spin. The leading critic of Microsoft's ubiquitous presentation software, Edward Tufte, author of several books on representing complex quantitative data, points to the role of PowerPoint misdirection in the catastrophic failure of the Columbia space shuttle. During a critical post-launch meeting the potential for serious wing damage at takeoff was masked by a bland slide title disguising the alarming reality nested within dumbed-down bullet points below. PowerPoint, Tufte says, has come to signify authority, and audiences numbed to an endless succession of slides are rarely given an opportunity to critically evaluate what is being presented.

In February 2004, at the Moscone Center in San Francisco, thousands of PowerPoint slides will flash before the eyeballs of the most important workers in HIV research, as the annual four-day Conference on Retroviruses and Opportunistic Infections (CROI) unfolds. A quirk of scientific conferences is that the papers deemed most important are presented as oral slide talks, while the rest are relegated to the poster hall where the data is hung on row after row of poster boards. The paradox is that important data from the slide shows, which can sometimes have an immediate impact on clinical practice, may never appear on paper; they are ephemeral and often difficult to reference. Meanwhile, the lesser, paper-based presentations usually contain much more detail; they can be studied at length and their authors examined in the gauntlet of the poster hall. It's as if the reward for scientific excellence is to receive less scrutiny and accountability.

The data from a featured oral slide presentation may eventually appear in a peer-reviewed journal article or, if a drug company has an interest in the data, it will be repackaged into ever-slicker slide shows and shown again and again at seminars and sponsored educational forums.

But until the study appears in print, the only source for the data is the conference abstract, which may have been written many months before the conference. Often, data presented in the slides can be substantially different from that in the abstract. This leaves third-party reports by online sources such as Medscape, NATAP or TheBody.com as the only way for people who missed the conference to find out what was presented.

Since CROI is a top-notch conference, the PowerPoint presentations on display will typically be well organized and checked for errors. Often, where data has been produced with pharmaceutical company investment, the slides will be carefully crafted by contract or in-house medical communications experts, who are very sophisticated in seeing that the correct nuance is given to each bit of data displayed. An amateurish slide show may have the narrative force of Rocky II, but this won't be as effective as a display in which the hand of the communications expert has been carefully effaced and the data appears to glow on it own. For a product in the pharma pipeline, a slide talk at an important conference becomes a key plot point in a larger tale of home run inevitability that culminates in FDA approval. Fortunately, there are always sharp scientists in the seats who will pop up to the microphone to skewer any egregious spin — or put in a plug for their own company's interpretation.

Good scientists instinctively look between the slides for what is not shown. They know that, while the process of viral entry can be boiled down to three bullets, in reality the shifts of energy and matter that move these molecules into their crucial alignments are contingent on forces near and far in space and time. Such evidence-based, complex mental maps do not lend themselves to frame-width sentence fragments.

PowerPoint, like industrial agriculture, is not going to go away soon. But maybe one conference session could be reserved for presenters who prefer to use a chalkboard.