I’d love to know what areas specifically that computers are undervalued in scientific pursuit. I am sure that’s true in many fields, but it would be interesting to see where from your perspective. (Also, what is your perspective? Are you a working scientist? Just curious.)

In biotech, interestingly, there are some important instances where computers — or at the very least, automation — can tip to the “overvalued” side. The most obvious example: gene-disease links.

We’ve all heard about “the gene” responsible for producing various diseases in individuals — diabetes, Alzheimer’s, obesity, schizophrenia, depression, etc. But in early 2005, a study reported that as many as 95 percent of these gene-disease links don’t actually exist.

The main reason cited is the speed and efficiency with which computers can sequence and analyze genes. Reading gene sequences can take about a day now instead of several months, but because the sequences are analyzed by software without a human reality check, some of them get linked randomly to a disease in a statistically significant way.

Everybody is eager to publish a new link, but as the study authors said, they don’t usually submit the “Ooops! Our bad” stories, and journals don’t often print retractions. And so a reductionist view of biology persists.

Just a warning flag about the reliability/fallibility continuum of technology …

The good news. A lot has happened in the past couple of years, so the prognosis for the communication aspect of computing is a good one.