dimanche 22 octobre 2017

Transparency real and imagined

from the all very well in theory department

Back in August I came across this blogpost—a somewhat generic homily on the importance of openness and transparency in scientific communication. I share the author's optimism that open access to the scientific literature will provide better validation of published work than traditional systems of peer review. But unless this superior access is actually exploited by knowledgeable users, the potential gains may not be realized. Though a blinking cursor beckoned me below the line, the author has, in his or her wisdom, failed to publish my comment. So it must appear here:
Transparency is superior to trust—as long as some relevant person(s) actually exploit(s) the transparency. Look at how long that ssl flaw hung about in Debian, for example: https://pinboard.in/u:juliusbeezer/t:security/t:opensource/
That was all open code, utterly vital to the security of hordes of crucial servers run by the world's top-most geeks, and therefore, every internet user. But the problem sat there for two years, apparently.
That's an extreme example that did get fixed. Transparency is necessary yes, but unless it's actually backed by readers/critics/reviewers/coders/experts actually looking through the windowpane afforded by it, its value is only rhetorical.
It does mean that the guards can guard the guards and we can watch the guards guarding the guards though. Or maybe McGregor-Maywether.