The problem the article highlights is that there is software that is widely used in many GNU/Linux distributions that is effectively *unsupported*. There's nobody behind it, accepting bug reports, or actively maintaining the code. This is the exact *opposite* of the Linux kernel developers!
I'm no fan of the GNU/FOSS communities in general, and certainly not its tedious politics, but, for once, Linus Torvalds is not the problem here.
The problem is one we used to call "bit rot". Old code, in common usage, that is a time-bomb waiting to happen. Not just because of potential security risks either, but because an important API that code relies upon might change tomorrow, causing all sorts of major headaches.
Such code thus also acts like a drag on improving such APIs, holding back development and forcing programmers to jump through additional hoops to keep that old code happy. And it's those hoops that cause hackery and kludges to appear in code, opening up wonderful new possibilities for bugs and security issues.
It took them long enough, but Microsoft have finally learned to say, "No!" to continued legacy support in Windows precisely because of problems like these. (Apple never said "Yes!" in the first place; they explicitly state that each major new release of OS X may break old software.)