Note that this comparison is really old, from mid-2014, while the tested dietlibc is from 2009 (although not much changed since).

Good point - it should definitely be dismissed without any further explanation. What possible relevance could anything from 18 months ago have to today?

What? I never even implied anything like that. It just should be taken with a grain of salt, as quite a few points will be out of date (e.g., supported features, code size, …).

From the text: "uClibc 0.9.33.2 (Buildroot 2015.02)"

In all fairness, this can't be from mid-2014 if they used a library build from 2015.

So why are both musl and glibc from mid/early 2014, and dietlibc from 2009? That's an awful lot of cherry-picking versions with no stated justification.

I think it's because those are the freshest stable versions the author could use? uClibc 0.9.33.2 by itself was released in 2012. I guess he had a reason for using the BuildRoot 2015.2 version.

Nice comparison. It would be interesting to see how different BSD's libc compares to these.

The footnote indicator † is not resolved in this document, leaving us to speculate as to what the author could possibly have meant by "2.0M†'.

I'm guessing it's the inclusion of iconv:

> size totals for glibc include the size of iconv modules, roughly 5M, in the “Complete .so set” figure. These are essential to providing certain functionality, and should be installed whether static or dynamic linking is being used.