There is also a reason why we don't typeset books in monospaced fonts, but oh well...
You could argue it is easier to read (and better for the environment too) but it wasn't that long ago when fixed width was the norm (typewriters, dot-matrix printers, early micro-computers / mainframes, etc) and people read just fine.
A fair amount of the older workforce did grow up with typewriters and, later, dot matrix printers (or even teletypes) in their offices. Many on even started their careers using computers that only supported monospaced typefaces. I remember the first time I saw variable width fonts on a VDU (as they used to call "monitors") and it took me a second or two to adjust to typeface because most of my text consumption up to that point had been with fixed with characters.
Or to use an analogy: horse-drawn ploughs might have pre-dated tractors by hundreds of years but that doesn't mean I have more experience with seeing horses ploughing fields than tractors.
Personally I find fixed-width very hard to read for regular text.
For example, compare these two documents:
http://www.laputan.org/pub/papers/aim-349.pdf
http://www.larcenists.org/Documentation/Documentation0.98/r7...
I know which one I prefer when it comes to typography and readability.
That doesn't make me agree with you any less, though. Some words are just unnaturally wide in monospace.
The only true meaning of natural text is the meaning given to the words; but the only true meaning of code is the way punctuation is organised around superficial natural texts within the elements.
In books, I agree. Even Gutenberg's press had variable width "types". However not everyone reads books regularly and fixed width was all that was available in offices (well, that or handwritten documents). So for a great many people like myself, fixed width typefaces were still the norm.
Fixed width was used because the toolchains for it were cheaper (with computers, that included software that allowed to visualize document appearance.)
In the era when typewriters and dot-matrix printers dominated, sag, casual office documents, professional print still used proportional type for all the same reasons that that is the norm today, and that it was for centuries before typewriters, much less dot matrix printers.
I wasn't saying variable width didn't exist. My point was that a great many of us used to spend hours each day (many still do, in fact) working with fixed width and we managed just fine. The reason for my point is because some people on here love to overstate and exaggerate the effect different type-faces have on readability when often it's more of an issue with familiarity (eg if you only ever read fixed width typefaces then variable width types can sometimes take a moment to adjust to).
To illustrate this point, use a variable width typeface in your code editor. It's still perfectly usable but man is it jarring when you're used to monospaced fonts.
You haven't previously clearly articulated your point, it wasn't clear if you were being wrong but relevant (i.e., claiming that there was no significant usability advantage to proportional fonts for text documents) or making an accurate rebuttal of a strawman argument (i.e., correctly claiming that fixed-width text has non-zero usability in rebuttal to the claim that was never made that it is completely unusable.) As such, I made the generous assumption that you understood the context which favored assuming the former, but you seem to have instead settled on the latter position.
> To illustrate this point, use a variable width typeface in your code editor. It's still perfectly usable but man is it jarring when you're used to monospaced fonts.
I recently transitioned to Input Sans, a proportional coding font, for all code editing full-time, after having used monospaced fonts for coding pretty much exclusively since elementary school, for a little over 35 years; it was a big improvement and the transition was not jarring at all.
Also code inherently has a structure whose readability is increased by monospace fonts.
Unless you are programming in FORTRAN or COBOL, I very much contest that claim