Mommy, where did straight apostrophes come from? (was Re: [ANN] Lua 5.3.0 (beta) now available)

[Jay Carlson <nop@...>]

On Oct 24, 2014, at 3:57 PM, Richter, Jörg <Joerg.Richter@pdv-FS.de> wrote:

>> Is there any historical reason for that use of the backtick in quotes? 
>> I always found it kind of odd. Does its use came fro TeX ?​
> 
> It is related to old X fonts. See http://www.cl.cam.ac.uk/~mgk25/ucs/quotes.html

Markus Kuhn has done many great works to advance the cause of internationalization. However, that page is extremely misleading, especially if it is understood as a historical account rather than a call for change. For example, he writes:

"The two ASCII characters 0x22 [["]] and 0x27 [[']] are supposed to represent the neutral glyphs commonly used on typewriters. They should *not* be used as directional quotation marks."

This is true under Unicode/10646. It is an interesting normative judgment. But it is also contrary to the de facto standardization of glyphs on nearly every character cell ASCII device prior to Unicode. Let's start with VGA:

http://commons.wikimedia.org/wiki/File:Codepage-850.png

Oh, you don't trust some random .pcf file Linux ships? How about IBM?

http://www.vintage-computer.com/vcforum/showthread.php?44256-Original-IBM-MCGA-VGA-Text-%288x16%29-Font-from-PS-2-Model-30-%28Revision-0-09-02-86%29

http://www.mediafire.com/view/23hpqqhkhr22x2x/IBM_VGA_8x16_Font_%28PS2_Model_30_BIOS_Original_09-02-86%29.png

For a quick survey of what 8-bit computers were doing:

http://damieng.com/blog/2011/02/20/typography-in-8-bits-system-fonts

The Atari 800 is the only neutral apostrophe among the full-repertoire machines. The Commodore Amiga had a curly apostrophe; for example, Matt Dillion's DME and DTerm originally used `' for grouping rather than ().

I'll add the VT100/VT220's curly apostrophe, because the appeal to the authority of "many classic terminals such as Siemens/Nixdorf" is extremely misleading if not unfalsifiable. 

http://vt100.net/dec/vt220/glyphs

I've got a LA36 "DECwriter II" and Televideo 925 in storage. Guess which direction their apostrophe points. (Bonus points if you remember why the Hazeltine 1500 would be interesting in this discussion.)

The careful reader may notice I am avoiding the Apple II and ADM-3A. Both archetypical devices were originally upper-case-only machines, and as a result had no backtick in their original repertoire. I don't remember what the factory lower-case adm3a did anyway.

In addition, Kuhn shows a Windows NT 4.0 DOS prompt using Lucida Console. This is cherry-picking of examples. Most NT DOS prompts were using bitmapped fonts modeled on MCGA/VGA glyphs, both out of habit and because screen resolution was low enough that non-ClearType Lucida was suboptimal.

The Mac has had neutral apostrophes since the beginning. I stopped looking after getting as far back as Windows at 3.0. These systems could afford neutral apostrophes in system text, as they had "smart quotes" available in their proprietary code pages.

So why did X change the glyphs it distributed? I don't know the internal history of x.org and friends, but my memory is that the glyph transition wouldn't have happened without Markus Kuhn *pushing*. As one of the lua-l Unicode weenies, I suppose I have to agree with his logic of compatibility. As a programmer in many environments which were designed on character cell terminals and used [[`']] as a matched pair, I'm still annoyed it was not handled better for X's programmer users. 

Nobody particularly wanted to make this change since it involved redrawing glyphs in the hundred(s?) of X11 bitmap fonts. As for the aesthetics of those freshly amateur-drawn bitmap glyphs of apostrophe, I plead that the following is designer jargon, and hence appropriate language:

THOSE GLYPHS LOOKED LIKE ASSSSSS.

-- 
Jay