[zgramana]

Surprised to see the author suggest that the C64 keyboard did not include the graphics glyphs on the keys.

Both the VIC-20 and the C64 both had them, albeit printed on the front vertical face instead of the horizontal face like the PET.

Moreover, the VIC-20 and the C64 keys include the box glyphs as well so you could see what character you’d get via CTRL and the ‘C=‘ (“Commodore”) key.

The Wikipedia page* states PETSCII includes 192 characters, but the author seems to suggest it’s at most a 7-bit charset. I was a little surprised to find that out after reading the article.

Fun walk down memory lane.

* https://en.m.wikipedia.org/wiki/PETSCII

[masswerk]

The idea was more about the arrangement of the character set, which isn't always obvious on the VIC-20 or the C64, because of the differences in keyboard layouts as compared to the original PET 2001. (E.g., both the VIC-20 and the C64 are lacking the numeric keypad and what are adjacent blocks on the PET is dispersed over a linear arrangement. So there's no discernible logic in the code, nor in the keyboard. The same is true for several other keys, like those around the space key of the PET.)

Regarding the number of actually assigned codes, mind that the basic code, regarding unique displayed characters, is single case (6 1/2-bit) in two banks (MSB set for shifted characters => 0x60 x 2 = 192). However, the mirrored regions are actually unique codes in BASIC (as far as CHR$() and ASC() are concerned).

Sorry, if I failed in making this clear.

Edit: Arguably, the ortho layout of the PET 2001's chiclet keyboard also helps in identifying logical blocks of graphics characters.

[rzzzt]

The location is dependent on the model: the C64C had the symbols on the top of the keys, the breadbin on the side.

[rachelbythebay]

FYI: https://commons.wikimedia.org/wiki/File:C64C.jpg

Lots and lots of revs, apparently.

[xenomachina]

Yup. Early revisions of the 64C had the glyphs on the fronts of the keys, but this was later cost-reduced to having them on the tops.

This is also one of the differences between the "Aldi C64" (breadbin case, but beige keys with front glyphs) and the C64G (beige breadbin case, but beige keys with top glyphs). Commodore loved to mix and match components, too. Some C64Gs have the brown keys of the earlier breadbin C64.

[rzzzt]

Do you know if Commodore manufactured any two styles of these in parallel intentionally, or if this was a matter of making use of existing housing/keyboard assembly/motherboards?

The corresponding article on C64 Wiki says that side symbol keyboards were used during the first year of production, and also includes a picture of the top glyph version: https://www.c64-wiki.com/wiki/C64C

[xenomachina]

My understanding is that the "main line" was breadbin, 64C with front glyphs, then 64C with top glyphs. The Aldi 64 was concurrent with the breadbin, but mysteriously used the keys from the early 64C. The 64G was concurrent with the late 64C. The Aldi and 64G were only available in a relatively small area (mostly Germany), rather than internationally.

Commodore was notorious for mixing and matching parts, though. Even the very early breadbin 64s would sometimes ship with VIC-20 function keys (orange instead of dark gray) and vice-versa. I don't know if this was due to poor quality control, or an intentionally sloppy supply chain management. Perhaps a combination of both.

The 64G especially looks like they were scrambling to meet demand any way they could, as the key colors and even the LED colors vary. I'm guessing the beige breadbin was used to increase production speed by taking the old breadbin injection molds out of retirement. They aren't old cases that they had in inventory, because breadbins weren't that color previously.