Let me apologize first if this should go to a new thread as I can't really figure out where it belongs.
I computed, using my own keycaps database, the 78 keycaps to be added to the US ANSI alphanumeric kit to cover the following additional languages: DE, UK, SE, NO, DK, IT, ES, PT and came up with the following differences with wodan's graphical representation (posted above), and came up with the following differences:
(1) R1 3 · # es(E03) instead of R1 3 # es(E03). cf.
https://en.wikipedia.org/wiki/QWERTY#Spanish: "some characters required only for typing Catalan and Occitan ... the interpunct (punt volat/punt interior, used in l·l, n·h, s·h; located at Shift-3". Also note that tere is no latin keyboard that comes with R1 3 #.
(2) R2 E € es(D03) instead of R2 E es(D03). cf.
https://en.wikipedia.org/wiki/QWERTY#Spanish : "On most keyboards, € is marked as Alt Gr + E and not Alt Gr + 5 as shown in the image. However, in some keyboards, € is found marked twice". This suggests that R1 5 % € se(E05),it(E05),es(E05) could be dispensed with, as R2 E € se(D03),it(D03),es(D03) are also present.
May I add my own remark about building a minimal international ISO kit: I would not go for a set of keycaps that does not contain all glyphs needed both for typing in a given language and for arithmetics and coding in the commonly used computing idioms.
That is, I require at least, on the one hand, all upper (or lower) case letters without diacritics, diacritics (dead keys), if any, when all of their occurences in the given languages are not provided by single letter+diacritic glyghs, and punctuation marks; on the other hand all digits, arithmetic operators, and additional glyphs require for coding.
The latter amounts to the following 30 non-alphabetic glyphs (sorted by ASCII codes):
! " # $ % ( ) [ ] & ' * + , - / ; < = > ? @ \ ^ _ ` { | } ~
where | may well be displayed as ¦ on a keycap.
When computed following the above rules, and using only keycaps already defined in standard national layouts, I came up with a set of 43 additional keycaps needed to cover UK, DE, SE, NO and DK. Then, with the aim is to provide a set of keycaps suitable for typing in the above five languages, SE/NO/DK can be covered using the SE layout (which is actually is a SE/FI layout), since this is the closest nordic layout to the DE layout and the one usually provided by major manufacturers (Realforce, Filco, Poker), plus the two additional alphabetic keycaps (R3 Ø and R3 Æ) in both the DK and the NO layout. The total number of keycaps is thus brought down to the following 36 keycaps:
R1 ˆ ° de(E00)
R1 2 " ² de(E02), R1 2 " uk(E02) being a possible alternative
R1 3 § ³ de(E03), R1 3 # £ se(E03),no(E03),dk(E03) being a possible alternative
R1 6 & de(E06),se(E06)
R1 7 / { de(E07),se(E07)
R1 8 ( [ de(E08),se(E08)
R1 9 ) ] de(E09),se(E09)
R1 0 = } de(E10),se(E10)
R1 ß ? \ de(E11)
R1 ´ ` de(E12),se(E12)
R2 Q @ de(D01)
R2 Z de(D06)
R2 Ü de(D11)
R2 + * ~ de(D12)
R3 Ö de(C10),se(C10)
R3 Ä de(C11),se(C11)
R3 # ' de(C12)
R4 < > | de(B00),se(B00)
R4 Y de(B01)
R4 , ; de(B08),se(B08)
R4 . : de(B09),se(B09)
R4 - _ de(B10),se(B10)
R1 ` ¬ ¦ uk(E00)
R1 3 £ uk(E03)
R3 ' @ uk(C11)
R3 # ~ uk(C12)
R4 \ | uk(B00)
R1 § ½ se(E00)
R1 2 " @ se(E02)
R1 4 ¤ $ se(E04)
R1 + ? \ se(E11)
R2 Å se(D11)
R2 ¨ ˆ ~ se(D12)
R3 ' * se(C12)
R3 Ø no(C10),dk(C11)
R3 Æ no(C11),dk(C10)
Which differs from wodan's 35 keycaps proposal by adding R1 ^ °, R1 § ½, and R1 ` ¬ ¦, plus having R1 2 " ² (for full DE coverage and consistency with the next keycap) instead of 2 " (which is thus missing from my UK kit), R1 3 § ³ instead of R1 3 § (which does not exist in any latin keyboard that I know of), E (standard ANSI) instead of E € (assuming everyone well knows € is AltGR+E anyway), and R2 M (standard ANSI) instead of M µ.
I hope some future day a designer would also add the following 7 keycaps, to provide for easier typing in spanish, portuguese and french by switching a few keycaps and using custom layouts (ca stands for the French Canadian layout):
R3 Ñ es(C10)
R3 Ç pt(C10), to be also used as es(C12), which the standard defines as R3 Ç }
R2 Ç ca(D12)
R3 È ca(C11)
R3 À ca(C12)
R4 Ù ca(B00)
R4 É ca(B10)
Hope this opens way to further discussion. Meanwhile, I guess non-British/German/Nordic users will have to stick to either ANSI US-International or ISO custom layouts (I plan to build mine on the UK layout based on keyboards+keycaps availability+costs consideration).