Necropost warning!
Heh, bored today, flipping through spreadsheets I remembered this one. Never got anywhere with it back then. Have tried a little correlation today, which has two interesting results:
- I think I've decoded numbers 1-8 (or 0-7, or some likely sequence) in column 9, and maybe some others
- The cipher may repeat every 16 characters(!)
It still seems to be a permutation cipher. Even given a mere 16 columns, that's 96! * 16 possible keys. Of course the permutation wouldn't be picked from truly any one possible, and presumably the permutation is merely rotated each step, if we are to believe the reference to Enigma. (Presumably then, there is one wheel with 96 positions, which moves through only 16 steps before resetting? Or at best two, the second one moving only one carry step.)
How does one write such a permutation? Is it enough to say column 1 is x[1] * p^1 (for plaintext x and permutation p), col 2 is x[2] * p^2, and so on? (Where "multiply by p" might be implemented by simply running the left-side operand through a lookup table.)
I'm a bit less sure about what the plaintext even is; the periodic (repeats every 10 lines) section seems
too periodic to be likely SPICE? Unless HSPICE allows component instances of the same name? Take these two groups for instance (lines 308-327):
!KVO*I2v$/D<%d:)X#;Y
8xafmQLh(/e*0.P)6
:2TOq+H2F0
o7cK-'B+5$u)*[10
o7k((E*Y*x8
_:=AHzX;,W6
:26/c((E*v9b8
_:]7*5!jq
$96UybV+m$B+L
T#2)il
!KVO*I2vi/D<%d:)X#;Y
8xafmQ,huj)90.P)6
:2TOq+H2F0
o7cK-'Bt(>R)*[10
o7k((EJEs=8
_:=AHzXvJ(6
:26/c((E*j*>8
_:]7*5!jq
$96UyjuIc2BjL
T#2)il
The per-char diff, 10 line spacing, is:
0 0 0 0 0 0 0 0 69 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 -32 0 77 59 -60 15 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 73 -13 26 -35 0 0 0 0 0
0 0 0 0 0 0 32 -20 73 -59 0
0 0 0 0 0 0 0 59 30 -47 0
0 0 0 0 0 0 0 0 0 -12 -15 -36 0
0 0 0 0 0 0 0 0 0
0 0 0 0 0 8 31 30 -10 14 0 63 0
0 0 0 0 0 0
If the statements are something like "Mnmos33a 42 47 0 0 MYNMOS", surely the "33a" or "42" or "47" have a lot more entropy? Yet the first columns with variation are the 6th and 7th. Which isn't terribly far from my example statement, that's true, but there doesn't seem to be enough characters remaining on the line. Most of all there are three
identical lines in this passage; either they're repeated component instances, or they're comments. But they have different starting characters.
The second part is line 23,
$$y<=-91-pXj+981:$u'(:!%-pXj+981:$,'(:/bJ5eU)*/2j/<2'kA25!j+(/2$pX#/tDA0
See it? Column 9 on, wrapped by 16:
-pXj+981:$u'(:!%
-pXj+981:$,'(:/b
It's not even full length either. S'pose it could be chance, if rather low. Though, flipping through a few other libraries I have, I do see what are most likely long strings of commented "#", encoded in repeating units. Doesn't seem to be any long repeats in this text unfortunately. I'm still not sure how helpful that would be, given that a known string still only makes a few trips through the permutation matrix -- I'd need a full 16x96 block of sequential characters to do that trivially.
Attached is a plot of frequencies; gray cells are unused codes, '1's are single occurrences, white cells are multiple occurrences, yellow are >10 and red are >40. (Codes are by row, 32 to 126; columns are ordered same as the ciphertext). The first two columns do indeed look very different, low entropy suggestive of comment ("* ") and line-extend ("+ ") patterns. The plot is left-heavy because there aren't many long lines. Interestingly there are 11 completely unused codes (across all columns): 32, 34, 38, 63, 64 and 123-126 do not show up in the ciphertext (codes below 32 aren't counted, and if 10/13 are any example, don't seem to be encoded). Additionally, 78 and 110 are only used twice.
Also saw a hint that ";" shouldn't show up in plaintext because of a bug in some past version of HSPICE, for whatever that's worth.
Incidentally, I think I originally had a use in mind for this (plaintext), but that's long since been forgotten. What's more, I also seem to have misplaced the original file, or headers, etc... So, pffbt, this is just for kicks.

Full ciphertext:
https://www.seventransistorlabs.com/HSPICE_Decode_Source.txtTim