@AC re. Joke
No. I was serious and yes, I know that 8 is 010 and 9 is 011.
I was presuming that the person was using a calculator which worked in octal and decimal (and thus had 8 and 9 and point keys) but which was in octal mode, so that when they were typing in something like 18.49, the calculator actually registered 14 (neither the 8, the 9 or the decimal point would have registered). Would get the sums very wrong.
If you had actually bothered to think of the mechanics of it, you would have understood.
By the way. I think that your floating point octal to decimal is incorrect.
When writing non integer octals to one significant digit, the numbering would be
0.1 octal, which is 1/8 (0.125 decimal)
0.2 octal, which is 2/8 (0.25 decimal)
0.3 octal, which is 3/8 (0.375 decimal)
0.4 octal, which is 4/8 (0.5 decimal)
0.5 octal, which is 5/8 (0.675 decimal)
0.6 octal, which is 6/8 (0.75 decimal)
0.7 octal, which is 7/8 (0.875 decimal)
1.0, which is 8/8 (1.0 decimal)
So in Octal 0.5 + 0.2 + 0.1 will equal 1.0, which it needs to do in order for arithmetic to work.
The first significant digit after the octal point (geddit) is 1/8th's, the second is 1/64ths, the third is 1/512th's and so on.
This means that by casual inspection, 0.44 octal HAS to be larger than 0.5 decimal.
By my calculations 12.44 (octal) is (1x8) plus (2x1) plus (4/8) plus (4/64), which makes it 10.5625 (decimal) or 10.56 rounded to decimal pence.
I can't see how you got 10.14. Even if you had worked in pence, 1244 octal is 676 decimal.
You got the 0.95 decimal correct, however.
You could do the exact arithmetic if you worked in pence or cents. Non-integer arithmetic in any base other than 10 hurts my head.