There are three Mr Boston.bin files floating around the net.

CRC:

bd4da4c2        8k.bin
f257151b        8k+tweak.bin
0110ea4d        10k.bin

MD5:

1c95cf8fcd2c98db747208ade464a44f  8k.bin
6fdfca6c084078100a05fabd7f3db78f  8k+tweak.bin
3cb1aa90bb145054cc8661609090fd77  10k.bin

==> 8k.hex <==
00000000: 67 20 47 43 45 20 31 39 38 33 80 01 41 f8 40 10  : g GCE 1983..A.@.
00000010: d0 4d 52 2e 20 42 4f 53 54 4f 4e 80 00 bd f1 8b  : .MR. BOSTON.....
00000020: 7f c8 25 7f c8 26 86 01 b7 c8 56 bd f1 92 bd f2  : ..%..&....V.....
00000030: a9 bd f3 54 ce 00 b5 bd f3 85 bd f1 af ce fe 76  : ...T...........v
00000040: bd f6 87 bd f1 aa bd f2 89 be c8 25 8c 01 2c 7d  : ...........%..,}
00000050: c8 56 26 d7 86 01 b7 c8 56 7f c8 25 7f c8 26 7e  : .V&.....V..%..&~
...
00001fe0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  : ................
00001ff0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  : ................


==> 8k+tweak.hex <==
00000000: 67 20 47 43 45 20 31 39 38 33 80 01 41 f8 40 10  : g GCE 1983..A.@.
00000010: d0 4d 52 2e 20 42 4f 53 54 4f 4e 80 00 bd 20 19  : .MR. BOSTON... .
00000020: 7f c8 25 7f c8 26 86 01 b7 c8 56 bd f1 92 bd f2  : ..%..&....V.....
00000030: a9 bd f3 54 ce 00 b5 bd f3 85 bd f1 af ce fe 76  : ...T...........v
00000040: bd f6 87 bd f1 aa bd f2 89 be c8 25 8c 01 2c 7d  : ...........%..,}
00000050: c8 56 26 d7 86 01 b7 c8 56 7f c8 25 7f c8 26 7e  : .V&.....V..%..&~
...
00001ff0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  : ................
00002000: bd f1 aa 86 ff 97 03 cc 01 00 dd 00 cc 98 7f 97  : ................
00002010: 0b d7 04 bd f3 54 7e f1 92 bd f1 64 8d e2 7e f2  : .....T~....d..~.
00002020: 72                                               : r


==> 10k.hex <==
00000000: 67 20 47 43 45 20 31 39 38 33 80 01 41 f8 40 10  : g GCE 1983..A.@.
00000010: d0 4d 52 2e 20 42 4f 53 54 4f 4e 80 00 bd 20 19  : .MR. BOSTON... .
00000020: 7f c8 25 7f c8 26 86 01 b7 c8 56 bd f1 92 bd f2  : ..%..&....V.....
00000030: a9 bd f3 54 ce 00 b5 bd f3 85 bd f1 af ce fe 76  : ...T...........v
00000040: bd f6 87 bd f1 aa bd f2 89 be c8 25 8c 01 2c 7d  : ...........%..,}
00000050: c8 56 26 d7 86 01 b7 c8 56 7f c8 25 7f c8 26 7e  : .V&.....V..%..&~
...
00001ff0: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  : ................
00002000: bd f1 aa 86 ff 97 03 cc 01 00 dd 00 cc 98 7f 97  : ................
00002010: 0b d7 04 bd f3 54 7e f1 92 bd f1 64 8d e2 7e f2  : .....T~....d..~.
00002020: 72 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  : r...............
00002030: 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00 00  : ................


The original rom appears to be "8k.bin" at 0x2000 bytes.

8k+tweak is the same as 8k, except that the "Jsr Init" in the header is patched
out and 0x21 bytes of replacement code rather strangely added at 0x2000 (rather
than finding an unused area in the initial ROM such as 0x1FDF) making it 0x2021
bytes long.

The strange jump sequencing strongly suggests the rom was patched twice by
different people.  Presumably the first patch had JSR $2000 at 0x001D.

I'm fairly sure all the zeroes at the end of 8k.bin are padding and that the
patches could have been put in there.

Here's the modified code from 8k to 8k+tweak:

10c10
< L001D               JSR   Init_OS                 ; 001D: BD F1 8B
---
> L001D               JSR   $2019                   ; 001D: BD 20 19


; I believe this is an earlier patch:

L2000               JSR   DP_to_D0                ; 2000: BD F1 AA  set DP=D0
                    LDA   #$FF                    ; 2003: 86 FF
                    STA   <$03                    ; 2005: 97 03     [D003] = 0xFF
                    LDD   #$0100                  ; 2007: CC 01 00
                    STD   <$00                    ; 200A: DD 00     [D000:D0001] = 0x0100
                    LDD   #$987F                  ; 200C: CC 98 7F
                    STA   <$0B                    ; 200F: 97 0B     [D00B] = 0x98
                    STB   <$04                    ; 2011: D7 04     [D004] = 0x7F
                    JSR   Reset0Ref               ; 2013: BD F3 54
                    JMP   Wait_Recal              ; 2016: 7E F1 92  ... and return to our caller.

; The current patch enters here then calls the older patch above:

L2019               JSR   Init_OS_RAM             ; 2019: BD F1 64
                    BSR   $2000                   ; 201C: 8D E2     call the code above at L2000
                    JMP   Clear_Sound             ; 201E: 7E F2 72  ... and return to the initialisation code at 0x0020

10k is the same as 8k+tweak padded out to 0x2800 bytes

gtoal@linux:/tmp/cleansweep/crctest$ diff 8k.hex 8k+tweak.hex
2c2
< 00000010: d0 4d 52 2e 20 42 4f 53 54 4f 4e 80 00 bd f1 8b  : .MR. BOSTON.....
---
> 00000010: d0 4d 52 2e 20 42 4f 53 54 4f 4e 80 00 bd 20 19  : .MR. BOSTON... .
512a513,515
> 00002000: bd f1 aa 86 ff 97 03 cc 01 00 dd 00 cc 98 7f 97  : ................
> 00002010: 0b d7 04 bd f3 54 7e f1 92 bd f1 64 8d e2 7e f2  : .....T~....d..~.
> 00002020: 72