“Imagine being given the job of designing a book that anyone could read but no one could copy,” George Trepal wrote in Ahoy! in October 1985. “Making software copyproof is a lot like this.” A computer has to read a program from its disk to run it, so anything the computer can read, a copy program can in principle read too. Disk protection looked for the places where that breaks down: marks on an original disk that a drive could read, or at least notice, but that an ordinary copy would get wrong.
Trepal knew the field from the inside. He mentioned “a few other tricks (including the one my company uses) that I’ve promised not to write about”, and warned that “people are reluctant to talk”. This entry follows what publishers did on the Commodore 64 and the Amiga, and the copying tools sold to undo it. The wider story, from tape loaders to manual look-ups, is in Copy Protection.
Errors on purpose
The first protected disks carried damage. “Virtually all copy protection takes advantage of the fact that the drive and computer hate anything abnormal,” Trepal explained. “Normal copy programs won’t copy errors because — well, because they’re errors and are therefore WRONG!” A publisher reformatted a track the program never used, or spoiled a sector’s checksum, and the game’s loader asked the drive for it. If the drive reported the error, the disk was an original; a clean copy failed. “The error is always confined to a portion of the disk not used to hold program information,” Trepal added, because a program with an error in it could not be loaded at all.
Inside Commodore DOS (1984) put it more bluntly: “disk protection involves the deliberate corruption of a given track or sector.” Its authors ranked the errors they found on protected 1541 disks. A whole track with no sync mark (error 21) and a single sector with a bad checksum (error 23) were the commonest, and a partly formatted track was “the new kid on the block”. The loader that checked for them was the real target, so it was hidden “through an autostart feature, the use of unimplemented 6502 op codes, encryption, or compilation”.
The same idea reached the Amiga. Amiga Disk Drives Inside and Out noted in 1988 that “some older copy protection systems which are still in use” relied on blocks whose data was intact but whose checksum was wrong, so that AmigaDOS could not read them and “simple copy programs cannot duplicate these programs, or correct the checksum”.
Errors had a cost. When a 1541 met a bad sector it knocked its read head against the end stop to realign itself, which a user heard as “about a second of machine gun fire”. Protection read that way “involves a great deal of head banging, which can indeed cause head misalignment”, Trepal wrote: “All programs copy protected with block reads work to send your drive to the repair shop.”
Off the edge of the format
The 1541 made more ambitious tricks possible because it was, in Trepal’s words, “a full computer with ROM and RAM and a dull little filing program that runs constantly”. A program could send it new code and tell it to do things Commodore’s DOS never did. Trepal’s 1985 list covered the main ones:
- Extra tracks. A standard disk used 35 tracks, but the head could go further: “Track 36 is the easiest to work with and track 40 is the hardest. Since a normal disk copy program only copies 35 tracks, vital program information stored on track 36 is lost.”
- Half-tracks. The 1541’s stepper motor moves the head half a track at a time, so data could be written between the normal tracks. The head writes too wide a path for this to work cleanly; Inside Commodore DOS found “the magnetic path produced by the read/write head is just too wide”, and Trepal noted that “the half-tracked track tends to destroy one of the tracks it’s between. But that’s taken into account.” A copy made on a well-aligned drive missed the half-track; on an ordinary drive “the head wiggles and picks up some information from the normal track and some from the half track”.
- Renumbered tracks. “Two tracks 34’s for example.”
- Missing sync marks. Each sector starts with a sync mark. Without one, “a normal copy program can never find where it is and hangs up.”
- Extra sectors hidden after the last sector of a track.
- Changed bit density. The 1541 records its outer tracks at a higher bit rate than its inner ones, and “by altering the bit density of a track it can be made unreadable”.
Two ideas were still ahead of practice in 1985. Writing a sector “with a magnetic strength halfway between a 1 and a 0” would make it read differently each time, so a loader could read it several times and treat changing results as proof of an original: in Trepal’s words, “a technique that has been proposed”. Burning a tiny hole in the disk with a laser was “really a technique for hard sectored disks”, not for Commodore’s soft-sectored ones. By the late 1980s the protection systems had names, and the copiers advertised which ones they could beat: in 1988 the Renegade package was sold as “Capable of Archiving RapidLok Protection”.
The copier trade
Every trick had a market on the other side, and it predated the Commodore 64. In July 1982 MICRO listed Lock-It-Up 4.1 for the Apple II at $195, “a sophisticated copy-protection system including over thirty state of the art protection features”, which claimed to beat “any standard or ‘nibble’ copiers including Locksmith 4.1”. On the IBM PC, a company renting out copies of WordStar agreed in 1984 to protect every rental disk with PROLOK.
Trepal sorted the copiers into two kinds. Block copiers read a disk one block at a time; nibble copiers read and rewrote whole tracks, which meant they did not “have to worry about such things as synch characters” and were “harder to protect against”. A copier written in BASIC took about 25 minutes, a machine-code nibble copy “(including half-tracks)” about eight, and fast copiers about three. Both sides advertised in the same magazine. The issue of Ahoy! that carried his article also sold Disk-Lock, “a professional copy protection system” that could “place protection on any track from 1 to 40” for $39.95, and, from Central Point Software, Copy II 64, also $39.95, which “copies half and full tracks” and “even protected disks in just 2 minutes”. Central Point already sold Copy II Plus for the Apple II, Copy II PC and Copy II Mac.
When protection outran the nibblers, parameter copiers stopped trying to reproduce it. Kracker Jax, advertised in December 1986 with the line “ARE YOU CAUGHT UP IN THE COPY PROTECTION ARMS RACE?”, sold volumes of “parameters”, each a small program for one named game. Ahoy!’s reviewer explained in January 1987 that “the working commands in a parameter file are no more than a series of block reads and writes which change specific bytes of the program disk”. With the protection removed, the disk could be copied normally, and put on a hard drive, which the reviewer noted “original copy protected programs” generally could not.
Amiga and Atari ST: disks made by machines
The 16-bit machines changed where protection came from. Commercial disks were made on duplicating machines, such as those from Trace, which wrote whole tracks from a description of the disk and could write “information that could not be reproduced on the target machine”, as Jean Louis-Guerin put it in his documentation of the preservation format IPF. A home drive wrote at one fixed rate; a duplicator could write anything its description asked for.
In December 1986 Dave Devenport, programmer of the QuickNibble copier, told Amazing Computing that “the most common protection I’ve seen on the Amiga is what I call a ‘sync track’, where they use a non-standard index sync, and write an extra-long track. The Amiga can read twice the density that it can write.” A long track holds more data than a drive writing at its normal rate can fit into one turn of the disk, so a copy made on an Amiga could never hold it all. Rob Northen’s Copylock, first used on the ST in 1987, turned the same asymmetry into a product: one sector written with a different bit timing, which a drive could read but not write, and a check that timed the read. Its keydisks each carried a serial number, which the game could use as data. The preservation format IPF has its own entries for other named Amiga schemes of the time, including Amiga versions of Speedlock and a system credited to Adam Brierley. The Speedlock name is better known from Spectrum tapes; see Copy Protection.
The copier makers answered with hardware. In January 1990 an advertisement for Synchro Express, a $69.99 add-on for the Amiga and ST, conceded that “some programs now have very sophisticated protection including long tracks, Incryption, etc.” and made the duplicators’ argument its own: “whatever the programer comes up with, the disk then has to be commercially duplicated”. Synchro Express, it claimed, “transfers data at MFM/TTL level, just like the duplicators”.
Protection also made disks fragile in a new way. The SCA virus, one of the first Amiga boot-block viruses, copied itself onto whatever disk was in the drive, and Amiga Disk Drives Inside and Out warned that “it was possible for it to destroy the important copy protection information on a commercial disk, therefore making the disk unusable”.
Preservation
The same features that stopped copying stop ordinary disk images. A standard image stores the decoded sectors a normal drive would return, not the track lengths, bit timings and deliberate errors the protection checks. The Software Preservation Society’s IPF format was designed to record a disk as it was mastered, “the disk data, how it is stored (and hence how to write it back)”, read with low-level tools the society developed, published through the KryoFlux project. Its format documentation lists density types for Copylock, Speedlock and others, and “fuzzy” data blocks for the bits that deliberately read differently each time, the technique that was only “proposed” in 1985. An IPF image can be written back to a floppy or loaded into an emulator such as WinUAE, so a protected original can run as it was sold.
Sources
- George Trepal, “Copy Protection: The Guarded Secrets”, Ahoy! issue 22, October 1985, pp. 58–60.
- Disk-Lock and Central Point Software (Copy II 64) advertisements, Ahoy! issue 22, October 1985.
- Richard Immers and Gerald G. Neufeld, Inside Commodore DOS (Datamost, 1984), pp. 122–123 (protection) and p. 199 (half-tracks).
- New software listing, MICRO issue 50, July 1982: Lock-It-Up 4.1.
- “More on loaning programs”, The Stack, November 1984 (reprinted from Small Computers in Libraries, April 1984): PROLOK.
- Computer Mart, Kracker Jax advertisement, COMPUTE!’s Gazette, December 1986.
- Review of Kracker Jax, Ahoy! issue 37, January 1987, pp. 67–68.
- Renegade advertisement, Ahoy! issue 54, June 1988.
- Show report, Amazing Computing vol. 1 no. 9, December 1986, p. 106: QuickNibble and Dave Devenport.
- Grote, Gelfand and Abraham, Amiga Disk Drives Inside and Out (Abacus, 1988), pp. 128 and 179.
- Synchro Express advertisement, AmigaWorld, January 1990, p. 85.
- Codetapper, “An interview with Rob Northen” (web, secondary): http://www.codetapper.com/amiga/interviews/rob-northen/
- Jean Louis-Guerin, “Interchangeable Preservation Format (IPF) Documentation”, version 1.6, April 2018 (web, secondary): https://www.kryoflux.com/download/ipf_documentation_v1.6.pdf