I know I might be disrespecting a sacred cow on this one and I'm thirty years late to the party anyway... but has anyone actually managed to use the ABC/CDE packets compiled by William Yu back in the days?
The Basic Source Repository fails to display most packets, and The Official(r) ABC Express Reader(tm) ("Accept no substitutes!") programmed by Yu himself produces garbled data more often than not. I got so frustrated that I delved into the files and noticed that the companion files (IDX/UPD) are not only utterly unjustified (as they can be rebuilt from the ABC/CDE packets) but the packets themselves (ABC/CDE) often includes faulty data that makes chunk localization hazardous.
I found an old text file from 1995 in which he claims: "Number of bytes - number of lines of code = bytes in packet," which is "usually" correct. Unfortunately, since chunks are stored contiguously and the only information provided is their size, getting one entry wrong will automatically desync all subsequent chunks. For instance, in ABC1995.CDE, the metadata for MAZEGEN.BAS declares 294 lines and a source file of 11,835 bytes, implying the packed file is 11,541 bytes long. However, the actual size in the packet is 11,538 bytes. That's the 9th chunk, meaning the following 282 chunks become inaccessible because we're 3 bytes off.
There is an overkill way to retrieve the individual chunks but I suspect it wasn't intended to work that way. Seriously, was this thing always broken or did it work at some point? When I see ABC packets mentioned, it's only in a positive light. I feel like I'm taking crazy pills.
Oh, and a Happy New Year to everybody!