> Intentionally or not, Phil Katz achieved "embrace, extend, extinguish" with the .ARC format.

I recall a text file circulating BBSes when I was a kid, around the time of the move from PKArc and PKXArc, to PKZip, supposedly detailing the visit of the lawyer to Phil Katz's house.

IIRC, the file said that some evil corporate lawyer came to PK's house, PK was out, his mom didn't know why the lawyer was there, served him tea or lunch or something while they chatted, and then at the end, evil lawyer surprises the mom or PK with evil legal papers.

At some point, PKZip appeared, and then downloads on BBSes started switching from `.ARC` to `.ZIP`. (There were a bunch of other packaging&compression formats, and you'd have a collection of the tools to extract whatever the download you found and got was in. But, IIRC, for awhile, `.ARC` was the most popular on free BBSes that I used, and then `.ZIP` became the most popular.)

I don't know whether the stories of the legal action had anything to do with the move from .ARC to .ZIP, but I could believe it, on the free BBSes. They were generally run by hobbyists who could afford the computer and a phone line, and the time to give the BBS a character, and I got the impression that just wanted to make something and share. As far as I know, there wasn't any implicit subtle self-promotion, career development, etc., like there is in many things today, and a lot of corporate behavior that we consider normal within tech today, I think wouldn't have been considered acceptable. So, especially given the stories circulating about ARC vs. ZIP, it's really easy to imagine a "f-word those SEA guys" reaction from the altruists running BBSes in their dens.

> I don't know whether the stories of the legal action had anything to do with the move from .ARC to .ZIP, but I could believe it, on the free BBSes.

That's my very clear memory of what happened: PC Magazine and other trade rags reported that people were so pissed off at SEA that they switched en masse to .ZIP; it became a network effect, in that very soon everyone was using .ZIP because everyone was using .ZIP.

I was a junior IP lawyer back then; it was a real eye-opener that if a company asserted its IP rights in what was perceived as a bullying way, the resulting PR blowback could destroy the company itself as people voted with their dollars. That was decidedly not what new IP lawyers were taught; much of the IP world regarded (and still regards) IP rights as virtually sacrosanct, to be prioritized over practically all other values.

SEA told that the PKware source code still had their spelling mistakes in the comments, which was part of why they were a bit indignant by the whole thing.. But this is a story about one mom-pop biz vs another, but that didn't really come to light until years after.

> SEA told that the PKware source code still had their spelling mistakes in the comments

That's like the Apple v. Franklin Computer case from the early 1980s: Franklin made a computer with an OS that it said was "compatible" with the Apple II; Franklin claimed to have independently developed its own OS, but it turned out that the Franklin code included not only what must have been thousands of identical lines as the Apple code, but also the name of one of the Apple programmers, James Huston, as well as the name "Applesoft." [0] (If memory serves, I read somewhere that the names were embedded as values of no-op string variables so that they wouldn't be stripped out during compilation, as comments would be.)

[0] Apple Computer, Inc. v. Franklin Computer Corp., 714 F. 2d 1240, 1245 (3d Cir. 1983), https://scholar.google.com/scholar_case?case=100632041256965...

Elsevier will likely find this out the hard way one of these days.

To add to the recollections, this was also the time that "look and feel" lawsuits were going around -- Lotus suing Paperback Software, Apple suing Microsoft, etc. Most of the BBS community was against look-and-feel and patent lawsuits ("The right to code"), and there was a court injunction preventing the parties from discussing the sea/pk lawsuit. But the pk side leaked, so that was the only side that was heard (and I remember believing at the time that sea was the "bad guys", until I learned a bit more of the nuances of the case, and now think that both sides were both right and wrong).

So that was a major driver for the "arc to zip" conversion, just like the "burn all GIFs" day.

There were a number of utilities that would convert from arc to zip and that really sped up the conversion - and the zip compression made significant differences in the transfer time so it was worth the effort to convert just for that reason. Some BBS software (or doors) would convert uploads from arc, sea, zoo etc to zip on upload. Between the two the move away from arc happened more or less overnight.

One if the failings of zopfli (IMO) is that there isn’t a utility to walk a directory and find embedded deflate streams and decompress them and there isn’t any indication that zopfli has touched a file so it is a brute force search to try and recon press a file. I’m also surprised that deflate with 64k windows never caught on even though zip has supported it for years and it’s a really easy extension to the existing format that just uses a reserved bit.

Dean Cooper's post on the PkARC vs SEA topic is interesting. He was another person working on an archiver in the same time period. http://ascii.textfiles.com/archives/2398

A while back I emailed Thom Henderson for the ARC source code and he sent it to me. I planned to port it to Python for kicks but never got around to it...

Isn't Python a poor choice of language for computationally intensive code?

I just wanted to do it to learn how the code works.

According to SEA founders, he had ripped the source code too, having intact variable names and such. Apparently, first version of PKArc was pretty much ARC with assembly optimizations added.