Hi Branden, G. Branden Robinson wrote on Tue, Mar 29, 2022 at 06:26:14AM +1100: > At 2022-03-28T12:11:32-0700, Larry McVoy wrote:
>> Just pick one, we picked GNU make, not my favorite by any stretch but >> it runs everywhere. Build the makefiles for that make and call it a >> day. > Some people will sign a thousand EULAs before touching one copylefted > byte. That doesn't describe Ingo, but it does describe some of the > people he works with. I corrected that misconception in an earlier message. > Not long ago, Robert Elz was carrying on over at the Austin Group > list[1] about how he was blockaded from implementing "realpath -e" in > NetBSD because reading GNU coreutils source code would contaminate his > eyeballs. The "viral" GPL would infect his eyeballs and then his brain > and then everything he wrote for the rest of his life would be under the > personal control of Richard M. Stallman. > [1] Issue #1457. > https://www.mail-archive.com/austin-group-l@opengroup.org/maillist.html Not sure which message exactly you are referring to, but i think it doesn't matter. If the essence of your caricaturing paraphrase more or less matches what Robert intended to say, that would indeed be a relatively extreme desire to adhere to a cleanroom approach. I call it extreme because ideas cannot be Copyrighted, only text can, so i don't fear reading glibc or coreutils code before writing my own code from scratch. If others want to be more cautious and apply a cleanroom approach - shrug, let them be happy. Then again, i rarely read glibc and coreutils code anyway because it feels relatively hard to read to me, typically being much more complicated than average BSD code and containing copious amounts of #ifdef on top of that. So black-box testing is usually a quicker way for me to figure out what GNU code does than reading it. Either way, i don't think cleanrooming is in any way related to the question we are discussing here, which is whether code should be written in portable ways where feasible. Yours, Ingo
