Date: Wed, 24 Dec 2003 12:55:57 -0800 (PST) From: "Bruce R. Montague" <brucem@mail.cruzio.com> To: chat@freebsd.org Subject: Re: A bit of trivia: what does usr stand for? Message-ID: <200312242055.hBOKtvcr000327@mail.cruzio.com>
next in thread | raw e-mail | index | archive | help
Regarding: > ... reference creat(2). ... > > I think it was Kerningham that said that given the chance > to do it all over, he would have called it create(2). I used (and "operated") IBM 1403 printers, teletypes, and various other impact printers of the same era, drum printers, band printers, etc.. The 1403 was arguably the best, and most widely used, mainframe printer of them all. Output from other companies machines was often printed on 1403s (via tape transfer). Invariably the print slugs that printed characters on all impact printers (and teletypes) wore out unevenly, reflecting frequency of character use. If used for normal text, the letter "e" was always the first to go (reflecting its well-known English frequency). The highly-used characters would start "fading" as their hammer/slug or equivalent began to lose its sharp relief. This was not a minor problem. An all-new print chain was like Christmas! Band printers and chain printers, such as the 1403, had multiple sets of the character set in their "chain". The 1403 chain allowed individual print slugs (characters) to be replaced (handloaded) when they became worn. I believe you could alter the frequency of the slugs to match expected use by using the "control tape", although this didn't seem really commonly done. The 1403 chain somewhat resembled a machine-gun ammo chain or bike chain. Anyhow, programmers often tried to reduce the use of common characters to avoid wearing out that character. This sometimes was designed deeply into things. For instance, the use of "." for the indirection operator in the BLISS language (equivalent to "*" in C, but much more heavily used because all variables were "pointers") was done for this reason (I heard Bill Wulf say once that's why he did it). So it is possible this common practice created a bias against "e" in Unix names. - bruce
Want to link to this message? Use this URL: <https://mail-archive.FreeBSD.org/cgi/mid.cgi?200312242055.hBOKtvcr000327>