Google Goes Back 20
This past Monday morning I was talking with my pal Michael Stern, the part-time hacker and full-time CEO of Information Markets Corp. Among other things, Michael shared his relief that Google hadn't yet indexed the Usenet newsgroup archive back any farther than 1995.
Later that day I received an e-mail from a contact at Google giving me a heads-up about the next day's announcement that their newsgroup archive now goes back to 1981. I'm sure Michael isn't the only one who cringed when he heard the news.
How about Bill Gates? Turns out the very first mention of Microsoft is this one, posted to net.general:
The June issue of BYTE magazine has a fairly long article on XENIX by Microsoft's XENIX product manager. Mostly, it's a standard "What's a UNIX" paper, but it also describes some of the enhancements they are adding to V7. The most important is support; additionally, they are going to add a fair amount of hardware error recovery (bad block handling, parity and power fail interrupts, etc.), as well as record handling, shared data segments, synchronous writing, improved interprocess communications, networking, and languages: Pascal, BASIC, FORTRAN, and COBOL.
Of the thirteen posts mentioning Microsoft in 1981, most involved UNIX. One post to fa.unix-wizards almost poignantly points out the problem with closed-source code:
You CAN'T in general assume people can go look at the source to figure out what an error means. Aside from the obvious fact that there are lots of users who don't speak C, recall that to get Unix a (non-Bell) site has to sign in blood a promise not to give out the sources to anyone. Many sites (especially University Computer Centers, who don't trust their users) protect /usr/src. (I know of one comp ctr at an unnamed major university in Berkeley that doesn't even make df suid because "It's none of your business if the disks are full". Since they also overbook disk quotas, which they have, this is a fun place to try to get any work done.) An even worse case is typified by Onyx, Microsoft, and other turnkey Unix system sources who don't let you have ANY sources. This is how a small business can afford to buy Unix - to not pay for a full source license. These poor guys can't go look at the source, and I can't tell you how many times I've wished I could in order to pinpoint some bug.
Other goodies in the archive include:
So what else is back there? You tell us. If they're fun, interesting and publishable, I'd like to run some of them in the UpFront section of the next issue of Linux Journal (with credit where due, of course).
Doc Searls is senior editor of Linux Journal.