Sunday, October 28, 2001 1:57:26 PM While I was interviewing to teach the Unix/Linux course, Ian Fisher asked me "Is Linux Unix?" Although I answered "Yes," the question stuck in my mind as I prepared the course materials. One thing that has bothered me is that apparently no company has submitted its Linux distribution to The Open Group for UNIX certification. The Open Group has made a statement that Linux is like UNIX, and at one time it extended an invitation to distributors to make a submission. Apparently, nobody stepped forward. Well, then, what about the UNIX source code, the distribution that descended from Bell Labs, whose owner should receive royalties from all that derive their operating systems from it? That would be the distribution currently possessed by Caldera. A study of the Caldera web page indicates the "one true UNIX," descended from System V Release 4, is alive and well. Lo and behold! Caldera is issuing the "one true UNIX" kernel with a personality package that allows all Linux applications to run on it! They call it Open Unix 8. What's in a trademark, especially if the trademark becomes a generic label? Has "UNIX" become a generic label? What if it isn't spelled in all capital letters? Has anyone sued a company for claiming its system is UNIX when it is not? Apparently so, early on. Hewlett-Packard did not sell UNIX, it sold HP-UX. DEC, may it rest in peace, did not sell UNIX. It sold OSF/1, Digital Unix, and Tru64 UNIX (OK, they did use the brand near the end. Talk to Compaq if you want these products.) IBM never sold UNIX, it sold AIX. Sun Microsystems never sold UNIX, it sold SunOS and Solaris. Silicon Graphics did not sell UNIX, it sold IRIX. U.C. Berkeley did not release UNIX, it released the "Berkeley Software Distribution." Actually, they did call it Unix for a while. 4.3BSD was a basis for many popular "Unix" operating systems in the 1980s and 1990s. The answer to "What is UNIX" only tells a tiny part of the story if we look to The Open Group brand name and trademark. Unixes have evolved, under license and out of license, ever since the beginning of time (January 1, 1970, for the uninitiated. See the Reference Manual entry for time(2)). All the major vendors have endorsed Linux. I was worried about Sun Microsystems, until I discovered that they were one of the founding members of Linux International. Apparently Linux development has not been in-house there, but Sun has cooperated with independent Linux developers. And of course, everyone is familiar with Sun's StarOffice product for Solaris, Windows, and Linux. Bill Gates had better be glancing over his shoulder now and then. What did all these operating systems have in common that allowed programmers and system administrators to call them "Unix?" What did that 12-year-old girl in Jurassic Park see when she exclaimed, "I know this -- it's a Unix system!" All the Unixes have been multi-user, multi-tasking operating systems. All of them have had a user interface descended from or similar to the Bourne shell. All the "Unix" operating systems I have known used a sequence of shell scripts to configure themselves on startup. Almost all of them have had graphical user interfaces since the late 1980s. All of them have had a fairly consistent set of commands, descended from "the programmer's workbench," that anyone could use. All of them have had system facilities that made the underlying hardware irrelevant for application programmers. This combination of features has been the core of what standards-makers throughout the history of UNIX have tried to codify. How far back have the standards-makers been trying? Since the beginning of time (q.v.). AT&T published the System V Interface Definition. ANSI published a standard for the C programming language. The IEEE published POSIX. Vendors joined together in the Open Software Foundation. X/Open, which became The Open Group, and somehow also obtained the right to enforce the UNIX trademark, published the X Portability Guide. The list goes on. Is it necessary to have the UNIX brand on your operating system for a programmer or a system administrator to recognize it? Of course not. They will recognize it because it has the core functionality of a UNIX system. They will recognize it because they can use it. How do you know a duck? If you look in Birds of San Francisco, you will find two species. Hawaii's Birds lists four species. Do I need to buy a book and look up a bird to know it's a duck? I don't think so. There's a popular quote, "If it walks like a duck, and quacks like a duck, it's a duck." The operating system I have been studying, Red Hat Linux 7.1, appears to me to walk and quack more like a Unix duck than some operating sytems that have been certified by The Open Group. When I look at Red Hat Linux 7.1, I get the feeling that girl in Jurassic Park expressed: "I know this -- it's a Unix system!"
Wednesday, October 10, 2001 2:14:35 AM It looks like Tim Mullen agrees with me about the IIS replacement idea, almost point for point.
Sunday, October 07, 2001 12:50:00 AM It had to happen sooner or later. I had been reading about the Nimda virus for two weeks, deciding not to upgrade my MicroSoft Internet Explorer because the patch was just too big for my phone line to handle conveniently. Then, last night, Norton AntiVirus noted, "you have a virus. Repair?" "OK," I clicked, and it was all over. At the next reboot I had no boot track to boot from. No disk label to show where the data was supposed to be. "No operating system," the laptop observed quietly. Reminding myself to remain calm, I took this event as a blessing that would teach me more about installing MS Windows 98. fdisk and format got me to the point where I could tell that the hard drive wasn't damaged. Then the CD that held the install image, the IBM "Product Recovery CDROM," mysteriously became unreadable. After the laptop booted from it, the install program would claim "invalid media in drive D." Oh, crime in Italy. The drive was neatly partitioned into two partitions, so I decided to exercise the system by installing Red Hat Linux 7.1 on the second partition. This was something I had just done the day before, and I was familiar with the procedure. After Red Hat was installed, I was explaining the situation to Margaret when she asked to see for herself. I was in a state, and it seemed every time I tried to boot from the IBM CD something different would happen. You know, the baby is never sick at the doctor's office, and the Windows 98 install ran perfectly while Margaret was watching. After installing a pair of operating systems, upgrading the MSIE didn't seem like such a big deal. It was done in 40 minutes. After that, several hours of reinstalling software packages from the net ensued. Data was reloaded from web servers and diskettes. I restored the bookmarks from a copy after several interesting attempts, and then threw them all away and reloaded them from an export image. I restored the address book, or part of it anyway, from an export image. There's a lot of aftershock when you reinstall a production system. Many of our files are gone, because we only backed up the most critical ones. A lot of files that weren't considered critical are turning out to be. License files. Dial-up configuration files. A well-formatted address book. Data that is only a memory. It's been 26 hours. I pictured all the consultants in the world who were restoring what they could for their clients, sometimes three or four clients per consultant, and I thought of all the downtime. I thought about all the little issues that would come up in the next ten days. Certainly I had planned to learn something different this weekend, but it looked like this experience was in my cards. Some good things have come out of this incident. I have learned that my antivirus program doesn't stop data corruption, it only detects when the data has already been corrupted. I have learned that even with partial backups, restoring the system is still a lengthy process. I now have a very clean installation and the latest MSIE. I appreciate how much Margaret cleaned up the hard drive last year. I have about 600 MB of free space I didn't have before. My modem connections run faster now, because I'm connecting to a different POP. I know for certain the modem is a WinModem. As they used to say on NYPD Blue, "be careful out there."
Return to David's Home Page
Your comments are welcomed by David R. Dull,
ddull@netcom.com.
(C) Copyright 2001, by David R. Dull.
All Rights Reserved.
