[DMcCunney]

Originally Posted by devilsadvocate

Yup, everything in *nix is a file. I thought (think?) that was the preferred method of keeping things simple/more efficient, but that will likely depend on specific usage.

It was a radical rethinking the the premise. In *nix, as you mention, everything is a file. A file is a high level abstraction. It can be created, opened, read from, written to, closed, and destroyed. It may be a collection of bytes on a disk, or a device like your terminal. It doesn't matter - you use the same semantics.

I remember a letter to one of the trade journals back when from an IBM systems programmer who read the early papers on Unix, went in to his boss, dropped copies on the bosses' desk, and said "How could we be so wrong for so long and nobody told us?"

...So I read up on this because, being a slacker, I use phpMyAdmin for MySQL and can choose which backend I use (also, I never paid attention). The choice was always MyISAM since it was always for small non-mission-critical stuff like a class project, my DC++ hub, etc. Oracle had always been my choice when it came to full-blown, monster-resources and damn the torpedoes RDBMS but they're quickly becoming Big Bad as a result of 1) letting OpenSolaris die without so much as a memo, and 2) suing Google over some maybe-or-not Java APIs in Android. I will be studying InnoDB on my own time in the future; thanks for pointing out the difference.

You're welcome.

Oracle also owns MySQL, as a result of acquiring Sun, and there was concern in MySQL circles about Oracle trying to kill it off. I'm not too worried: the folks who use MySQL in the first place wouldn't be likely to go with Oracle instead, for cost reasons if nothing else, and MySQL is open source. If Oracle tried to kill it, the most likely result would be a fork of the last open source code and a new product under a new name carrying on where it left off.

The problem is "What's an RDBMS?" Relational databases have their foundations in the work of Codd and Date, and ISAM predates that. Earlier databases on mainframes were hierarchal, not relational, and that described ISAM (and VSAM), as well as products like Cullinane's IDMS. Use of SQL as a query language does not an RDBMS make.

I'm in the beginning stages of learning this stuff myself. I just started wondering when I discovered the default file type in MySQL was ISAM how truly relational it was, and what InnoDB might offer that MyISAM doesn't. For the sorts of stuff MySQL gets used for, it doesn't seem to have made
a real difference, and for the real Big Mutha projects where you're managing terabytes of data, Oracle may be the only realistic solution, but it's still a worthy question. There are lots of database products out there. MySQL has the majority of the web based market, but you've got things like Postgresql and Firebird in the open source world, and there are still a few Oracle competitors out there like IBM's DB2, Sybase, and SQL Server.

...who also invented the C shell (csh), the vi editor, the chroot jail, the Network File System (NFS), and probably Skittles just to prove he had a sense of humor. The guy's bio should be required reading for any system designer or programmer.

Yep. I was never all that fond of csh (I preferred the Bourne/Korn shell), but I grew to actually like vi, once I understood the design concepts. Bill talked in an interview back when about trying to create a full screen editor that was usable over the 300 baud dial up modem connection he had from his dorm to the system he was working on, and a lot of vi makes more sense when viewed in that light - minimize what the user has to type to perform actions, and optimize screen updates for low bandwidth.

The Open Group now maintains it and sets the standard for the Single UNIX Specification, while SCO vanished into bankrupt obscurity as a lesson learned when you mess with the big boys, and Novell got in bed with Microsoft.

Yep. The Single Unix Specification is what Spec 1170 turned into.

SCO began as a second source for Microsoft's port of UNIX System III as Xenix, and became the main source when MS decided to wash it's hands off *nix and go entirely with home grown efforts. The growth of Linux savaged them: what they offered was a commercial X86 version of Unix System V, and why buy it when you could get Linux free?

SCO's moves against Linux were apparently prompted by a VC board member who told SCO CEO Darl McBride his real business was intellectual property, and he should go after Linux instead of trying to sell SCO UNIX. I think the hope was IBM would buy SCO, or pay them a lot to shut up and go away. IBM chose to do neither, had a large interest in Linux itself, and chose to fight. The issue isn't quite dead and buried, as the last legal formalities haven't finished playing out, but SCO's claims have been pretty much rejected across the board in court, and the VC board member dumped his holdings and left, so SCO is belatedly getting back to trying to sell and support SCO UNIX. They finally got a new major release out the door and are trying hard to preserve their installed base, but I doubt their long term prospects.

Novell pretty much had to get into bed with someone. Back when Noorda was running it, he acquired Word Perfect and Quattro Pro to compete with MS Word and Excel, got USL to provide an OS that could compete with NT, and already had Novell's core LAN offerings and expertise. He appeared to be trying to build a company that could compete with MS across the board, but never did meld his acquisitions into any sort of coherent whole or seem to have any long term plan about what he would do with them. Noorda got the boot, Corel got WP and Quattro, SCO got USL, and Novell was left defending a shrinking customer base and trying to reinvent itself.

As a famous rock band once said, you gotta give the people what they want.

Assuming you can figure out what that is. Too much these days seems to take the form of Microsoft's "Shove it down the market's throat and do lots of marketing to convince them they want it."
______
Dennis