Mad Dog 21/21: A Lion in the Sand
by Hesh Wiener
In November 2001, a team of French archeologists found the bones of a mummified lion in the Egyptian tomb of Maia, who had been the wet nurse of the pharaoh Tutankhamun, popularly known as King Tut. They believe the big cat was interred along with the little nurse about 1430 B.C. At first, they didn't let the cat out of the bag. But shortly after New Year's Day, at almost the precise moment Microsoft announced it was going to preserve support for Windows 98 for far longer than previously declared, the Egyptologists published their discovery in Nature.
The archeologists, Alain Zivie, Cecile Callou, and Anaick Samzun, pointed out that their lion was an extraordinary specimen. First, the male lion was huge, even by King of the Beasts standards. Second, it was very old when it died. Third, the lion had not been hunted down: There were strong indications that it had been a pet, or at least that it had been kept in captivity and treated with great, care, respect, and admiration. It was, perhaps, the Steve Ballmer of the Royal Menagerie, but more tame.
Microsoft usually hunts down and disposes of old software the way Egyptian royalty hunted lions. IBM usually acts the same way. But, like Microsoft, IBM from time to time relents.
During the past few months, for instance, IBM decided it would extend by about nine months, until this coming September, its willingness to sell z/OS 1.04, which is the last version of z/OS that will run on 31-bit mainframes. All subsequent versions require 64-bit hardware. The ultimate 31-bit version of z/OS will be supported until the end of the first quarter of 2007, giving the buyer about three years to run old systems with IBM's backing.
Similar policies and migration pressures exist for OS/400 users, who can only get IBM support for Version 5 and can still buy an upgrade to V5R2 from only the last release of Version 4. Users who don't move ahead will have their systems mummified, as far as standard IBM support is concerned.
Of course, Microsoft may not have to provide as much support for its older operating systems now that purloined excerpts of its source code, or at least portions of its source code, have found their way to pirate sites and file swapping systems on the Internet. The unauthorized publishing of this code (see our report on it in the Guild Companies Windows & Linux newsletter) rattled the whole computer industry, particularly any independent developers who might have previously gotten their hands on such code by various means.
Tomb it may concern: Archeologists in the big cat catacombs
There is a lot of Microsoft source code floating around outside of Microsoft, all of it supposedly governed by nondisclosure agreements, but nobody really believes that all of the people working with that code have fully honored the deals their companies signed.
What is clear is that this is the most visible dumping of such code into the laps of hackers and script kiddies, and that it will most likely bring about a big increase in the level of corporate paranoia at the world's biggest software company. At the moment, it doesn't look like the incident will affect Microsoft's plans to phase out older Windows programs or, alternatively, to provide additional support for such products.
The support situation is the same for every user or proprietary software, and not just in the case of operating systems. The only software where vendor support doesn't disappear after some arbitrary date is open source software, and that's because there is, in a sense, no vendor and, therefore, no vendor support. Users of open source software are, unless they contract with specialist support outfits, are on their own.
But being on your own with open source software that has its source code and documentation is not the same as being on your own with proprietary operating system or applications package.
First of all, open source stuff exposes its source code and includes documentation details that are kept private by commercial software suppliers, so right away any firm or individual that wants to support an open source program has complete information about the package. Commercial software vendors feel, with considerable justification, that it is necessary to keep a lot of information about their code under wraps.
Then, the open source development process includes public argument, making available not only a lot of information on how things work but also why they work the way they do. Commercial software vendors provide some extra information as part of their support programs, but quite a bit of the thinking that goes into the development process is kept secret because the vendors fear too much exposure of their internal effort would engender business risks.
Additionally, open source software packages don't move ahead because some vendor wants to churn the installed base to increase revenue. Commercial software packages may move ahead because a vendor really has made some substantial improvements, but that is a hard case to prove to the skeptical buying public. At the very least, this is an open issue and it is very possibly that cynics who say financial motives far outweigh technical goals in the product release cycle of commercial packages are right, or appear to be right.
When the software in question is an operating system and the vendor controls both hardware and software development cycles, there simply is no way to settle arguments about what announcements are carrots and what are sticks, even if both buyer and seller agree the newer versions of various products are superior to their predecessors.
The entire process is made all the more dramatic in appearance if not in fact by the marketing process. Vendors describe their new products the way religious folk in ancient Egypt may have described their deities. This doesn't seem to cause too many problems when one software deity is put aside in favor of a newer major release. The ancient Egyptians didn't act the same way. At various times different deities may have been more or less prominent, but, from our distant perspective, it does not appear that last year's favorite god was simply discarded when a new one caught everyone's attention.
Still, the spiritual aspect of cats, including big cats, did seem to vary in importance over the centuries. Lions were not merely admired. Their qualities of greatness were not merely idealized. No, lions and other cats were important in a number of ways, and so important that the Egyptians came to think about beings that combined human and leonine characteristics. Such a creature is called a sphinx, and the biggest statue of one that is still around, in a somewhat eroded state, sits with its human head and lion's body near the great pyramids.
The Great Sphinx: A creature in what was once the pride of an empire
Perhaps the sphinx inspired software companies, which have many of the characteristics of lions, to attempt to show a human face at times, particularly when they fear punishment in the marketplace or the courts for their most predatory behaviors. Microsoft, IBM, and their ilk could never persuade anyone they were pussycats, but they can, at times, be humane enough to preserve a satisfactory relationship with most customers.
This could, however, become more difficult in the future than it has been in the past.
Computing has reached a level of maturity that makes it less compelling for users to move quickly to the next great thing. The cost of migration, retraining, and other concomitants of progress can offset to a significant extent the benefits of that next generation of iron or code.
The impact of rising migration costs and reduced incremental technical advantage is not just on individuals or particular organizations. It is great enough to be an issue in whole economies. When the money spent to move ahead turns out to exceed the value of any gains or savings due to the forward migration, whole economic systems suffer from what might be called technology inflation. The megahertz and other measures of computing power grow but the results don't grow as much, just as larger nominal economies may, due to inflation, yield less actual wealth.
What might be needed here is some kind of option that is currently not available in the world of commercial software, one that would allow customers to stay back or more ahead in accord with actual conditions rather than the pace the vendors seek to enforce.
Suppose, for instance, copyrights on software became null when a vendor discontinued support. Further suppose any owner of a software copyright would be obliged to make public its source code when it lost its protective copyright, making it possible for the same mechanisms that provide support to open source software packages to operate in the realm of formerly commercial products.
Open source software has not prevented users from migrating to new releases of Linux, of various language processors, of database engines or of any other open source code. The very same situation ought to apply to commercial software. That is, users would move to new, copyrighted versions of various programs if the new packages provided acceptable value. And where vendors failed to provide enough incentive to move ahead and tried to force migration by killing support for old versions, users would have a somewhat easier time holding back until their software vendors sweetened the deal by technical or economic means.
Of course things don't have to move in this direction. It's entirely possible that all the support firms and software companies growing apace in Asia will fail to see any advantage in changing the rules of computing in favor of their future prosperity.