Tuesday, July 20, 2004

Telirati Newsletter #45

Overall, I find it eerie to reread what I woute five years ago. For an industry that supposedly moves at lightning speed, the predictions presented below are remarkable mainly for the fact that their outcomes are still unfolding.

Still, it is possible to tell that most of these predictions have come true. Moreover, it is also possible to say quite clearly that Microosft has failed to use the previous five years to take actions that would have been effective in minimizing the competitive threat from Linux and other open source software.

One can also see that the importance assigned to these threats has held up: Price hasn't dislodged Windows from the desktop, where Windows pricing is low enough to fail to deter any customers. It is the other attactions of open source software that matter more.

Telirati Newsletter #45: Categorizing the Linux Threat (Opportunity)

How much of a threat is Linux to Microsoft? What are the main opportunities for Linux? In answering these questions, it is best to keep in mind that there are layers to the answer, and some layers matter more than others even though the theories behind each layer are supportable. Here, we attempt to order the layers according to importance:

Big threat: Linux is good enough. Linux is good enough for a lot of purposes. Serving simple Web content over moderate-speed links is simple, and Linux is more than good enough for this purpose. Inexpensive desktop PCs, or used PCs, are good enough for this purpose. If you need to put content on the Web, Linux is good enough. As broadband access explodes, a lot more people will find low-end PCs in combination with Linux ideal for this purpose. The vast majority of computers and the OSs they run are not applied to anything remotely approaching their highest and best use nor to the limits of their capacity. It should worry Microsoft that good enough might be good enough for most people.

Big threat: It's better than stealing. A lot of people got into computer software for the same reason people become drug pushers: they don't want to pay for their own habit. But not everyone can cadge free goodies. Linux, in addition to being good enough is free or near to it. As makers of commercial software squeeze out freeloaders and thieves, these people will have to make a decision on how to go legit. In this, Linux has the distinct advantage of not having a commercial entity associated with it that can be embarrassed by this market.

Big threat: It's better than cadavers. Students can cut Linux open and start hacking away without having to be part of the select few organizations with source access to Windows OSs. If students come to be familiar with the anatomy of Linux more than Windows, Windows will lose a key slice of mind share among the most fertile minds.

Moderate threat: IBM likes it. And so do all the other also-ran software makers. The problem is that there are a lot of these guys out there and they might get it together enough to add significant momentum to Linux as a platform that supports a full suite of software. And don't get all worked up about how IBM has higher software revenues than Microsoft if that is still true. Computer Associates is a very large software company, number two or three by most measures, and I defy anyone to tell me how that might be relevant to this discussion.

Moderate threat: Responsiveness. One of the most visible fruits of community development is multi-platform portability for Linux. It runs on the x86 platform, of course. And Alpha, so that it covers all the ground NT covers. Now add PowerPC, MIPS, and a number of miscellaneous platforms. The reason this isn't a large advantage is that the x86 architecture is so dominant. But it does draw in all the makers of platforms Microsoft doesn't have time for. To counter this threat, Microsoft has to reduce the cost of maintaining NT and it's successors on a variety of platforms, which is right now in the tens of millions of dollars per platform per version.

Moderate threat: The vast hordes of Linux coders. Only a small minority of these coders matter. If Microsoft can reform its culture, and there are some indications it will try, Microsoft can more than meet the challenge of open source development.

Moderate threat: Open source. Important, but vitally important to only a small minority of users. If you are running a defense organization outside the U.S., you would want an OS you can inspect for the presence of exploitable features. This is no joke. It has been reported that the Chinese government will not use Pentium IIIs because of the PSN and related features. Microsoft can blow it here by incorporating content protection features into Windows that will create the same kinds of security concerns that Intel inflicted on itself in the Pentium III. If this does happen, more people will justifiably ask: is my computer capable of betraying me?

Moderate threat: Stability. Only a small minority of people leave their computers on all the time. Microsoft has an opportunity, with Windows 2000, to meet the threat of Linux stability, which, in no small part comes from the fact that most applications of Linux are exceedingly simple. Still, it is not outrageous to suggest that a Unix-like OS would be more stable than Windows 98, and Microsoft will have to show it can make systems that inspire confidence before it becomes commonplace to leave one's computer in charge of one's house.

Not much of a threat: User interface. The Linux community is only now discussing user interface in something other than dismissive tones. Even now, various attempts at a user interface are regarded as interchangeable, and the discussions in most Linux fora still treat the topic as if it were a check-off item, like USB support. Wrong. The OS exists mostly to support the user interface, and the theory and architecture of the user interface are the most difficult aspects of OS design. Windows 98 is manifest proof of this. Apple understands this (and may produce the first Unix-derived OS that embodies such an understanding), and so does Be. Unless Microsoft goes on the long march to Bob II, Linux won't catch up. More likely, Microsoft will come up with at least a few interesting UI innovations while Linux is still learning the basics.

Not much of a threat: Price. Other than the aforementioned freeloaders and brigands, the price of a Windows OS has never been a barrier for any important customer that can move a significant number of machines. PC makers have common cause with Microsoft in maintaining price levels and margins, and, when they shift, to adapt to new levels by segmenting the market and pricing appropriately. Consumer preference is far more important than the difference between the cost of a Windows OS license and zero.

Copyright 1999 Zigurd Mednieks. May be reproduced and redistributed with attribution and this notice intact.