by Rik Farrow <rik@spirit.com>

Rik Farrow provides UNIX and Internet security consulting and training. He is the author of UNIX System Security and System Administrator's Guide to System V.

This special edition of ;login: focuses on Windows NT and the USENIX conferences that took place in Seattle in August. For some of you, the mere mention of NT may be reason enough to shelve or immediately place this magazine in the circular file. But I hope that what we have chronicled here will provide a better understanding of both Microsoft Corporation and its offspring, Windows NT.

The conferences were both close to overbooked ­ over 300 attendees at each. The focus on a single vendor's product, and the degree of commercialism in some presentations, were very unusual for a USENIX conference. Part of the problem is simply that NT is still the new kid on the block, even though NT, in some form, has been around since 1992.

I got what I wanted: more information about NT, particularly its security. I also learned more about what people are doing to integrate NT into their existing networks, the problems they are having, and how to solve those problems. These solutions appear in the summaries and online versions of the papers </publications/library/proceedings/usenix-nt97/> and <.../nt-sysadmin97>;.

Some of those solutions appear in more detail in this issue. Samba is one of the best UNIX/NT integration solutions, and there are two feature articles about Samba. Dave Korn's U/WIN and Softway's OpenNT provide programming interfaces to NT that will be familiar to any UNIX programmer. In addition, these feature articles helped me understand the different environment faced by programmers moving to NT from UNIX.

The conference also marked the first time I had come face to face with Microsoft, in the form of product managers, engineers, evangelists, and summer interns. I could tell that these people had pride in what they were doing ­ that they were quite similar to the rest of the software development community. But there was also an edge that was unfamiliar and puzzling to me.

Microsoft dominates the desktop software market and is striving for the server market as well. One Microsoft employee made no bones about this during the conferences, saying, "We want 100% of the market." Not only is this attitude disturbing; it also was proven to be dangerous for the computer giants of the seventies, IBM and DEC.

But there was another aspect of the Microsoft edge that intrigued me. There was a degree of defensiveness that bordered on fear. Why would the most successful software company in the word be afraid of anything? I found a reasonable answer in an unlikely place: American Heritage Magazine of Inventions and Technology (Fall 1997), which included an interview with Edward Tenner, a scientific historian and author of Why Things Bite Back: Technology and the Revenge of Unintended Consequences (ISBN 0-679-42563-2). Tenner said:

The difference between Intel and Microsoft and the older giants [AT&T and IBM] is that these organizations have much shorter product cycles, and they have done far less basic research. They are more sensitive to tremors in the marketplace, more vulnerable potentially to competitors.

Although many people were amazed at the swift change from a position that ignored the Internet to one that embraced it, Microsoft's quick turnaround strongly affirms Tenner's statement. AT&T could spend years designing a telephone that would then have a product life a generation long. And both AT&T and IBM leased their products. They were not relying on a steady stream of sales, but on income from leases.

We learned a lot about NT during the conferences, and Microsoft had a chance to learn from us. The need for command-line interfaces for NT rang particularly loud and clear, to the obvious surprise of the product managers. To me, the goal is a two-way flow of information ­ to learn about features of NT that are unique and useful and for Microsoft to learn from the many years of experience of programmers and system administrators attending the conferences.

NT is an undocumented operating system, according not only to Microsoft employees, but to source licensees as well. This is because it changes too quickly for documentation to keep up with it. In reality, this means that the interfaces to the operating system, its networking, and authentication mechanisms are secrets. I do not find this comforting, but I am cheered by the efforts of programmers, like the Samba team, which unveil the hidden aspects of NT and help to integrate it with the rest of the computing world.

The conferences revealed more than anything that there is a dearth of detailed knowledge about a major operating system, NT. Microsoft-certified engineers understand GUIs and procedures ­ they do not have the depth to solve the really difficult problems. Those solutions will come from the members of communities such as USENIX.