Andrew Schulman, senior editor at O'Reilly & Associates, has a history of embarrassing Microsoft. He us the author of "Undocumented Windows" in which he suggested there are hidden calls that potentially slow down rival ISVs in developing Windows applications, and "Unauthorised Windows 95" where he took the lid off the fact that DOS lives on in the OS. Now, he has published an article that suggests Windows NT Workstation and Server are virtually the same product. PCDN spoke to him last night about Microsoft, NT marketing, the Web and the law.
A lot of the reaction seems to have been `how come nobody saw this before'. Why was that?
Well, people had noticed it, they just hadn't drawn the full conclusions. If you do an Alta Vista Web search, sure enough you'll find a couple of articles on using NT Workstation as a Web server that says it works just like NT Server. After the article was published on Friday I had a couple of e-mails saying `we already knew that'. It was even more similar in 3.51. I can only presume they put in the override to stop someone finding out about this.
They have been saying that NT Workstation is not suitable to run a Web server. They had actually put a piece of code in to prevent you having a certain number of TCP/IP connections, claiming that that's put in to protect the customer because it's not going to work well. If you have [your Web server] set up as Workstation that's not the case. The difference is just some settings and its not clear that any of those are beneficial to a Web server. It would be beneficial if you have it on something better than a T1 line but most people aren't using that. There are very minimal differences.
Your problem is that they haven't been straight in their marketing of the two products.
If they want to come out and say 'Look, we want to set conditions on the use', fine, but you have to tell people what they're getting. To third-party vendors like Netscape or O'Reilly they say NT Workstation is not capable of doing things. That's the basic problem. They want you to buy NT Server and then you have to use IIS and the other Microsoft Internet products. They had [charts showing] this difference based on whether you were running a LAN server. The problem is they're talking about LAN servers not Web servers, and really overlooked that they're completely different. It shows up a real ignorance in their marketing people.
The bottom line is that they wanted to charge more. Novell for a number of years charged per different numbers of users and that has always been legitimate, but on the Web can you charge people for every hit they have? Then you're getting into what we call the `Web tax' and that doesn't seem fair. You wouldn't put restrictions aren't things that aren't even yours like TCP/IP, HTTP and sockets. Some of these things were developed using US government money.
Have you heard back from Microsoft?
Not from Microsoft, no. I mean, I know people there. Hoping to hear something soon. At O'Reilly we have some contacts at Microsoft and they had said when all the fuss about the 10-user limit was about that in 3.51 the changes are minimal but in 4.0 they `re going to be completely different kernels and so on. I guess they'll respond tomorrow or hope it'll blow over. They'll get somebody to write something.
O'Reilly is collecting something of a reputations a Microsoft basher. Is that fair?
We're a technical publisher. We have software such as Web Site, our server, but primarily we're known for publishing books and this is just part of our role as a publisher.
This isn't the first time Microsoft has been accused of abusing its power. What's your feeling about how Microsoft should be handled?
I certainly don't have the libertarian `hands off' position. On the other hand I don't think `kill Microsoft'. I'd just like Microsoft to change the way it does business, be more honest. I don't even think it's necessary for them to act that way. If they had a level playing field they'd still do fine.
OK, so that's the fantasy. I think in the US, given the importance of software in the economy, there's really got to be some regulation of some sort. If Microsoft can't wake up and say: `Gee, it really would be better to do this', then someone should make them.