Microsoft had revenues of $8.6 billion for the fiscal year ending June 1996. Not bad for a company Gates and Paul Allen formed in 1975.

On December 12, 1996, Gates' Microsoft stock holdings totaled 282,217,980 shares. Multiply that by $116 -- the late afternoon price on April 14 -- and his net worth on Microsoft stock alone is $32.7 billion.

Several facets of Gates' Microsoft have always impressed me personally:

• The way he organizes his people. Microsoft hires smart people, puts them into small teams, gives them a very clear vision of what he wants them to do and then empowers them to get things done. Despite its growth, I still find Microsoft easier to deal with than many of the much smaller companies in our industry, and certainly a lot easier to deal with than its direct competitors, like Novell, IBM and Netscape.

• Its feeling for the long-term. I see many companies launch a product, find the first or second version is awful, and then drop their support, let the Its feeling for the long-term. I see many companies launch a product, find product languish or drop out altogether. While Microsoft also drops products and changes directions (viz. the latest thrust to the Internet), it has a stick-to-it-iveness that makes life comfortable for a writer. I like the fact that Microsoft gets it really right the third time around. But I like more the fact that it's willing to go it three times, no matter how painful or expensive.

• Its people work hard and get things done. I've had email exchanges with Microsoft employees at midnight on Sunday night. I don't get them from other companies. Other companies ask, "What's the deadline?" Microsoft just produces.

• Microsoft seems to listen. We in the press see an awful lot of good and bad software. Often we would like to use that software ourselves. Often we'd like bug fixes and improvements. I get the impression that Microsoft listens to my suggestions. (Maybe, it's masterful Microsoft PR?) I don't often get this impression from many others.

• Microsoft is intellectual and literate. For a writer, Microsoft is always "a good story." When it releases a new product, it backs the release with extensive Technical Backgrounders, Reviewer's Guides and heavy documentation.

• Microsoft pushes through "standards." What computer telephony lacks most is standards. CT has come from telecom, a world of proprietary OSes and closed boxes. Though NT, TAPI, MAPI, Outlook, Windows, MS-DOS, etc., are all proprietary to Microsoft, they are far more "open" than anything we've ever seen in telecom. Without Microsoft, there would be no computer telephony industry.

Computer Telephony: How does Microsoft see the future of computer telephony?

Bill Gates: Voice and data have always lived separately. That's limiting. Data can add richness to voice. Voice can add richness to data. Bringing them together into computer telephony brings enormous benefits. We'll see those benefits explode soon.

In a few years, we will distinguish voice and data by the input and output devices rather than the network and transport technologies.

Look at the phone system today. It is not fundamentally different than 30 years ago. The sound quality is better. There is call waiting, and voice mail. But the way most people interact with the phone has not changed.

Applications such as Outlook and NetMeeting bring the inklings of major changes in the way people will think about communicating. You can place a voice call in the same way you send email, by clicking on an address book entry, or have Outlook handle your voice mails like it does your emails and your fax mails.

Use NetMeeting to share voice and data on the same line. Two of you can now view and talk about an Excel spreadsheet or PowerPoint presentation.

It will get better. The data network -- including and especially the

Internet -- will become the "Command and Control Center" for the voice network. Wizards on your PC will handle incoming calls (data, voice and / or video) almost as intelligently as secretaries and personal assistants (when we all had them).

Gates: Applications like these depend on a network infrastructure capable of simultaneously carrying real-time traffic like voice and video, and traditional data for file, database, and Internet access. This is why we are building support into our operating systems for network quality of service and more efficient handling of real-time data streams.

When we have this kind of integration, all kinds of problems that have been really hard in the past, like the big call centers that handle catalog orders or customer service requests, become a lot easier.

With the Microsoft Internet Information Server, you can now develop and customize Web-based user interfaces that combine database access with powerful call-control features. This makes call center functionality cost effective for small companies instead of just large corporations as today.

CT: Why should people care about TAPI? What makes TAPI unique?

Gates: TAPI abstracts hardware from software. This is its most important benefit. It will do for telephony what Windows printing did for printers -- make them effortless to install. It will lead to an explosion of innovation (new phones, new communications devices, etc.) -- just as it did in printing.

TAPI enables telephony software apps written by one company to work well on telephony hardware provided by another company. This will make telephony hardware much more useful and explode the market for telephony hardware.

TAPI is still the only telephony API that's built in at no additional charge to any family of client and server operating systems. It enables any type of call model -- from first party desktop to third party client server. This gives TAPI great customer and developer flexibility.

TAPI supports a wide variety of telephony features. Each release of TAPI broadens those features. TAPI epitomizes what open telephony is all about. TAPI enjoys wide support in the industry. Microsoft is very committed to TAPI.

CT: What is the extent of Microsoft's commitment to TAPI?

Gates: TAPI is the cornerstone of our computer telephony platform. We're delivering TAPI 2.1 very soon. We're now in a beta with 2.1. As we look ahead to the next release of Windows NT (i.e. 5.0), we plan to extend TAPI's capabilities to even more broadly encompass traditional telephony and Internet-related telephony. We will support frame relay, ATM and, of course, IP Telephony.

By the way, we have more people working on computer telephony and TAPI in Microsoft than ever before. We'll shortly be announcing our fourth TAPI BakeOff.

CT: What is TAPI's timetable?

Gates: TAPI 2.1 is now in beta. We intend to make TAPI 2.1 available as a Web release in the late spring or early summer time frame. We'll release the product when it passes our internal testing metrics and our customers tell us the product is ready.

CT: Most PBXs do not have a TAPI service provider. What can we do to speed this process? Or is there another solution? The UnPBX, for example?

Gates: Actually, many vendors have already developed or are close to completing TAPI 2.0 service providers for their PBXs but are holding off shipping them until they can test them with the TAPI 2.1 remote service provider.

There are other solutions for client/server support, and some vendors have chosen to go that way in the interim. Now that the web release of TAPI 2.1 is on the horizon, we expect to see many vendors going into beta test with their TAPI 2.0 service providers.

We are total believers in the concept of Windows NT-based UnPBXs, as you call them. We fully expect that over time smaller phone systems (PBXs and key systems) will be replaced by UnPBXs.

These UnPBXs, also called communications servers, will also provide IP/PSTN capabilities and integrated messaging. They'll be a lot easier to manage. Computer telephony using client/server TAPI with traditional PBXs will continue to be important for the foreseeable future. There are a lot of big PBXs installed.

CT: TAPI 2.1 does not work on any clients except for 32-bit Windows machines. Sun's Java promises to bring other platforms on top of TAPI. What is Microsoft's response to this?

Gates: We think Java is a great programming language for certain applications. As you are probably aware, Microsoft's Internet Explorer has one of the best implementations of the Java virtual machine. Java does not have much functionality, however, and it's slow. For mission-critical applications, developers will continue to require the performance and functionality of the Win32 APIs. We are committed to supporting TAPI and evolving it to meet the needs and challenges of and Internet telephony and browser integration.

If Sun or other companies want to expose TAPI 2.1 functionality through a Java mapping layer, that's great.

CT: Currently Microsoft offers an array of APIs to address media processing. Are there plans to merge them with TAPI?

Gates: This is part of what's coming in the Windows NT 5.0 time frame. We intend to more tightly integrate media streaming and call control. We'll be able to talk about more of these technical details in a few months.

The key thing now we're encouraging within the telephony development community is to complete your TAPI 2.1 service providers and apps, because these are important foundational items as we get into the Windows NT 5.0 time frame.

CT: What about TAPI 3.0? We hear it's going to be very browser based. What are its features? Its timetable?

Gates: It's premature to talk in detail about the next version of TAPI beyond 2.1. As we've said, the evolution of TAPI will make it easier and more seamless to use Internet or traditional PSTN networks. That would imply easier integration from within browsers.

We'll have more details about this in a few months in the form of a specification. For the time being, developers should press ahead by using TAPI 2.1 for traditional telephony apps. As well, they should use the NetMeeting 2.0 SDK interfaces and ActiveX controls for Internet telephony applications such as web call centers. I assure developers their work today will not be made obsolete by TAPI 3.0 or any other OS changes for some time to come.

CT: Windows NT is fast becoming The Telephony OS of Choice. What's the future here?

Gates: We've very excited about this. We're seeing Windows NT Server showing up not only in what you call UnPBXs, but also in new form factors that happen to look like PBXs or central office gear.

For example, on the remote access front, US Robotics and RAScom now offer remote dial-up servers built on top of Windows NT Server. When you look at these systems, they look like other remote access dial-up systems. But because they're running on Windows NT Server, other applications, such as Proxy Server, security apps and telephony apps, can also be running there.

Harris is providing a great example closer to home in the telephony space. They have Windows NT Server running on a card that slides into their carrier-class central office telephone switch. In this case, Harris is using Windows to quickly enable other third-party telephony services.

When we speak to telephone companies (the old ones and the new ones) the plea we hear most often is: Give us a platform to deploy new services as fast as possible. We can no longer wait years or even months. We must have these services in weeks.

We're going to continue investing in telephony and other network communications support in Windows in a big way. We want to make it easier to develop and deliver great, compelling solutions that work well in the PSTN and Internet space.

CT: How will Windows NT become the 100% reliable OS which the telephone industry demands?

Gates: Windows NT was designed from the ground up to be a very robust and secure OS. We know these are critical issues for carriers. We also know that scalability and performance are also critical. In many respects we're there now with Windows NT Server 4.0, but the offering gets even better in the Windows NT 5.0 time frame, especially with the Active Directory Service.

The Active Directory will work very well with the next evolution of TAPI to make managing the network (voice and data) easier and less expensive. It will also let you quickly figure out how to connect and talk to another person. Today, a telecom manager has to deal with a multitude of databases every time the company gets a new person or a person moves. Those databases include PBX, voice mail, call accounting, internal and external e-mail, LAN identity and security access, etc. Tomorrow, one entry should do it. Active Directory ties neatly into our thrust to Zero Administration Windows at the desktop.

CT: What technology will Microsoft bring to the Public Switched Phone Network (PSTN)?

Gates: Our operating system products are a foundation -- especially Windows NT Server. We're extending not only telephony support in Windows. We're also enriching networking. This interests public carriers.

Features like Point-to-Point Tunneling so carriers can offer virtual private networking and out-sourced remote access service. "Steelhead" (a code name) for routing and remote access, and built-in ATM support. We have BackOffice server applications, such as Exchange Server, Proxy Server, and SQL Server, that are used by carriers.

There are also specific products we've developed for use public carriers. These let them deliver services to their customers -- Microsoft Commercial Internet System, Merchant Server, and so on. We also have Internet Explorer, which carriers can use and customize for their services.

Over time, the phone network and the computing networks will converge. As a result, public carriers will be able to leverage many of the technologies we offer to provide great, new services to their customers.

Most of the public telephone network providers are also Internet Service Providers and/or have announced plans to provide Internet Telephony service -- as one way of completing an incoming call to busy lines.

We released NetMeeting 2.0 to the web on April 28. At that time over 35 other companies announced interoperable, complementary, and companion products. NetMeeting 2.0 provides standards-based audio and video conferencing.

NetMeeting is bundled with Internet Explorer and will be included with future releases of Microsoft operating systems. We believe it will become the ubiquitous Internet conferencing platform of choice. Many of the public telephone network providers are considering basing their Internet Telephony offerings on NetMeeting.

CT: How will Steelhead affect the world of Virtual Private Networking?

Gates: We already support VPNs today with Windows NT Server and Workstation 4.0. We're extending our support for VPNs. Today, we enable PPTP from a Windows NT Workstation client to a Windows NT Server machine. We have in beta right now PPTP support for Windows 95 clients. This will bring VPN support for RAS clients to millions of users.

With Steelhead we're enabling VPNs to work from one server to another server. That means a company could choose to connect its branch offices and headquarters together using the Internet as a VPN.

We've been piloting this for a few months in a few domestic and international sites. There's a great business case for doing this, especially with international locations. Other companies support Point-to-Point Tunneling Protocol on their dial-up platforms, and we've published the PPTP spec so this can be implemented on other operating systems.

We're working with Cisco and other members of the IETF [Internet Engineering Task Force] to define the next step in tunneling -- Layer Two Tunneling Protocol, which is a natural combination of PPTP and Cisco's initial approach for tunneling, Layer Two Forwarding. That effort is moving along nicely.

Steelhead is more than just server-to-server VPNs. Steelhead is a unified RAS and routing service on Windows NT Server. It supports lots of great IP and IPX routing capabilities. Steelhead is a platform for value-added development. We expose lots of APIs in Steelhead so companies can add routing protocols, change the user interface, and enrich the management of the machine.

With Steelhead we're applying many PC industry business model elements to the internetworking space for the first time. We think many companies will see opportunity in this.

CT: How do you see point-to-point tunneling and other Steelhead attributes changing the way we do inter-site communications?

Gates: Steelhead's Point-to-Point Tunneling will forever change the way companies connect remote offices together. Rather than install expensive leased lines or forego connections to some remote offices, a company will be able to "tunnel" traffic across the Internet -- securely and inexpensively.

The common thread with both of these related offerings is customer choice, increased flexibility, and lower cost. In the future, networks will carry data, voice, and video. Microsoft and its networking partners, such as Cisco Systems, will each provide the pieces necessary to make this vision a reality.

CT: How do you see the Internet integrating with the PSTN? What new services?

Gates: It's hard to predict with accuracy. There are some very cool applications available now. For example, unified messaging can be accessed from your client PC across an RAS connection. Web call centers really increase the level of service a company can extend to its customers.

We expect lots of interest in applications of IP-to-PSTN gateways. Several of these have already been announced, based on Windows NT and compatible with NetMeeting 2.0, from companies such as Lucent and eFusion.

CT: After the Internet, it seems the last frontier is the PSTN. What is Microsoft's vision of how we will place and receive telephone calls?

Gates: The technology underlying the telephone will dramatically improve in areas such as messaging, interactive voice response and so on. I believe more and more people especially in business will gravitate toward their PCs as a rich device for information and call control. For example, pointing and clicking on a name in my personal contact list to make a call, as opposed to manually dialing the phone number, is a feature that Outlook offers users today.

I want to speak the name of the person I want to call and have the PC understand this and instruct the network to place the call over whatever

network makes the most sense. That means the call could go across the PSTN or across the Internet, to their office or cell phone. I want the system to be able to find that person, wherever he or she may be. All this should happen transparently to me.

The key thing here is to provide customers with an easier, more compelling experience and to save time and money. These technologies also give people more control over how they reach people or are reachable by others.

CT: Microsoft's work on T.120 and H.323 (two ITU standards in NetMeeting) are very promising. Can we expect the same standards corroboration with the Enterprise Computer Telephony Forum's (ECTF) S.100 specification, for example?

Gates: Microsoft supports formal standards when it helps customers. T.120 and H.323 are important international standards because they define interoperability between systems and there are no other standards with similar scope and capabilities.

S.100, on the other hand, is more of a hardware and driver model that operates within a single system. There are several solutions in the same space (such as MVIP).

The architecture of Windows NT, along with Microsoft development tools and developer kits, makes it relatively straightforward for S.100 proponents such as Dialogic to provide implementations of S.100 components and drivers on Windows NT without the need for a great deal of participation from Microsoft.

Application developers will be able to access these drivers and components through TAPI and related interfaces.

The real issue is one of opportunity cost. S.100 addresses the needs of a relatively small number of developers of specialized equipment and applications, and those needs can be addressed very well by the suppliers of the hardware platforms in that market. Microsoft will focus its development resources on tools -- such as NetMeeting -- with a much broader market and field of application. We'll make it as easy as possible for S.100 hardware and software developers to use Windows NT as a platform, even if we don't ever ship an S.100-based application or driver ourselves or participate in ECTF activities.

CT: Can Outlook be used to manage your voice mail as well as your e-mail? What is the benefit of using Outlook as a platform for Unified Messaging.

Gates: Outlook is the ideal way to manage your messages and communication, whether e-mail or voice mail. A number of companies are already offering Unified Messaging products that use Outlook and Microsoft Exchange Server.

Bringing voice mail into the Outlook inbox allows you to use Outlook's great viewing and information management tools (like filtering, sorting and grouping, customized views, and find items), uniformly across all your messages.

CT: Outlook was designed as a messaging platform. Many of our readers would like to use it as a platform for real-time control of telephone calls at the networked desktop. Does this make sense?

Gates: It makes sense. Outlook was designed as both a great messaging (i.e. e-mail) product and a way to manage your personal information, including contacts, tasks, calendar information, etc. These capabilities are integrated into a single desktop information management product.

For telephony support, using Outlook today, you can dial telephone calls directly from your Outlook contacts using TAPI. Upon dialing, Outlook will log the call in the Journal, allowing you to take notes during the call and know exactly where to go to find those notes in the future.

Outlook can also be extended by other software companies who want to integrate telephony products with Outlook or by corporate developers who want to link Outlook to existing in-house telephony systems. Outlook includes an "object model" that lets programmers use Visual Basic to access the contacts, journal, inbox and calendar information in the Outlook database.

Programmers can also use MAPI to extend the Outlook user interface and build sophisticated real-time applications. With these tools, quite a few companies are beginning to market "call control" applications that use Outlook to store information, taking advantage of the integrated user interface and unified inbox that Outlook provides.

CT: How do you feel about speech recognition?

Gates: This is a very compelling technology that always seems to be just around the corner in terms of wide scale use. We're investing quite a bit into speech recognition and text-to-speech, because we see this as another natural way of interacting with the PC.

Speech recognition has lots of value in telephony. This is one of the ways that computer telephony can offer a different, compelling user experience that can incite people to deploy the technology on a wide scale. This gets back to our discussion about broadening the market for CT.

It's exciting to imagine dialing into my messaging application using my cell phone and have my urgent email messages read to me with very human text-to-speech. There are already several compelling products today that use these technologies to deliver carrier-based services. It is beginning to gain presence in some customer premise-based products, but we're really at the beginning of this market from what I can tell.

CT: You believe these tools can be dramatically improved?

Gates: In the past, companies were virtually all alike in terms of how well they managed these things. The tools were well known, and so there wasn't much differentiation. But my thesis is that with the incredible advances in technology -- both computing and

Technological change has been pretty unbelievable. When Microsoft got started, computing was a million times more expensive than it is today. Computing was simply a central tool, a tool for keeping track of business accounts or reservations. Only the largest organizations could have this tool. In fact, it was a clear advantage for very large companies.

That factor of a million did more than simply drop the price. It changed the very character of how we think about computing. It changed it to be a tool of communication. It enabled people to individually think about how they manage information so that everybody has a machine and great flexibility in dealing with that information.

CT: You talk about the "digital nervous system."

Gates: When I say the nervous system of a business, what I mean is the way a business deals with events -- planned events like yearly budgeting or sales results or unplanned events -- competitive activities, unhappy customers.

And there are all the different systems -- the meetings, the paperwork, the way the information workers are organized, the way that information about customers is stored, the budgeting system, the coordination system. All of these are the nervous system of the company.

No matter what business you're in, it's my claim that the excellence of that nervous system determines your competitiveness.

CT: Where will it go in the next 20 years?

Gates: Over the next 20 years, the cost of computing will fall again by a factor of a million. Communications won't fall that fast. But it will fall fast. The good news here is that the demand elasticity for both computing and communications is quite amazing.

Other markets where you find efficiency anywhere near this, you have a tendency to shrink the market. When radial tires came along, they lasted longer, and so they were quite a bit cheaper. But people didn't respond by driving around a lot more. So the total revenue of the market shrank.

When people looked at computing and the exponential impact of Moore's law, the doubling in performance every two years, they were really worried about what would happen to the computer industry. Would it also shrink?

In fact, the computer and telecommunications industries have grown because the ways we can use this power have proliferated even faster than the performance improvements. My belief is again that this next factor of a million will be absorbed in the same way. So computing and communications are very different to any other industries on the planet.

As we use PCs for communications, we need to think about more than just the speed of computation or even the size of the memory or the size of the disk. We need to think about how these machines are connected together. If there's anything that will slow down the pace of revolution it's if we don't get these high-speed, low-cost connections in place.

I'm personally very optimistic about the demand elasticity for high-speed connections. I believe that phone company investments in ISDN and ADSL will pay off in a major way, even more than they expect them to. I think that PC cable modems coming in and providing users additional choice will spur even more competitiveness in this space.

Once you have tasted fast access to the incredible resources on the Internet, you'll never go back to slow speed dial up again.

CT: Where is it all taking us?

Gates: When we take the improvement in speed and the improvements in communications and put them together with what the software industry is doing, I'd really say there's an opportunity to do almost anything businesses are interested in doing with information.

If you can imagine how information would flow in your company, how it would automatically be analyzed and compared and summarized and brought to the attention of the people who need to know about it, I'd say building a system that fulfills your wildest dreams is possible today.

It's possible with development costs that are really quite modest. We no longer talk in terms of many-year projects -- two, three, four years of development or a huge backlog of activity. Today, if the right infrastructure is in place, people can build computer applications in four or six months. And the beauty of these applications is they can get piloted in a few days. So the user is very involved in saying how they want it to work and how they want it to tie in.

Now, this does assume a basic infrastructure for the knowledge worker. It assumes every knowledge worker in the company has a personal computer that's reasonably up-to-date -- say, about a three-year hardware cycle. It assumes they have common multimedia productivity tools for spreadsheets, documents and presentations and that those are shared not only within the company, but also with the key partners, with the law firm, the accounting firm and the key vendors you work with.

It assumes there's electronic mail where your email, faxes, and voice mail all come into one place. And it's an extremely reliable email system, so people can count on it working not only for simple text messages but also for sending rich compound documents and voice mail.

It assumes these machines are connected to the Internet, so that not only the knowledge and information inside the company are available, but the company's total communications also connects them out to customers, suppliers and the Internet in a very rich panoply of many media -- voice, video, imaging, data, etc.