MicroTimes Interviews Jim Clark from Mosaic Communications (1994)
They discuss how to commercialize the internet.
from the October 17, 1994 issue of MicroTimes
I previously published an interview with Jim Clark when he worked at Silicon Graphics.
Mosaic: The Future of Publishing?
Jim Clark on commercializing the Info Highway
By Mary Eisenhart
“If you don’t change every year and a half, you're dead,” says Jim Clark, in his personal corollary to Moore’s Law.
Even by Silicon Valley standards, Clark's recent past has seen some rapid change. From the time he founded Silicon Graphics in 1982, Clark was regarded as the company’s chief visionary and charter of its course; he spent much of 1993 involved with such projects as SGI’s venture into the emerging market for set-top boxes, designed to bring interactive services to cable TV customers.
In February of this year, he resigned as chairman of SGI, and in April made headlines again with a startup—Mosaic Communications Corporation. Its mission: enabling a full range of commercial transactions and information exchange on the Internet.
Clark’s cofounder in the venture is Marc Andreesen, who only last year graduated from the University of Illinois. While still an undergraduate, Andreesen worked on the staff of the university’s National Center for Supercomputing Applications, where he was simultaneously excited about the wealth of information available on the Internet and appalled by the difficulties that beset users trying to access it. His response to the situation was to gather a team of his fellow students to write what's now called NCSA Mosaic—an easy-to-use, point-and-click navigational tool for the Internet.
Freely distributed since last year, NCSA Mosaic now has an estimated two million users and has resulted in an explosion of publishing on the World Wide Web.
Mosaic Communications Corporation, which includes most of the original software's authors, is building products that enhance the functionality of the original Mosaic in strategic ways, notably speed (which, on PC-class machines and ordinary phone lines, is prohibitively slow in the freeware version) and security, utilizing cryptography from RSA Data Security. The first products, due to ship in October, are Mosaic NetScape, client software for Mac, Windows, or X, and Mosaic NetSite, UNIX-based server software.
According to Clark, the ability to transfer money in a secure manner will, for the first time, enable a full range of commerce on the Internet. In particular, it will solve a problem that's beset the publishing industry since the dawn of the digital age—a commercially viable way to sell content to subscribers online. He elaborated on all this when we visited Mosaic’s offices in Mountain View. (Mosaic Communications Corporation, 415/254-2601)
Let’s do some catching up... Quite a bit has changed since I saw you last —you were on a panel with some cable company executives at the cable convention last year, and there appeared to be some serious culture clash.
It’s still going on, actually, I think. Somewhere along the way I concluded that cable guys were a noninteractive culture nondigital culture. And they were talking about interactive digital television. [laughs]
They'll get there, it’s just a very hard thing. They have no digital culture-all the suppliers in that industry are all still dealing with technology that's forty years old, basically. Scrambling analog signals. Doing funny things with analog signals. The right way to do all that stuff is to do it digitally and then encrypt. Digital is so much cleaner.
It’s really no fault of theirs that they're coming along so slowly, it’s just that they're in the wrong kind of industry. They have this huge one-directional, broadcast-mentality infrastructure that has to be converted to a two-directional, switched digital system carrying data, television signals—TV pictures and sound.
You look at that, as I finally did, and say, “I don’t see how they're ever going to get it.”
But it was impossible for me as an employee of Silicon Graphics, most especially as the chairman and the person who was the spiritual leader of the place, to be standing up and saying, “I don’t believe that this is going to win. The PC is going to win.” I couldn't even acknowledge that myself. I had to get out of there in order to acknowledge it.
When I was there, I believed what I was preaching. But once I got out and disconnected, it became obvious to me that the PC was winning. My head cleared fairly quickly. [laughs]
I resigned from the board with absolutely no plan specifically, other than I wanted to do something that didn't require that I have to say, “I believe in everything that SGI does no matter whether I was part of the decision or not." And as you know, if you're a board member you have to do that.
Board members are basically endorsers of what the management wants to do, and I didn’t want to have to do that, so I resigned altogether, to give myself freedom to think clearly, and freedom to go pursue things that weren't necessarily in SGI's best interests. Not that I'm trying to look for things that aren't in SGI's best interest, but I want the independence and freedom.
What causes you to believe that the PC is going to prevail over TV?
It's tied up in two things—interactivity and digital.
From a practical standpoint, what does that mean?
That means that the technology required to do television on a PC is going to be an easier thing to accomplish than the culture shift and the infrastructure upgrade required to switch the television into an interactive digital device.
Going for the PC is the fact that it’s already interactive and digital. What's missing? That you incorporate video signals. Silicon Graphics was demonstrating that that was possible digitally; now you can see all that stuff on PCs, the more advanced ones just coming out. Integrated video is going to be there on all of them, it’s a no-brainer.
So, given that you have integrated video on a PC, what do you have? You have a smart television.
The computer is going to evolve into that stuff much more readily than television will evolve into an interactive digital device, and it’s because of all this inertia.
The truth is that it's going to be some merger of the two still, but the necessary new investment is something that computer people know about. The existing television people aren't real up on the kinds of investment, both software and hardware, that have to be made in order to make it work, so there’s a built-in advantage for the computer orientation.
The PC comes at it with a computer orientation. Traditional television people don't.
Computer people are also used to moving faster, living in a world governed by Moore’s Law.
That's right. If you don’t change in a year and a half you're dead! [laughs] That’s my version of it. Clark’s Corollary to Moore’s Law.
So when did you begin to think that Mosaic was more interesting than settop boxes?
I was fairly unclear on what I wanted to do until I began to talk with Marc Andreesen. Marc began to enlighten me about what the Internet was about and what Mosaic was about.
How did you come to connect with him?
I just mentioned, “I’m looking for smart guys to partner up with,” and a friend of mine said, “One of the guys who wrote Mosaic is out here on the West Coast right now. You might want to contact him.”
I said, “Well, how do I reach him?” He went over and typed in “Mosaic” and bingo! there was Mosaic running on the screen. And he said, “Just find your way,” and he walked away.
So I began to poke around and found Marc’s email address and sent him email, very brief, asking him to meet with me. And so he did.
Little did I know what I was getting into. I went down and had breakfast with him. He let me do most of the talking; he just kind of modestly said that yeah, Mosaic had been his idea, and he pulled the team together to do it, He didn't seem to be too charged about it.
So we went off doing other things. We went off looking for set-top boxes, trying to do a deal with Nintendo, and all these different things. The more we é meandered around out there, the more we realized that hey, this is all not going anywhere in any short period of time.
And meanwhile, Mosaic was continuing to explode.
We started talking about it, and once we really hit on the idea of what we wanted to do, we said, “God, we've got to go out and recruit those guys.” So we did, real quickly.
We now have ten NCSA employees. This is NCSA West.
So what are some key things people will be able to do with your product that they can’t do with the public domain version of Mosaic? Take me, for example, with my Mac at home, my Plain Old Telephone Service, and my 14.4 modem.
One thing is that you can get a reliable Mac client, bundled with a dialer and all the registration stuff to get you on through an Internet access provider. Once you're on the Internet with this, you can easily browse the net and get quick response.
The very first alpha version of our stuff is ten times faster than what was out of NCSA; we totally rewrote it and restructured the code to deal with the network more efficiently, deal with images more efficiently. There's another factor of somewhere between five and ten that we'll get if people make certain changes to their source documents.
That's a big change. If you're on a 14Kbit modem, it’s prohibitive right now.
That's the main thing in the first release. Then there’s security. Users may not think about it, but merchants who want to get on the network, and have liabilities associated with taking credit card numbers and the like, do. This first release of the software establishes encrypted protocol between the client and the server, using RSA Data Security public key cryptography.
Currently, if I’m a supplier, I will connect to the Internet given I have information I want to give away. But what if I want to charge for it? That’s what this software will do; give you the mechanism to privately conduct commerce between users and suppliers,
Right now there's about two hundred small Internet service providers, who get little tiny amounts of money, because not many suppliers will spend much money to put their data on the net, because they have no way of getting paid for it.
Getting paid for it aside, though, the net in its current state is a huge vanity press. One of the things people have found most seductive about publishing on the net, say, creating their own Web pages, is that they don’t have to go through an editor or a publishing house, they just put it up and it’s there.
The explosion of information like that presents immense problems for those who want someone to prefilter something. One of the reasons I, as a reader, would go to a publisher is that I assume they've done the first-course filter.
The net is going to enable people to put their stuff online and anybody to get it. If they've got a way to chargefor it, there is going to be the occasional rare bird that gets a hit out there.
But I think, with regard to books and the like, part of a publisher's function is to be a filter. People are going to have an easier alternative to the fact that sometimes they seem like an arbitrary filter.
Ever since he’s been talking about Xanadu, for instance, Ted Nelson has been stressing the importance of editors. You subscribe to Trusted Editors X, Y, and Z to find stuff that interests you and present it. Because there’s so much Stuff to slog through.
Yeah, so much junk. I think that’s probably true. It'd be chaotic.
There's another issue too, another example of why people may go toward the publishers. It may become prohibitively expensive to do it yourself.
The first thing that’s going to change very rapidly is that those who want to give information to the net are probably going to have to pay. If I’m a supplier to the network, I’m going to at least have to pay according to the size pipe that connects me to the net. If I want to connect-with a little garden hose, then I can at most deliver a garden hose’s worth of information, which governs the rate at which I can make money.
Let's say I’m supplying and charging people to read—I mean, no one’s going to give it away, right? So I set up a subscription service, which is what our software is going to enable you to do. Anybody can set up a subscription service.
That’s where the crypto comes in, among other things?
Right. You'll be able to govern who gets access to your service.
Once you've done that, you have to decide: How big a pipe are you going to feed to the net? That governs the rate at which you make money, in some very direct way. Those who own big pipes will be able to make money at a faster rate. By definition, large publishers will own big pipes to the net, because they’ll be amortizing the cost of that pipe over many of their people.
So the real challenge for an isolated small publisher is: How big a connection to the network? He won't want to give that information away and let it be duplicated hundreds of thousands of times, because he loses all his money. If he intends to make any money at all, and I think all writers intend to make money as a rule—
Either that or they're just ego-driven and they want an audience.
Well, if they’re ego-driven, then fine, they get their stuff out there and they’ve got a name, but everyone's copied it, and...
I think that in the end everyone has to have some economic motivation, because economic motivation is how you pay for your food when you go to the grocery store.
Publishers can buy a big pipe to the net and then they can be selective about who they let occupy some of the bandwidth of that pipe. And when they have a hit, they have the whole pipe there.
If you are a random small publisher, and you have a hit—by definition, a hit is when a lot of people are buying it—if you've only got a small pipe, how can a lot of people buy it? You've got delivery problems.
I think there's going to be a kind of economic boundary around this thing. None of this is magic. Internet as it is today is kind of magic, because it’s been subsidized by the government, and it’s very embryonic in its current state, but when you leverage it up into a viable economic operating entity called The Internet, and you have Internet operators, who make money, then the economic motivation is for big publishers to get a connection to the net that is big, because they can afford it. They're ‘supplying their data to many, many different kinds of readers.
That's one part of it, and another is that the name provides a place to go. Random House, or maybe a virtual bookstore. I guess a person can go to a bookstore independent of the publisher and say, “Will you carry my product?” But even a bookstore is going to say, “Well, who are you? I already deal with Random House and McGraw-Hill and Simon & Shuster and ten thousand other guys.” There is a limit to what they can do. They can’t take random stuff over the Internet.
So you think that the wild crazy openness we have right now is more or less a blip in evolution, and that the structure of publishing on the Internet is going to look very much like the structure of print publishing now?
Yeah. There will be entities that will be publishers, and they will be aggregators of content. And filters.
Right now, garage guys can do things they’ve never been able to do in previous history. Do you think that’s going to go away?
Well, when Leland Stanford strolled into this valley, he said, “I’m going to claim that!” And he did. And he kept it. Wouldn't it be nice if you could walk out there today and say, “I’m going to claim that!”?
Frontier times end, is my point. They do. Space gets staked out, mechanisms get in place, legal contracts get bound—then you have a thing that’s a little more rigid. Now it’s wide-open frontier time, but the rules are going to be established. They’re going to be established with encryption.
It can't be a wide-open free-for-all, because no one makes any money. Even now, the reason the Internet is starting to be economically viable is that companies find it useful to give away brochures, to give away information on the net. So lots of companies are buying a presence on the net today, and paying whatever it takes, and giving away their product literature.
The next step is the ability to control who gets access to what information. You can only do that with encryption. In the first phase the server, the merchant, has a signature, and that allows you to create encrypted sessions; the second phase is authentication of the client, the user, and then you’ve got point-to-point secure communication, authentication each way.
When that’s totally bulletproof, rock-solid secure, you’ve got commerce in the broadest sense—people able to do wire transfers, initiate stock trades, and the like. That’ll come over time on networks like this.
Do you remember Personics?
No.
Personics was a company that, dunes the ’80s, decided to offer a legitimate alternative to the dread Home Taping Menace—instead of making your own party tapes off your records, you could go to the record store, select songs on a jukebox-like device, and a custom tape would be generated just for you. You'd pay by the song, at different preset prices. The record companies got paid, the artists got paid.
Eventually Personics died, for what seemed to be two reasons. The machines were too hard to use; also, for the most part they could only get the little independent labels to put their stuff on the system. The independents loved it—apparently their sales quadrupled in the record stores with Personics machines—but the major labels simply regarded it as a threat to their traditional sales model.
Has anything changed?
No. People are going to be hyper-paranoid—” How can I protect my intellectual property?” is think it’s going to take a while for them to get really comfortable.
But we're talking about the traditional people who think of their stuff as intellectual property, like the music industry and the software industry. They're going to be careful.
Look at the vast numbers of publishers who have material that’s only time-sensitive value. They’re not going to be hypersensitive. Just so you can give them a small amount of protection, that it can’t be copied for about a week, for the most part they’re happy. Where the value of information drops radically with time, those people are going to be out there big-time, because they deliver more quickly, it’s time-sensitive information; it’s going to be very, very valuable. We're talking about the market of people like yourselves, publishers who want to get an issue out there. You want your readers to get it and be able to get it quickly. Two months from now, it’s not nearly as valuable. Time, Newsweek, those kinds of publications. That's going to drive the whole thing initially.
Now National Geographic tends to publish timeless things. Those who have information, intellectual property, which has timeless value, are the ones who are going to be real concerned about network dissemination of their intellectual property, because of the speed with which it can happen. They're going to be slower to adopt, and there are going to have to be better protections and safeguards in place. What we have to do is create better and better mechanisms to prevent copying.
But compact discs today are copyable. You have a perfect rendition. I can go out today and I could master-copy, in a day’s work, intellectual property that gets sent out on a compact disc.
The big difference is how much work is required. [laughs] When you make it so that copying takes a lot of work, people are going to be able to publish this way.
How is Mosaic likely to affect advertising? When he talked to us a couple of months ago, George Gilder said that advertising is going to become more demand-driven, and also that in some cases Gees is part of a publication’s value-added.
A lot of what you think of as advertising is TV-based. There is a brainwashing that’s gone on in the last two or three years, that we all hear, because we all watch television. Even a lot of focus of the news, in terms of mergers and the various things that go ae is things like “CBS is pene bought.” “NBC is being bid for.” “Time Warner is..
But those are elatnieh malt picihesses ina Fantiy sort of global way. The total motion picture industry is only a five-billion-dollar industry. Cable TV is twenty billion.
You start examining the world’s publishing industries, and you've got a much bigger collective industry. Because print media is a oe business. It just isn’t in our face all the time.
Well, exactly what Mosaic is, in its initial phase, is a way for the print industry to turn a page, ‘kind of turn over a new leaf. It’s not the TV model at all—it's more the magazine and newspaper model.
[points to USA Today page] These guys make a lot of their money on advertising.
We make a lot of our money on advertising.
Exactly. You're in the print media business. When people read your publication, there are certain kinds of advertisers that go to you. You don't, as a rule, have Revion taking out ad space in your publication.
We do have a sushi bar...
But not lipstick. My point is that publication types are really vehicles for delivering advertising to a target audience. Yachting magazine—pick it up sometime. Well over half the magazine is advertisements for sailboats. Two or three articles and some ads, and the entire second half of the thing is “This boat's for sale.”
The same is true of golf. When you read a golf publication, guess what kind of ads are in it? Golf-related ads.
That kind of advertising is just a fact of life. In fact, you want it. One of the reasons you even read these special-interest magazines is, “Well, I think I'll check out what the latest golf clubs are.”
A publisher like IDG—all their PC and related publications have tons of ads about memory upgrades and modems and every conceivable kind of add-on device. It’s big business.
Do you think audiences will get any better at filtering and distinguishing between an advertising statement and an editorial statement, or do you see the lines getting so blurry you won't be able to tell?
I think there are overt attempts to deceive in that regard with these special advertising supplements. I’ve been confused numerous times—I’ve been reading along and say, “Gee, this is weird, it all seems to have the same names,” and then I'll notice in the fine print that this is an advertising supplement.
I typically just brush past those, because I don’t need that kind of deception. I don't mind ads being in publications. It doesn’t bother me; I sometimes want them to be there. But I don’t want somebody trying to deceive me that this is an objective statement.
That's a personal statement, but I think most people are going to be that way. What we expect out of journalism is objectivity. Objectivity means that if you think it’s a great product, great, say that it sure measured up in your mind. It’s not like you're getting paid to say that; you're being objective. And if you think it’s a terrible product, say so.
Where do existing online services fit into this? I gather that the commercial Services are interested in your product.
Various of them are. It’s kind of a dilemma for them. In order to offer an online service in the last ten years, you basically had to lease your network. Or build your own network—CompuServe built their own network, they built points of presence, modem banks, leased the lines from the carriers, but they were a network operator. Since they got that network up and running they decided they'd lease part of it to VISA for carrying VISA’s credit.
The notion of the network operator over the last ten years has not always been distinct from services operation. CompuServe has a services aspect to their business; it’s about two- thirds, and the other third is network operations.
AOL leased all of their lines and points of presence from SprintNet; Sprint operates that network.
But now people are going to be able to offer online services over the Internet, because the Internet itself has broken away, and it will be a viable standalone business on its own. It’s become a big enough business because so many people connected to it that no single company in the world could build a network as big as the Internet. It would be too prohibitive and expensive.
So if you've got encryption, security, the Internet becomes the means, the channel, the conduit, the network, the mechanism that connects users to suppliers of information. Since you've got users who are willing to spend money with suppliers, with security of transactions built into the software that works this network, the user can pay for services. And since the user will be able to pay for services, and that’s something unique and new to the Internet, the service supplier or information provider has got the money to pay the network operator.
It becomes a nice healthy cycle. The network operator then has the money to pay for the bandwidth or build the wires, the conduit.
Without some mechanism for the subscriber to pay him, the publisher doesn’t have a whole lot of motivation to get connected to Internet. He can't tell who's reading—or he might be able to tell, but he can't charge.
Today’s cash-flow model for the CompuServe-like service or the AOL-like service is that the content supplier or the publisher gets money from AOL. AOL owns the subscriber. The content guy is ead ee content to the operator.
Today's model for the Internet doesn’t have any way for the user to pay with money. But when you enable that, the supplier is willing to spend a lot more money; the user pays the Internet operator a little bit to be connected. The Internet operator could be a telephone company, because they own the fiber and the wire. They could be a network operator, and I think in large measure they will.
Do you think the new model will replace the old model, or will they coexist?
They can coexist. AOL can mutate to accommodate this model. They become a content aggregator, a place to go shopping. “Why do you want to go to AOL?” “Well, because I like it. It's got all this neat stuff. I find stuff at AOL that I don’t find if I just go out prowling around on the Internet.” They may very well coexist.
But it is a fundamentally different model. This [the new model] has a lot more viability because this is exactly your current publishing model. The same exact business model: users subscribe to things from these guys. Because they own the subscriber relationship, they can then tell their advertisers, “This many people read my publication.” So the advertiser will pay them more money to be in their publication. That's a very important component.
Your first two products are a client and a server. What about authoring tools to make it easier to actually produce content, the equivalent of PageMaker in the mid-’80s? Right now people can theoretically put content up, but it’s a pretty painful process.
You're going to need some adaptation of tools similar to what you have today for creating content. A Quark-like system that allows hypermedia linking. But Quark is a specific layout system, and it is very page-oriented. Current HTML is not page-oriented, and you don't describe layout—you describe a document and it gets laid out according to a set of rules. It’s a little different model.
There’s going to be a plethora of authoring systems, and you're going to have a hard time figuring out which ones to use. I think the ones that’ll win are those that accommodate a match between what HTML is today and what it becomes. They bridge the gap; they start creating enhancements to HTML.
There’s a huge frothy activity that’s going to occur over the next couple years, and what will come out will be somewhere in the neighborhood of five to ten new companies servicing this online electronic medium. It’s definitely as big as the early ’80s. Bigger, in fact—you're talking about a more rapid adoption because of the electronic nature of it.
Bear in mind that you're talking about mechanisms to charge for intellectual property that’s delivered over worldwide networks, the ability to charge for anything as small as a single product, to a subscription to a newsletter, to running an Internet shopping center. The software to do all that, the evolution of authoring systems, is going to happen very, very quickly because of the explosive nature of all this. When these systems are available and how quickly they get to market is going to stimulate how quickly the market grows, and we're in the midst of all that.
Right now you have to build distribution channels. Just think—once the whole system is “booted up” and everyone can do electronic commerce because they have public and private keys, what happens to the current distribution channels?
In other words, Fry’s ought to be online, big-time. If I were Fry’s I would be looking real quickly at how to set up a virtual store.
Do you see any convergence of this with wireless and PDAS and that whole market?
Kinda not. I think that’s a different world. It may merge together.
Do you think that will change when the General Magic stuff comes out, which is very oriented toward online commerce? The applications that people like Bill Atkinson talk about are very similar; perhaps they’re coming at it from a different point of view?
Who's Bill Atkinson?
He’s the author of MacPaint and HyperCard and the chief scientist at General Magic.
Well, let's just say that I’m pretty negative on PDAs. I don't get excited about PDAs. It’s questionable who needs it.
It’s like pen-based?
That's the best analogy. I really think that PDAs are like pen computing. There may be a time, and it may be ten years from now, when you just can’t live without ’em, but right now feel like I can live without them.
Quite honestly I’d rather carry my portable. I can't get by without my portable anymore—I carry it on vacation. Check in on email and so forth. But I can't see doing that on a PDA. I've gotten used to the keyboard and that’s an efficient way for me to interact. Writing is less efficient—I can actually type more rapidly than I can write. And I can edit.
Wireless is a different question. That is a larger issue. I can see portables becoming wireless. Let’s just go out a couple of years—I'd love to be able to bring my portable on the plane, and be able to get a little infrared feed, log on to the Internet from the plane. This will happen. That'll be wireless.
The Internet goes everywhere, because it’s transport-independent. I’ll be able to get online by satellite and by cellular, and do the same kinds of things I do now when I dial up with a 14Kbit modem.
You'll be able to pay for your connection when you've got it open, and it closes quickly; you'll be able to open and close it for asynchronous kind of traffic. Now if you open up a connection and they’re spewing video at you, then you've got a full-time, relatively high-bandwidth connection. I think that’ll stay wire.
Nicholas Negroponte of the MIT Media Lab has an inversion he likes to talk about—we used to be wireless for television and wired for phone and communication; now we're going to be wired for television but wireless for all other communication.
You'll need wire for video. You won't be getting it over the air because it’ll be on-demand.
In order to have interactive video today, you're asking too much. But you can have interactive static media, interactive print media and pictures. That's possible today over 14-Kbit modems. So you can start building the paradigms for interaction today.
Time marches on, the network gets higher and higher speed. The underlying transport gets higher speed. You get ATM. And suddenly TCP/IP, which is just this little vocabulary for communication, is riding on top of this and getting faster and faster. Then you start being able to stream video, stream audio, and you can use the same paradigm to initiate and video.
Mosaic is the navigator for interactive video. It’s here today; the same exact paradigm will be applicable to initiating interactive video. As soon as the network can support it, you've already got the software installed base; you don’t have to go out and create all these fancy browsers.
This is where the cable strategy of trying to build the video navigator on the basis of a system that doesn’t have any installed base is dead-end. The Internet has a huge installed base, and it’s going to be even more.
What computer ads would you like to see in the future? Please comment below. If you enjoyed it, please share it with your friends and relatives. Thank you.