Steve Jobs on OpenStep and WebObjects

Good morning. I'm Steve Jobs. I'd like to welcome you to Open Step Days at Object World. We've really enjoyed putting this together with OMG and we hope you enjoy the tracks today. We're also having a party, the OMG Open Step Party tomorrow night at the Center for Performing Arts across the street. We hope you can come. And we've got a really fun day today for you. How many of you saw the billboard coming into town? Any of you? Isn't that great? And after this talk at 9.45, there is a second keynote on Open Step and at 11 o'clock, there is a third keynote on web objects.

So I hope you can stay around for those, they'll be quite good. So we got a lot of exciting stuff to cover today. And what we'd like to do is talk a little bit about the past and the present and spend most of our time on the future. We've got a ton of new products coming out that we want to take you through today. So the past. The past for us is 1994. 1994 was our first year as a software company and we were extremely pleased with how well we did. We did $50 million in revenues and we managed to become profitable. We did a lot of... I know some of the people clapping are our shareholders in the audience.

We were able to ship a lot of new products and initiate a lot of new programs last year, but the one I think is the most strategic of all is that we really established a major effort in professional services because we realized that technology alone was not the most effective route to inject this into our customers, that if we could supply some technology transfer that would go a long way. And we became the largest object supplier in the world, a position that we really appreciate. Now let's take a look at the present. The present for us is the first half of 1995. The first half of this year we have completely refreshed our entire product line, actually starting in December of last year with Next Step 3.3 and carried on through today.

We've completely refreshed every product we have. We're not public so we don't release numbers except on an annual basis, but we have exceeded our revenue and profit plans for the first half and we're very, very happy with how the business is going. It's growing very nicely. We've added some really wonderful mainstream new customers to the list. We grew our consulting business, this technology transfer consulting business, 400% from January through the first half of this year and we have remained the largest object supplier. Let's take a look at a few of these in detail. On the product side, we shipped Next Step 3.3 for Intel in December. We shipped it for all the RISC platforms in first quarter. We shipped Next Step 3.3J in second quarter on all four platforms carrying all of that into Kanji.

We shipped Next Step Developer 3.3 on all four platforms and we updated our enterprise object framework which is the most profound piece of new technology we've shipped in several years. All of our major customers are using it as well as our portable distributed objects technology primarily adding support for digital Unix so that you can run Next Step objects on all servers from digital as well as all servers from HP and Sun. So we've been up to a lot in the product space. Customers, we've got a very long list of customers. We really love them and they sometimes give us a lot of hassle about how our products can be better which is what we love so we can go make them better for them. And we're very, very close to these folks. Now I couldn't list them all but I pulled a handful and just threw some in alphabetical order.

British Telcom is a new customer. On this list, Fidelity is a new customer, Merrill Lynch is a relatively new customer. On this list, NTT is a relatively new customer, Walt Disney is a relatively new customer. So this is just a sampling but a very, very blue chip customer list and we're very, very privileged to be working with these folks. And then object experts, this is our, we have three parts to our professional services. We have training and education, we have support and we have object experts. And object experts are our consulting organization that go on site with customers to help them train their people, architect their core objects and their applications and of course write some of those core objects if needed.

And we've grown this group from 4 to 21 people so far this year. I expect that it will go out somewhere in the 40s by the end of the year and I suspect it will hit 100 people sometime next year. We are absolutely moving into the consultive selling mode, following in the footsteps of probably Oracle would be the best example. These folks do technology transfer, they're a direct pipeline to next engineering so that when we have customers doing very state of the art stuff, they can get very state of the art information. And they really become a key component of success for a lot of projects that are customers. Lastly, Software Magazine ranked the top 100 software companies early this year based on last year's results. And in our first year as a software company, we were a third of the way up the list. We're the 67th largest software company in the world.

And based on last year's results in terms of growth rate, we were the fastest growing of this list. So we're very pleased with that too. Now we want to spend some time on the future. Where are we going? We're going to show you three of our major directions today. And we've got a bunch of demos. If everything works, it should be pretty enlightening for all of us. The first one is OpenStep. We announced the OpenStep initiative a few years ago. And OpenStep was a realization that what our real jewel was was our object layer, not our complete stack of hardware and operating systems and object layers.

So if we could just break that object layer out and run it on popular operating systems, a lot more people would want to buy it. And we've been at that work for the last few years. And it will be shipping in the first half of next year. And OpenStep has a lot of partners. Hewlett Packard, Sun Digital, each of them doing OpenStep work. Sun has announced their schedules. They will also be shipping around the end of this year to early next year. Others will be announcing their schedules shortly. But I think the thing that has captured most people's imagination is putting OpenStep on Windows. And I think a lot of folks have thought we were just, it was pure vaporware. And what we're going to do is show you that running today. OpenStep for Windows runs on Windows NT and Windows 95.

It has Windows look and feel. We made that decision very early on that although we think we know a little bit about user interfaces, when we're going into the Windows world, we have to be an excellent Windows citizen, both on look and feel and on interoperability with all the Windows standards, like print panels, save panels, full OLA integration. Matter of fact, I think we'll be one of the best OLA citizens in the world. And so this product is running. We're going to show it to you today running on Windows NT. It is going into alpha next quarter. It will be available with the Unix versions of OpenStep in an alpha release next quarter, in a beta release in Q1, which should be a very good beta release, and it will ship in Q2 of next year. And what I'd like to do now is just give you a very simple demonstration.

This is code out of engineering with a lot of debug code in it, so don't take it too seriously, literally, but let's see what we've got here. I've got myself a Hewlett Packard Vectra running NT here, and I'm just going to do two things. One, I'm going to run a simple draw document. As you can see, this is total Windows look and feel. Here's the program manager. And here's my draw document coming up. And I've got a simple NeXT logo on here, a Hewlett Packard logo, a Sun logo. Matter of fact, let me go down here to the file manager, and I can just grab something like a NeXT logo EPS and drop it on.

Ooh, color map problems. Well, there we go. I've got a NeXT logo, and I can resize it here and move it around and do all those nice things. So you get the idea. This is obviously the demo that we didn't put the most work into, but here's draw running on Windows NT with all your nice little Windows doodads hanging around. We have Windows menus up here, by the way, on our draw document, so we can bring up an inspector potentially. There's a line inspector. We can bring up a color palette. panel here. So here we are selecting colors. We can bring up a font panel, etc, etc. So you get the idea. This is, we are really running Windows here and we are really running Draw.

And that's the things that most people never thought they'd see together. Okay. It's funny, the people that have never used Next Step in the audience are going, well, what's so amazing about that? And the people that have used Next Step are going, my God, I thought I'd never live to see the day. In addition to, obviously, the user environment, we have all the developer tools running. And what I'd like to do now is show you Interface Builder running. Here's a simple window. And let's, I'm just going to build something very simple here. I think I'll just take a text field and drag it in. And I think I will go over here to bring up a font panel. Oops. And let's just go ahead and change this to a bigger font so we can all read it. Oops. Select this. All right. And let me drag a slider down here. And, oops. And let me bring up an inspector. Let's see here now. And let me select the slider. And I'm just going to set this to go from 0 to 100 to make this simple. And then I'm going to take my connection. And just like I always do in Interface Builder, connect it to the text field. And we'll just say take an integer value. And if everything is working right, I should just be able to say test this. And here we are running Interface Builder inside Microsoft. We're finding that this stuff runs really well on Windows NT.

And it is exactly the same code, exactly the same code that we are running on Unix with obvious changes for integration with the Windows system, drag and drop, the print, the save, et cetera, et cetera. And your application source code is exactly the same. The draw application has 100% the same source code as it does for OpenStep on Unix. So we have a truly portable environment. A lot of our customers want to use the mixed environment between Unix and Windows. However, we must also say that OpenStep running on Windows has gotten us into some projects that these customers that we never would have gotten into were we not running on Windows. So we're pretty excited about this. We have a tremendous amount of customer interest. And we think it's going to be a pretty successful product.

Okay. So that's OpenStep. You can see it running in the booth. And I think you'll be pretty excited about it. And it is on schedule. As a matter of fact, we're a little bit ahead of schedule. So OpenStep is one of the major pillars of where we're going. And the OpenStep partners, Hewlett-Packard, Digital, Sun, us, are really excited about it. Now, in doing OpenStep for Windows, we wanted to be a great Olay citizen. And so we developed technology to let our distributed objects look just like Olay objects. As a matter of fact, really become Olay objects so that we have seamless integration between Olay objects and our objects.

And that's all in OpenStep for Windows. When we got this all running, somebody said, my God, you just run this on two Windows machines and you have distributed Olay two years before Microsoft. We thought, hmm, this is very interesting. And so we are bringing a product to market at the end of this year, which is distributed Olay. And it's hot. And we couldn't figure out what to call it. We thought about calling it just distributed Olay, but we figured Microsoft would have a trademark issue with that. So we couldn't call it that. And so we thought about calling it D-Olay, but we figured that same old trademark issue might come up. And we thought about calling it Dole or Dole or something like that. That was probably a problem.

We thought about calling it Bob, but then they went and trademarked that. So we figured they couldn't have a problem with Bob Dole. But in the end, we decided to provoke a little controversy and we decided to call it D-Olay, distributed Olay today. And let me tell you a little bit about this product. Now, this is what we have been shipping for the last four or five years. As you know, Next is the only company to ship distributed objects in production, even today. And we were certainly the first four to five years ago. And what we do today is you can have any client, whether it be Nextstep running on an Intel PC, Nextstep running on a Sun or HP Workstation, objects can message other objects transparently that live on other machines over the network. And you don't have to worry about the networking gunk and you don't have to worry about finding them and you don't have to worry about anything.

It's just as if you messaged an object that was right next door. And what most people do with this is they message objects running on servers. And so we have our object model, PDO, running on all the servers from HP, Sun, and DEC. And this makes it very, very easy to build multi-tier client-server applications. As a matter of fact, Nextstep is by far the easiest way to build multi-tier client-server applications because of this completely transparent distributed object model. Well, what we have done is this. We've said you run Dolay on your PCs and we'll be shipping it for Windows NT around the end of this year. And there's no reason we can't ship it on Windows 95 first half of next year. You run this on both your machines and literally all your applications do is make an Olay call to Dolay and they message their Olay messages distributed to other machines. That's it. Nothing special.

They just make Olay calls. They have to know nothing about Objective-C. They have to know nothing about PDO. They have to know nothing about Nextstep. They have to know nothing about OpenStep. All they have to do is know how to call Olay. Well, what knows how to call Olay? Well, Visual Basic, Visual C++, Delphi, anything running on a Windows machine. And it's even cooler because now on a Windows machine, it's really hard to write a client-server app. It's really hard to write a distributed application if you want to move some of the processing to another machine because all of these tools I've just mentioned like Visual Basic and Visual C++ have no notion of distribution at all. Well, guess what you can do? You can do this. You can write your front end on a Windows machine running using Visual Basic, let's say. Call Olay. Dolay will automatically pass that Olay call over to a server, let's say, an HP server running PDO and message back-end Nextstep objects on that HP server.

And the front end on the client knows nothing about any of the back-end stuff. And so, generically, this is what you can do. You can have Windows machines talking to Windows machines. You can have Windows machines talking to servers, Unix servers running PDO. Or you can have it talking to NT servers running Dolay. Or you can have it talking to OpenStep or Nextstep clients which have PDO embedded with them. Any permutation here works seamlessly. And I'm going to demonstrate two of these permutations to you. I'm going to demonstrate Windows NT and I'm going to demonstrate a Windows NT client to a Hewlett Packard server running PDO in just a moment. Dolay is in beta now and we'll be shipping it by the end of this year. Okay? So let's see if it works. I have two NT machines here. I'm going to show you how to do it. I am going to be driving from the one on the left and I have a simple OpenStep on NT application which simply has a host name over the network, three buttons, one that says connect, one that says create chart and a third pop-up that selects the type of chart and I have a slider here. So the first thing I'm going to do is push this connect button. And when I push it, it's going to go out over the net and send an Olay message to this other NT machine right next to me, the screen on the right, and which is going to launch Microsoft Excel. So I say connect. And this is over a standard local area or

wide area network TCP IP so it even works over the internet. So there's Microsoft Excel and I can push this button and it loads in a file, creates a bunch of data and builds a chart. Okay? And I can say I want a different type of chart. I can say, well, why don't you give me, you know, a line chart? Or why don't you give me a column chart? Or why don't you give me a 3D pie chart? And I can change this and actually update values in real time. You see up there? OK, totally distributed, NT to NT. This OpenStep for Windows applications, a Nibware app. This Excel is obviously Excel. So there we have it.

OK, now I'm going to move over to this other machine. And I'm going to launch Excel again. And now the first demo was client to client. The second demo I want to run is client to server. And as we talked about before, it's really hard in the Windows world to write these multi-tier client-server apps. So what we've done here is we've used Excel again. There's no more generic Windows tool than Excel. And we all know Excel has no notion of distribution. But it does call Olay. And so we've got about a page and a half here of Excel code. Let me just show it to you. This is our Excel code here. We don't have a lot of it, not a lot.

And it calls Olay. And when it calls Olay, that gets converted by delay into distributed messaging. And it goes to PDO running on an HP server, which has got a database. And so I can make a very simple app. And when I say fetch locations, it's just pulling up all my locations here for me, maybe a bunch of employees. And I'll pick Hillside 2. And here's a bunch of departments. And I'll pick customer service. And here's a bunch of people. And you notice it just computed an average group salary for me. And I'll pick an employee. And this is all fetching the data over the net from HP servers running PDO with Excel, making only Olay calls.

Does this make sense? Isn't this cool? OK, so that is Olay. And we think early indications from some of our large customers is this is going to be really popular. The next thing that we'd like to announce here today is WebObjects. Now, Next has a long history with the World Wide Web, as you may know. Tim Berners-Lee, who is the inventor of the World Wide Web, who did that work at CERN in Switzerland, did it all on Nextstep.

And so the World Wide Web was actually invented on Nextstep. In addition to that, as you know, we have always been compatible with the internet in terms of all the protocols we've used for distributed objects, et cetera. And we finally decided it was time for us to do something on the internet. And so we thought, what could that be? Where could we add a tremendous amount of value? And we started looking at the internet. And the internet is fundamentally a mainframe architecture, if you will. It harkens back to the 70s, where a web browser, think of it as a 3270 terminal on multimedia steroids. That's what it is. And all the wiring's done for you with the internet. And the servers like the mainframe. And all the computation, all the applications, run on the server. This is a very simple model, but it has enormous benefits in its simplicity.

So we looked at the client. And we said, you know, clients are going to be free. There's probably not a lot of money in client software. So let's put that aside, because we like being profitable. And the internet's wonderful, but a lot of people are working on that. But we looked at the server. And we said, the server's where the applications run. And that seems to need a lot of help. And we started analyzing that. And we said, you know, there are four things that people are going to want to do on the server with the web. And let's examine what they are. The first one we call static publishing. This is where you make web pages and you publish them. And they don't change until you change them.

And this is what 99.9% of the web is doing today. So here's an example with the Mercury Center News Service. Here's another example with the Monterey Bay Aquarium. Everybody's making web pages. And that's wonderful. But no one's making much money on this yet. Because to publish static web pages, you don't really need to go out and buy fancy web server software. There's 100 public domain packages out there you can get for free and just start publishing your pages. Now, the second thing that we're starting to see an inkling of is where this ain't enough, where companies now want to start providing some interactive value on the web.

And we call that dynamic database publishing. That's number two. Now, perfect example of that, how many of you have seen the Federal Express website for tracking a package? A lot of you, I hope. You can type in your package number here and give it the country destination and push Submit. And it will look in its corporate databases, several corporate databases all over the place, pull that information together, and on the fly build a web page based on that data. Nobody ever typed it in and stored it as a static web page. It's computers building it on the fly with no human intervention. And you get back something that looks like this. Here's your air bill number. The package has been delivered. And here's a bunch of information about when it was picked up, et cetera, et cetera. So they go from that to that with no human intervention, with a ton of data from a bunch of databases in a bunch of different places.

That took them four months to write that out, by the way. Here's another example. A real estate company on the peninsula named Elan Penel. You can plug in how many bedrooms of a house you're looking for, the price range you're looking for in the cities, and it will come back to you and show you all of the houses in its database that meet that criteria all over the web. A third example, WebCrawler. I'm sure a lot of you have used WebCrawler. It's the second most popular site on the web. By the way, it runs on Next Step totally. America Online just bought this company that did this. And it runs all on Next Step today. And what you do is you type in keywords to search on. It searches through a massive index database and comes back with places on the web that you might go to find what you're looking for.

Again, looking through massive databases, building web pages on the fly, and coming back. This is dynamic database publishing. And we're going to be seeing this explode, we believe, next year as companies want to bring this kind of service value to their customers. The third thing that people are going to want to do on the web will take a little bit longer to happen, and that's commerce, where you can actually buy something over the web, pay for it with a credit card, have it delivered with Federal Express, let's say. Now, the interesting thing about commerce is that it's a massive job. Everybody focuses on the credit card problem. That's a small part of it. The larger problem is medium and large size companies have to tie their web into their existing order management system.

Why is that? Well, you can't tell somebody that a certain Walkman is in stock on the web when you just shipped it out five minutes ago to Circuit City. You've got to have one order management system to keep track of all this stuff. And everybody's order management system is custom, and you've got to write a custom app to tie the web into your order management system. That's why it's going to take a little bit longer to do real commerce over the web. And lastly, but not least, is internal web usage. We see this exploding now. What's this all about? Well, there's simple publishing internally, but that's not what I mean here. What I mean here is that IS departments have got a great idea. When they want to write a custom app, like a time off app, let's automate the vacation reporting in our company.

Let's write a time off app. Today, we have to write a version for Macintosh, for Windows 3.1, for Windows 95, maybe for Unix. We have to distribute these all over the place. Why not write it on the web and let everybody get their own web browser, and we only write it once? And it's fully distributable, so our sales offices can use it too. And people are starting to look at this in droves. And so internal custom apps on the web are going to be a very big market, we think. So these are the four things we think people are going to be doing with the web in the next few years, static publishing, dynamic database publishing, commerce, and internal custom apps. What's interesting about this is that three of the four require custom applications.

Static publishing does not, but dynamic database publishing does. As I mentioned, that Federal Express app took four months to write. The others took longer. Commerce, major work to tie the web into your order management system, and obviously, internal custom apps by definition. Well, customer us, that's what we've been doing for 10 years. So how can we bring value to this? And this turned out beautifully. How do we help people take input from the web? process it, go to databases, get data, bring it back, process it some more, dynamically create these web pages and send them back out to the web in a tenth of the time that it would normally take to write these applications.

Sounds familiar, right? Well, we decided to look in our bag of tricks and pull out some familiar tricks, so here's what we did. Let's start with PDO, our robust dynamic object model. Runs on every computer by sun, deck, and HP, as well as 200 PCs, as well as, by the end of this year, NT. So it runs on everything. We can have a portable product if we start with PDO. Now, then what we'll do is, as you may know, last year we announced a product called the Enterprise Objects Framework, one of our most profound pieces of technology. And what the Enterprise Objects Framework does is it allows you to make next-step objects, with no programming, have persistent and coherent storage with SQL databases.

So you can go fetch data out of databases, you can keep it persistent in databases with no programming on the part of the developer. And we really think that saves a ton of time. Our customers tell us that. What we have now done is written a sister infrastructure product to EOF, the Enterprise Objects Framework, the Web Objects Framework. And what WOF does is it does everything that EOF does for databases with the web. When HTML comes into the web, it automatically parses HTML, builds a graph of objects, and shoves the data in those objects. When you want to write something out to the web, it automatically takes objects and dynamically creates the web pages from them.

To where the developer only needs to worry about the logic, the business logic. And they do not have to worry about SQL, nor do they have to know or worry about the web. They just drop their objects into this framework, and everything just works. And so with this framework, objects don't have to know anything about the web, HTML, HTTP, nothing. They don't have to know anything about SQL relational databases. And because of PDO, they don't even need to know their location. Matter of fact, you can distribute these objects running on several machines as your web traffic increases. And so this is a major new framework for us, which we think is going to be very hot. And again, because PDO runs on all these platforms, Web Objects runs on all these platforms, we are in alpha on Web Objects now with some selected customers.

We've just recently gone into alpha. We're going to be in beta in Q4, and we will ship this stuff in Q1. And remember that PDO and EOF have been shipping for a long time. And so the majority of technology in this product is existing, mature, robust, proven technology. So what I'd like to do now is try to show you a demonstration that captures the power of this. What I'm going to do is I'm going to go to an industry standard Netscape browser running on Windows NT. And we have written a web application with Web Objects. And our web application, we wrote this in about two days because we're still working with alpha code.

It should have been possible to write this in a day, and we think it will be when the code is final. But this took us two days to write, and in parallel with that, about two days to make it pretty. So that will not go down once we finish Web Objects. And so we borrowed one of our customers here, Dodge. Chrysler's a big customer of ours. And we said, welcome to the Dodge virtual showroom, and here's all the cars that Dodge has, and here's all the price ranges, and here's the different model types. And you can select by models or prices or type of car, and you can sort by model, price, or type, and then you can display the car. So let me go ahead and pick every kind of car they make, every price range, this, and let me go ahead and sort by price, and I'll just say display cars.

And this is going now over the network to an SQL database somewhere around here. I have no idea where. And it's showing me all the cars that Dodge makes. So this is all the cars that they make. And this page was dynamically created on the fly based on my request. Okay? So, you know, we could just stop anywhere here. Let's go to the top. And, you know, I could pick one like a Dodge Neon, and it will show me about that car. Right? So let's go back. And I don't want to sort through all these cars, so I'm going to go back to the top here, and I'm just going to pick Neons because I really want a Neon, and I'm going to pick maybe from, you know, maybe $8,000 to $20,000, and all of the options.

And I'll say display cars here. And, again, it's constructing this page on the fly, and as you see, there's a lot less cars now. Right? So these are the Neons. And I'll say, okay, I want to pick a base model. So it brings me the information on that one. I can say, well, tell me about the colors. Well, I'm going to pick, you know, I want to be noticed. So there we go. And this is a base four-door sedan, so my base price is $9,500. Options are zero. So I'm going to pick Package B. So I'll go ahead and pick Package B. And it's going back, and actually, see, there was no... These pages aren't stored. They're being computed on the fly. And so Page B has been highlighted. It's been... Package B, I'm sorry, it's been added up here in the options to calculate a total. Package 25B has been grayed out because I guess all that stuff's in Package B.

I'll pick Package 25A, let's say, too. And now I'm up to, like, $14,390 for my under $10,000 car. And now what I'll do is say, well, I can't afford that. I better finance it, so I'll pick this little dollar bills up here for financing. And so what it does is it shoves all that into a newly dynamically constructed page for financing. So here's my... I get to type in my down payment, and I get to select whether I want 36 months at 5% or... I better pick 48 months at 6%. And what I'm going to do now is push calculate. Now, we did something kind of fun here. The web server is running on a Unix server, I think an HP server, a WebObject server is. And what it's going to do, it's going to do something fun.

Remember, we have full Olay interoperability, right? And we have full distributed Olay. So here's what the server's going to do. This Netscape client is going to message the WebObject server. The WebObject server is going to message this NT machine here with Olay and have it do the loan calculation and calculate how much... Right? Look at this. On a completely different machine, calculate how much the loan should be, shove it back to the WebObject server, and then create this page on the fly and shove it back over the network to the web client. Applause Believe me, there's, like, no other web server software in the world that can even dream about doing this today.

So this is incredible. Let's put a down payment here on it. Maybe I can cough up, like, maybe 5,000 bucks. And hit enter. And it should recalculate it. Again, we're using three machines here, one of them running NT, Excel, one of them running NT in a web client, and Netscape, and a third in HP running WebObjects. Okay, here we go. And now, let's go to dealers. And one of the dealers locally has this car in stock. Again, that was all computed on the fly. But, you know, I've got a weird color. So if I take away the color, maybe there'll be more dealers that have that car. Sure enough, there are, if I don't have to match my lime green.

See? So aesthetics do matter. So there you have it. Let's go back. I can step back through all these pages that I went through because they're all stored in my web browser. And this is WebObjects. We think there's nothing else that's even close. WebObjects is going to be available on all of those platforms that run PDO. And again, we're in alpha now. We're going to be in beta in Q4. And we will ship this stuff in Q1. And based on everything we've seen, it is going to be by far the most powerful advanced web server software available. And we think a lot of corporations are going to want to use it so that they can rapidly build these web applications that are going to get new products and services to their customers over the web.

Okay. Now, you've seen our future. Three things. OpenStep, Dole, and WebObjects. What's interesting is, these are not three random products. They have all been driven with a common vision, and they all work together. We just saw an example of that right there with WebObjects server messaging, Olay, Excel spreadsheet doing the loan calculations, messaging back. There's one common vision that has driven all of this stuff. And first, of course, we start out with our vision of objects, that objects can dramatically revolutionize the creation of software by a factor of 10, by an order of magnitude.

leave that even more than the day we started with this stuff. We then spent the last several years building and perfecting our infrastructure products, which have become equally important to the objects themselves. First of all, PDO. And we spent as much time making sure this stuff scales to 10,000 seats and more as we have creating it. We then created major infrastructure platforms of enterprise objects framework and now web objects framework so that it's not just objects anymore. It's objects in this extremely robust framework. And the way you write your applications now are primarily writing your core objects in this infrastructure framework.

That's where you spend your real time. It's architecting your enterprise objects. And guess what? When you're done doing that, you can write front ends now in three different ways. The first way is using those objects. You can write an OpenStep front end running on Unix, running on Windows. Of course, this is true of NextStep today. Most of our customers are developing in NextStep today. And they will just translate those applications to OpenStep when it becomes available early next year. You can write it in OpenStep. But we didn't stop there. Because using those exact same enterprise objects and infrastructure technology, you can now, using delay, write your front end with Visual Basic, Visual C++, Delphi, anything that calls Olay.

You can write your front end in, have it called Olay, and use exactly the same enterprise objects and infrastructure that you use with your OpenStep app. Or now, if you choose, you can write your front end with the web using the web object framework, exactly the same infrastructure objects. So we have customers now that are writing NextStep applications and who say, you know, after I'm done with this NextStep application, I'd like to bend some of this functionality on the web. Maybe it's a reservation and guest management system. I'd like to bend some of that reservation data on the web so customers can look on the web and see what's available. Who thought it would take six to nine months to figure out how to do that. And we've shown them it's going to take a week. Because all this stuff works together.

And lastly, being that this is Object World and being that it is sponsored by OMG and being that OMG's shtick is Korba and being that Korba is now on Korba 2, we thought we'd have a surprise bonus topic, which is Korba 2. Now, we have a lot of people who are We have the only distributed object system shipping in production today. And we have enhanced that distributed object system this year to be fully interoperable with Olay. As a matter of fact, the fastest way to create Olay services is to write them in NextStep today.

Because all NextStep objects just look like Olay to Olay objects. We've added that and it's shipping next quarter. We thought, why don't we use that same exact technology that we developed for Olay and apply it to Korba 2. And we are doing that. It is in process now. And we are going to ship it with the next major release of PDO next year. And at that moment in time, PDO sort of becomes the first universal orb. It can speak PDO protocols, which is the most popular distributed object protocols in the world today. It can speak Olay, which is the most popular object protocols of all, even though they're non-distributed. And it will speak Korba 2, an emerging industry standard.

And that's what we've done. And so what I wanted to do now is show you how we've done that. We basically said, PDO, Olay, let's add Korba. So I'd like to show you two demos. The first demo I'd like to show you is Korba 2 to Olay. And the way I'm going to do this is that OMG was kind enough for this show to set up a server, a Korba 2 server, that plays a game. It's a very simple game. And it keeps track of the registrants and their scores of this game. And so any app you write to play the game has to talk to this Korba 2 server. So we're going to use Korba 2 to Olay. And then we're going to use Korba 2 to PDO to play the game a little different way. And the way in which we're going to play the game the second way is not just to stop at PDO.

We're going to go all the way to the web. We're going to go Korba 2 into PDO. And we're going to come right up through the Web Objects Framework and play it using the web. So Korba 2, we're going to have support in PDO 4.0 for what you're about to see. It'll be the first universal orb in the world, fully integrated with Olay and PDO, and full access to all of the infrastructure that we've built over the last 10 years. So let me demonstrate this. OK. The first thing I'm going to demonstrate is Olay to Korba. Olay, because we've got that old Microsoft Excel spreadsheet here again. And we're going to go ahead and register.

And it's now connecting with that Korba server. And I'll type my name. And I'll register. And it is now connected to the Korba server. And now I can play this pretty simple game. You pick a number between 1 and 10, right? So I'll pick a number between 1. I'll pick 8. OK. Oh, well, I guessed it. All right. Let's try again. Oh, I get to register again. OK. We need some usability testing on this. This game here. OK. 2. Small. OK. Good. 6, 7, 8. All right.

8. Could they have rigged the game for me? I don't know. So here we are playing this game on this Excel server. Olay speaking with Korba. Now let's try playing it another way. Let's go back to Netscape. Let's go to bookmarks. And let's play the Korba game using WebObjects. So this is a very simple app. We wrote it in a few hours. And it's now speaking Korba. So register new player. And again, I'm running Netscape on an NT box here. So I have to click. OK. Pick a number between 1 and 10. I'll pick 1. I hope it's not 8 again. Let's pick 8.

I think they're making it easy on me. So again, as you can see, here we are doing full WebObjects down to PDO to Korba and back again. So I think you get the idea. So in summary, we think we're building the future of objects. We think it's really bright. And we're doing it with OpenStep, the industry's first open object layer. We're doing it with Dolay, distributed Olay today. And we're doing it with OpenStep, the industry's first open object layer. And we're doing it with Olay, distributed Olay today. We're doing it with WebObjects, we think by far the most powerful web server software.

And we're doing it with a common vision so that all this stuff works together. And this is what that vision is. If you want to learn more about this, I'd advise two things. One, the keynotes at 9.45 and 11 o'clock this morning right in this room will be followed by a short break after this talk. And secondly, we have a new web server up. We put it up this weekend, www.next.com. And it's got a lot more information than we've ever had up on our web server before about all these products. There's a button about OpenStep. There's a button about WebObjects. There's a button about professional services and our consulting. There's a lot of success stories on it, of course, NextAnswers, and all the hot news.

So I would encourage you to take a look at it. And I think what you will see is there's a lot of great stuff coming in the future of objects. So thank you very much for joining us this morning. Thank you very much. You