by Steve Gillmor

Interwoven manages enterprise content

feature
Oct 11, 200217 mins

President John Van Siclen and VP Kevin Cochrane discuss content management and the impact of Web services

INTERWOVEN’S CONTENT MANAGEMENT solutions help organizations manage a range of enterprise content, including documents, code, and Web content. Interwoven President John Van Siclen and Vice President of Product Management Kevin Cochrane met with InfoWorld Test Center Director Steve Gillmor and InfoWorld editors to explain how Interwoven differs from other content management players and how Web services and XML have changed the market.

InfoWorld: Tell us what you’re up to and explain how Interwoven is different from your average content management company?

021011hninterwoven_vansiclen.gif
Van Siclen: We announced [Team Code] in April, and it’s targeted directly at Web application management and the repository and virtualization and deployment services that are required for Web applications. Every one of our customers has something set up — usually it’s a secondary staging area, beyond the content staging area, where they try and bring the code together with the content through various scripts and test sequences, very manual. They are using some kind of system to at least version the code — PVCS or Visual Source Safe or something — and it’s sort of hit and miss. If it doesn’t work, it goes back and forth between the developers and the secondary staging environment until it works. At that point, they have scripts that try and push this code and content out to the production. And usually that’s through some kind of FTP arrangement that gets pulled down by the production servers and launched. Hopefully. [It’s] very error-prone and pretty fragile.

Team Code is targeted at the versioning and testing elements of the code, where it eliminates two or three steps in the process. It integrates the [code] developers with the content developers. Both can see each other’s changes and test their own functionality without breaking anything, so they know [whether] it’s going to work or not. We have all this functionality to proxy the actual production environment, the actual app server environment, and so we virtualize that and show all new changes in a real working context. You don’t have to hit and miss with the code changes. It’s going to be in general availability [in] looks like October.

InfoWorld: If I have business process integration and business process modeling tools all driven off a content management system and leveraging this Code Team approach, how does it all come together?

Van Siclen: We view enterprise collaboration as an opportunity for content management systems. We all have workflow systems, [but] we think about the content contribution, which is the other half for an enterprise. There’s the viewing time spent, [and] then there’s also the contributing or editing time spent. We view our world as that contributing, editing side. We see the need and the opportunity to enhance the collaborative facilities, because I believe it is the content management system that grows up into the collaborative framework.

InfoWorld: How have Web services and XML changed the nature of the game?

Van Siclen: I’ll give you the example of our Team Portal product, which [has] the ability to expose content services, versioning services, meta-tagging services, or deployment services through any enterprise application that can accept a SOAP, XML call and component. We’ve componentized major chunks and we’re now breaking it into smaller pieces and wrapping those kinds of content services in as Web services. It [makes] our integration with Siebel 7, PeopleSoft 8, Plumtree, with Epicentric, with the BEA portal, [and] with IBM portal, much, much faster. It’s a huge advantage. Same thing with the back end, when you talk about integrations with a variety of repositories, because companies don’t want to migrate everything into one master repository. That’s a waste of money today. But they want to provide our kinds of services, these pipelining services of discovering these assets and moving them out to enterprise apps. [And] those connectors to those repositories [are] much quicker through an XML, SOAP service.

InfoWorld: The content management space used to be a slugfest between you and Vignette and Documentum. How has it changed over the last year?

Van Siclen: We all still compete a fair amount, but things have sorted themselves as well. Documentum has clearly focused themselves back on the document management market, because that is growing well again, as opposed to totally stalled the way it was in ’99. I think that Vignette has a different challenge. Their challenge has been their schizophrenia as to what they do. Now that [former CEO] Greg Peters has moved on, that may straighten itself out. But they’ve been very clear that if you want to buy it all from one vendor, they are that vendor. We started to see that clarity and separation relative to Vignette about 18 months ago. I think a lot of it was [driven by] the industry analysts who would say “What do you want? Do you want to buy it all from one [vendor] or do you want a best of breed? [If] you want to use a WebSphere or Web Logic, then you want best of breed. Interwoven does that. You want to get all that from one [vendor]? That’s Vignette.” So it sort of cleared itself up and the market now is educated.

InfoWorld: Is there a fundamental difference between Interwoven’s approach vs. Vignette’s?

Van Siclen: [Our approach] is fundamentally different. In a database-driven environment like a Vignette or the document management companies, you have to retrofit your content to fit that data model. Our system is an object system, that’s why you can drag and drop any files in. Literally, in minutes you can have all the file system elements under management. If they’re database-driven, our system is schema independent so you can put all different kinds of schema layouts into our repository; it’s just a bunch of objects. It’s easy to get them in, easy to get them out. Last quarter SBC, a big customer that was a Vignette customer, became an Interwoven customer [because] it’s faster to move from Vignette to Interwoven and then rebuild your applications than it is to go from Vignette 5 to Vignette 6. The database-driven world really does require a lot of schema development, a lot of data modeling that our system just does not require.

InfoWorld: There’s this 800-pound gorilla in the corner called Oracle. What’s your take on what they’re doing?

Van Siclen: That’s probably the same question that BEA got a couple of years ago. [Oracle is] so focused on approaching this from a structured SQL database environment. This problem of unstructured information and managing it and moving this stuff out to applications has been around for a long time. Oracle’s fighting plenty of battles, and the one they’re currently fighting, the app server [battle], is a really big one. We have had a little bit of luxury; since we worry about contributors, we sit a little bit back in the development side. We’re not on the production side where the cash register gets rung big. It gives us time and opportunity to move and build out a platform that is quite robust, quite different than the database players. Over time I believe that the database players [will enter] this market and they’re going to come in the same way IBM’s coming in, which is through Data Manager, the DB2 group. They’re going to treat it as a repository play. Our focus as a company has always been on the services that are provided on top of the repository, where management is one of the services. We have a set of services for connecting to any repository, which is content processing. We have the enrichment services, which is our tagging services. We have the publishing services, which are XML-templating services, and we have the distribution services. We’re adding the code and content services underpinning all of that. It’ll be awhile before Oracle figures all those pieces out, and in the meantime [there will be] a lot of great cooperation with those guys. And with IBM, despite the rumblings from IBM, which is already probably a year or a year-and-a-half ahead of where Oracle is, in their thinking about this.

InfoWorld: How do you see IBM’s progress in terms of unstructured data and DB2?

Van Siclen: So far their progress has been about where Oracle’s is. It’s pretty light — more positioning than it is function, and it’s really driven by the DB2 group within IBM. IBM has some positive aspects to it relative to Interwoven, [and] then there are some challenging aspects. The most positive is the WebSphere organization: They understand the role of content management. The WebSphere sales organization that’s out talking about implementing e-business applications actually sounds quite sophisticated. They’re rebuilding some of the old enterprise applications with WebSphere as the platform. So they’re a great advocate.

InfoWorld: There seem to be a lot of open-source content management products out there — Sun, for example. How does that impact your strategy?

Van Siclen: First of all, Sun’s a good partner. We crisscross in a lot of places, [but] if you really look at who they watch, it’s Microsoft. So what they’re doing is trying to figure out how to drive toward that. There are opportunities to partner with them on a number of the things that we do. The content management open-source work at Sun is quite lightweight, like it is in a number of areas. And we have for awhile now supported a number of those open standards, if people want to use them and integrate those in with our system. But right now it’s nothing our customers or the CIOs that I talk to ask me about. They do ask me about our support for Linux, especially in the production environment where they have many small boxes now running their production environment, replacing Sun boxes. And we do support that Linux environment. We’ll have to see how it plays out and how the open source plays out as far as enhanced functionality in which areas. So right now [our] relationship with Sun is good. I think it can get better. There [are] some things that we can provide to help in their constant obsession with Microsoft.

InfoWorld: Everybody’s content management strategy grew up in an ad hoc way, and companies probably have half a dozen content management systems strewn across the organization. Can I put Interwoven at the center and have it manage everybody else’s content management system?

Van Siclen: The answer to that is yes. Most people don’t want to change [their] repository strategy. They would like to take the homegrown systems and get some business process under them, the workflows under them — because most of them don’t have workflow. If they’re going to replace the user interface, they want to have one that they can put together that mimics the user interface that everybody’s already using. The last thing [they] want to do is re-train. Some of our user interface strategies actually anticipate that [with] the portlet-like UIs that get snapped together in various ways depending on where somebody’s coming from and where they want to go. What people want is to normalize their index, their metadata about the assets, in one common framework. Maybe [have] multiple taxonomies that represent the view from different applications, but one set of vocabulary and index.

InfoWorld: There seems to be more happening at the edge of the network with data and content. Do you see synergy between products like EMC’s Centera and your framework?

Van Siclen: Yeah, there’s some interesting interplay there. We just caught up with some folks from EMC and I think that there’s some interesting things that can be done there. At the same time, I also have a lot of my customers that are trying to figure out how to reduce the amount of redundant storage that they have, which gets back to the re-use and re-purposing of individual assets rather than duplicating them throughout the enterprise. Like anything, the more we try and reduce the number of assets, the more we’re going to increase storage anyway. So there [are] opportunities on both ends.

InfoWorld: With XML and Web services we now can have thin clients that are rich. How do you see that changing the demands on the content management system?

Cochrane: One thing, one word is very critical when you think about Interwoven and that word is “open.” What is critical for Interwoven is enabling ourselves to be a data management platform for the enterprise, a true enterprise content management platform. And what enterprises are looking to do, where they have a dozen content management systems that they can’t necessarily rip out for political reasons, they still need to centrally define a set of core services: publishing services, metadata services, business process automation services, componentization services for transforming information into usable objects, and more. Those sets of services we look to afford to any enterprise application, be it a portal [or] a homegrown application that you wrote or a customer self-support application or your own content management system. Our entire UI layer is built upon our content services, which is a Web services interface that exposes metadata, publishing, workflow services, and more.

Moving forward, we will extend our API to include more componentized and modularized UI portlets. That’s what you see shipping from Interwoven today, with our Team Portal application. It’s nothing more than a couple of simple UI components that implement our content services. Our UIs moving forward are nothing more than simple components that expose our services. Services will be embeddable in your thick clients as well, so the integration that we have with the Microsoft Office environment implements our content services interface. You expose the final browser component of our UI framework, you expose the conflict detection component of our UI framework, and more.

The trick is we’re a set of services and we collect, store, manage, index, [and] componentize information captured from the end-user through our Web services [which are] integrated with Adobe FrameMaker, XMetal, Microsoft Office, StarOffice, and what have you. Then that information gets persisted lots of places. It’s going to get stored in PeopleSoft, it’s going to get pushed to Siebel, it’s going to get stored in a custom relational database schema that you have to support any one of your enterprise applications. Or it’s going to get dumped into our repository which will sit on top of an EMC box, a Network Appliance box, or we’ll write directly to something like an IBM Content Manager. We want to be able to take that same reusable object captured from an end-user when they do File, Save, and Submit in Microsoft Word, [and] componentize it and send certain elements to PeopleSoft, certain elements to Siebel, certain elements to 10 different custom enterprise application database repositories.

InfoWorld: Do you agree the average IT guy is very excited about Web services because it allows him to do .Net on the client and Java on the server and he’s just using it as a bridge?

Cochrane: That’s exactly right, that’s the one side. The other side is when you’re looking at the collaboration that teams of people do to create information for ultimate publication out to certain audiences, that type of peer-to-peer environment; we will be providing those services to our customers. That process of collaboration is what people do today in our environment with both Team Site and some of our new collaboration applications like Team Doc and Team Portal. As part of that collaboration, oftentimes I need to have my local workstation synchronized with the integrated content because I’m a mobile professional. If, for example, you’re working on your C drive and you drag and drop some asset into an Interwoven folder, I may need to have that on my laptop’s C drive as well, so that if I disconnect I can still work with that asset offline. With things like our Desktop Integration Suite, when you publish information you can persist it down to the user’s local drive as well. So it’s peer-to-peer in a sense, but it’s not pure peer-to-peer; it’s going to the server first. The trick is, when you put something on your C drive I still need to enforce things like security, business process, I still need to get the asset tagged. I don’t necessarily know that Kevin Cochrane should have that asset yet on his local drive, because it may need to go through an approval process first or it may not be suitable.

InfoWorld: Isn’t that similar to what Groove is doing?

Cochrane: Groove has a good thought, but it has the wrong idea. The great thought is that as soon as you’re finished with something, I have it on my own system and I can completely disconnect and have access to it. That’s a great thought. But the idea that information that you immediately put on your local system should be immediately available on mine is wrong, and it’s wrong for a couple of reasons. I need to enforce some control around that information you’re publishing; it may be very sensitive information. I need to understand what that content’s about, because I don’t necessarily know that Kevin’s the right person to access that information.

The major thing that our customers want is [for] information to be shared amongst everyone within the organization, but they do need to control the process by which information is shared and distributed. They’re looking to have a centralized set of services that enforce the business process, that enable the version control, that enforce things like componentization of the asset, and the tagging of the asset. The trick is you need to enable the peer-to-peer networking of content but you still need to have an intermediating central service that can determine whether or not this is something that should be transformed into XML and, if transformed into XML, componentized. And if componentized, who do the components go to? Do they go as XML into a relational database or do they get generated as a PDF and syndicated to Kevin Cochrane?

I think Groove [has] the slickest product I’ve seen in years. It’s very cool. And the great thing about something like Groove is, out of the box Groove works with our product. So if you wanted to do the peer-to-peer sharing of assets today using Groove on Interwoven, there’s no integration required, you just simply re-enable that. If you wanted to just do peer-to-peer, Groove is the best platform for pure end-user to end-user peer-to-peer. Interwoven and the resources that we’ll develop in our R&D don’t want to mimic and re-create Groove’s pure peer-to-peer functionality. Groove will work out of the box with Team Site; even if it becomes part of Office, we’re still going to leverage it. But where you’re really looking to transform profits and information, that’s a set of services that I personally don’t see someone like a Groove or even a Microsoft thinking about at a really deep level. That’s what our value-add is.