2005 Computerworld Honors Program

	Digital publishing of documents Cisco Systems, Inc. San Jose, CA USA Year: 1999 Status: Laureate Category: Business & Related Services Nominating Company: Adobe Systems, Inc.
	Automated web site management takes new content from authors and adds links, generates indices, places navigational buttons, copies to multiple sites and pre-masters CDROMs, easing the task of keep web site content fresh.
	What do you call a company that has been programmatically publishing to the Internet for over four years? The leader. The vision of Cisco Connection InfoManager (CCIM) is two-fold: one, to simplify, streamline, and automate the process of Internet publishing (and now authoring), and two, to ensure that our customers get the information they need. For over four years, Cisco has been exceedingly successful in both areas, even as the scope of what we deliver has grown. CCIM Today At its core, CCIM is an application that allows Cisco's authors to define: -- Which files should be published -- What format these files should be published in -- Where these files should appear, and to which other pages they should be linked -- Which Cisco web sites they should appear on -- When these files should appear on these sites Based on this input from authors, and based on logic in the CCIM system built on well-defined business rules, the CCIM system automatically renders files into multiple formats (such as HTML and PDF), links parent files to their children, generates indexes, places navigational buttons on files, and copies the resultant content files to several destinations, including different web sites and staging areas for compact disk pre-mastering -- all with no further human intervention. From Publishing to Authoring Cisco has just deployed enhancements to CCIM that cross over from publishing automation into authoring automation. In its first iteration, the system allows us to significantly simplify the creation of one of our largest books -- the Cisco Product Catalog. This document has over 150 contributors, and historically, gathering consistent, timely product information has been difficult at best. To streamline this process, we designed a series of web pages that guides the contributor through authoring the various sections needed for the catalog. When the contributor has finished entering his or her product's data, the information is automatically cataloged and checked into CCIM, ready for the managing editor to retrieve, edit, format, and publish. No more endless voicemails, emails, and hard copy mark ups to deal with! We're excited about the prospect of extending this capability of CCIM to encompass authoring for other materials, as well as our next step into changing our authoring environment to deliver more timely and targeted information to our customers.
	CCIM allows Cisco's web sites to be managed with far fewer resources than would be required if all files were posted (and maintained) manually. The materials published through CCIM make up nearly 60% of the Cisco Connection Online (CCO) website, or about 75,000 pages -- and this is just one of the web sites Cisco maintains. To illustrate: the CCIM application publishes about 20,000 objects (files, illustrations, indexes, etc.) a month. Each object may be rendered into multiple formats and may be copied to multiple web servers or other destinations. Therefore, the actual number of objects that are manipulated by CCIM is over 100,000 objects per month. Based on the average human work month, this works out to roughly 11 objects delivered to a web site per minute. Clearly, programmatic intervention to process all of this information is the only way to quickly and accurately deliver content to our customers in a cost-efficient manner. More importantly, the Cisco customer wins from our use of CCIM by getting the most up-to-date information available. Because it is easy to publish a document, writers are encouraged to update their documents regularly. This ensures that the customer who is accessing CCO has the latest and most up-to-date information, including any new features, bug fixes, etc. Using CCIM, a writer can post a new or changed file on multiple websites across the globe in a matter of minutes, and have the exact same output appear in every location -- or the output can be customized based on a location's specific requirements -- all determined by the logic built into the CCIM system. Using a system like CCIM is indeed very different from how most companies post electronic files, where there is either a full production group who handles the manual posting and linking of files, or there is no formal methodology or system, and so all files are posted by individual contributors, with the resulting inconsistencies in look and feel, navigation, etc. Again, because it is so easy to electronically publish a file, it is not uncommon for an author at Cisco to republish a document several times in a few weeks, rather than through more traditional methods, where a book might only be published once. The result is far more accurate and timely information for our customers. Do our customers like this? Absolutely: "In the last two years, the online documentation and support provided by Cisco has vastly improved. I think it is the best site, provided by a vendor, I have ever used. Don't stop!" "INCREDIBLE! Through using the online documentation and having two routers in front of me, I was able to install and configure both for frame-relay....and they WORKED! Thank you to Cisco for the EXCELLENT online documentation! I only wish other vendors had such concise, yet in-depth documentation on their websites. Cisco's website should be a model for all. Thanks again for the WONDERFUL web site and the online documentation."
	Internet publishing at Cisco has always been driven first by customer need and second by productivity enhancement, so the system started as a business problem to which technology was applied. Having said that, technology is tantamount to the CCIM project. We kid our vendors that it takes nerves of steel to work with us, as we will push the envelope on any technology we adopt. We simply outgrew our first Electronic Document Management System (EDMS), and are pushing the envelope on our current EDMS, as we are now endeavoring to build a system that manages links as well as objects. In true Cisco fashion, if we can't buy it, we'll build it, rather than wait for the herd to catch up. We are committed to our customers, and to the Internet as the delivery vehicle, and we will do what it takes to get where we feel we need to be. As we move forward, technology will become increasingly important to our system -- and indeed, CCIM was built to be flexible enough to embrace emerging technologies. Our third generation publishing system (the first iteration of which we deployed in August 1998) begins to build on SGML (Standard Generalized Markup Language, ISO 8879:1986) content, in order to deliver more formats (including XML) and more highly targeted information to our customers by building an SGML content repository from which to pull content for automated, targeted, custom documents. Further, using SGML will increase the potential for content sharing and other single- sourcing strategies for our authors, as well as effectively divorce our content from platform and application dependence. Many people question why we are using a 13-year-old technology (SGML) as the cornerstone for our system, rather than jumping on the XML bandwagon. There are actually several reasons. First, it will be a while before XML becomes "normalized" enough to build an elaborate publishing system on, and at this stage, there is no guarantee that even after a standard is deployed, it will be strictly adhered to by vendors (witness the current problems with HTML browsers). This type of problem plays havoc with systems like CCIM. Secondly, XML is "enhanced" for the web, and while CCIM's focus is Internet delivery, we still need to support print, HTML, and PDF documents. So, we see XML as a delivery output, much the way PDF and HTML are outputs today. The upstream system power is delivered by SGML. And while SGML isn't enjoying as much press as XML today, remember that XML is in essence a "souped-up" derivative of SGML.
	One of the most exceptional aspects of CCIM is that it decentralizes the responsibility for publishing documents to the writers, yet maintains a level of consistency in look and feel not possible without centralized programmatic processing of files. We feel that the writer of the document is in the best position to know where and when a document should be published. Further, the granularity of control that CCIM offers, in terms of selecting different sites to send a document to and the specific date to send the document, is certainly one of a kind, to the best of our knowledge. Our recently deployed web authoring tool for the product catalog is truly unique in that it goes far beyond the typical web forms-type of application. This tool can be used to write (or revise) an entire chapter through the web, and then transparently to the content provider, capture the content, mark it up in SGML, and check it into our EDMS
	The third generation of CCIM (using our web authoring tool for the product catalog) was successfully deployed in December 1998. Building on the second generation's successes, whose primary beneficiaries were Cisco authors (better infrastructure, far better performance inside and outside the firewall), the third generation extends the benefits throughout the Cisco organization by streamlining the authoring and data collection process to the 150+ contributors. After several necessary "baby steps," our next iteration of CCIM will most closely approximates our original goal three years ago -- deliver custom documentation to our clients, on demand. This entails building an interface for our customers, available on our web site, where they can request specific content from CCIM, and preview, organize, sort, and store or print the retrieved information -- in effect, allowing them to "publish" their own book with just the information they want. Based on the customer information we have collected over the years, this will be an immensely important tool, especially for our customers who use the IOS documentation. With over 12,000 printed pages, this tool will dramatically increase a customer's ability to locate and retrieve the precise information they want online, quickly and reliably. Once we build this application for the IOS writers, we will begin to roll out similar authoring/publishing environments to other groups, with the same end in mind: more accessible content for our customers.
	Without a doubt, this project has had its share of difficulties along with its successes. The team has seemingly tapped out Silicon Valley's sizeable resource pool for qualified personnel, gone toe-to-toe with several vendors on technology issues, and faced internal resistance partly due to a change in organizational structure. We've had it all! Yet, we have persevered and have had significant success -- after three years, the vision is more clear than ever before, especially as we begin to implement some of the core infrastructure required for the SGML repository, we learn of more and more people who share our vision, and "enabling" technologies such as dynamic delivery tools and XML become more and more viable. Indeed, one main area of difficulty has been in trying not to outdistance technology so far as to develop tools and methodologies that cannot map back into a vendor's offerings, once they catch up -- to push the envelope of what's available without having to constantly build new envelopes from the ground up that won't work with someone else's mailbox. For example, when we started this project three years ago, we were repeatedly told by technology vendors that we were unique -- that no one else had needs like Cisco, that no one else would want to build a repository of information like ours, that no one would require the depth of link management, component management, and attribute management as we. As time has passed, we discovered many kindred spirits who either have built, or are trying to build, similar component repositories for much the same purpose. This has led us back to these same vendors -- if we, your Fortune 500 customers, believe this is the future of information dissemination, where are your enabling tools? So, in this regard, one of our chief difficulties has been that business demands (ours and other companies') have surpassed the ability of technology in this area, at least for the time being. This means we need to build "backward compatible envelopes" on our own, but it also means that there is a business opportunity for those with vision. We're heartened by the fact that XML appears to be helping to crystallize this vision in the vendor community and make it more mainstream more quickly.