I am at the KMWorld conference in Santa Clara.

This panel presentation has three speakers: Chip Gettinger (VP, Customer Solutions, Astoria Software), Vernon Imrich (CTO, Percussion Software), and Mike Draper (VP of Information Systems, Network World Fusion).

Chip is up first with his slideshow titled "Collaborative XML Content Authoring and Management: An Astoria Software Case Study".

This is a case study about Texas Instruments, who has a very mobile, worldwide, workforce. TI realized that the delivery / collaboration problem they thought they had was actually an authoring problem.

Some of the business challenges: had been using a system from Interleaf to develop their technical datasheets for their DSP chip products. Datasheets were not going out as accurate as they should have. Conflicts with customers around errors in datasheets.

The other issue: redundant information. TI looked at how to make use of similar information. DSP chips have similar information across chipsets.

Lack of control of update cycle was another issue. Lots of changes from engineering for custom DSPs for customers, and lots of pressure to deliver sheets. They needed to deliver more datasheets, and also at a time when having to cut back on staff. Also, production cycles needed to be reduced.

Also, looking for more collaboration among engineers. They have engineers around the world contributing to this DSP datasheet process.

Also, PDFs were not good enough for datasheet delivery. They needed to deliver in XML, plus also print versions. And, needed to quickly update documents with updates from enhancements, etc.

Finally, they also needed to reduce the review cycle time by engineers.

Business justifications, primary ROI: improving engineering resources. Engineers were most productive designing chips, not producing documentation. So, separate documentation group that work closely engineers. But, engineers spending 15% of time creating documents, and 33% of that time in the review cycle.

Demand for easy to use authoring and review tools. Cohesive business process for authoring content, managing and reusing content, managing the review and approval of content, and automate the assembly and publishing of content. Also, key function for product lifecycle management, ultimately (in the future) including feedback from the customer for collaboration.

Legacy process with Interleaf, was very serial from authoring to review to annotation, all around paper and PDFs. Challenges: not easy to see all the comments from all the engineers, and keeping up with all the iterations of datasheets changes in the review cycle. One document, different engineering groups updating the same document.

Then, they would go into production and delivery process, originally with paper, then later PDF, HTML, and finally XML.

TI put out CMS requirements RFP, and Astoria and Corel partnered to create a solution. Their requirements were very authoring centric: wanted content authors to drive the process. Early on, TI made a commitment to XML.

TI wanted an online environment (web browser) for authoring and commenting. Good infrastructure for supporting people remotely, and didn't have to install client software on desktops.

Had typical CMS requirements: versioning, access control, security, etc. But, they also wanted to be able to do annotations to XML document through the web browser, and to have those comments stored with the document itself. Content was important, but context of document changes was even more critical.

They also needed a workflow component. Decided they needed very task oriented workflow management. Basically, they wanted to use email to notify people when tasks needed their attention.

Also needed paragraph level access control and security on XML documents, in terms of views, adding, editing, and deleting content. This also to allow partners to collaborate on documents.

Finally, they wanted a single source from which to produce HTML, XML, PDF, and paper. Industry standard XML interchange (RosettaNet) was changing, so they needed to be able to go from their source to whatever the standard interchange format was.

New process. From engineers to tech pubs all within the CMS. Engineers' word files imported into CMS, and review cycle all in the CMS. Corel XMetal used for tech authoring and publishing, to simplify working with XML documents.

The engineers started to really use it, and liked it. They wanted to change the content in the web browser as well. Altova Authentic XML editor Active-X component used to allow XML editing in web browser. All changes tracked in CMS.

TI implemented in a phased approach. Phase 1: proof of concept. 8 months, 5 people who designed DTD and started authoring a brand new product line, got 5 engineers to use the tool for review cycle). Worked out things that didn't work, replaced them with things that did.

Phase 2: test group pilot. Did some production datasheets. Lasted 6-12 months.

Phase 3: full production. Rolled out toolset to all the appropriate folks in the organization, and have phased out the legacy system. Converted ongoing product datasheets into new system.

Key lessons: understanding the users process was critical, especially with people who don't work together everyday. Relationship with IT group who dealt with network administration issues with new system.

Expected and unexpected results: reduced content review and approval cycles, which did have visibility at the executive level. They had executive sponsorship, but needed to have quick ROI, which they got. Streamlined collaboration process between engineers and tech pubs.

Sharing of information is growing at a great rate. People look at what other groups are doing, and new forms of sharing across datasheets is occuring.

Increase the independence it takes to make change in the organization. TI's orientation was not to be tied to old process, but instead to think about the business requirements and create a new process.

Demand for digital delivery has gone up. PDF is now second, and digital data (XML) is now the principle format for document delivery. But, they still need to have paper books. TI is evaluating their metadata to help automate the assembly of new, unique documents. Content reuse, where what used to take months to assemble manually can now be done automatically with a quick review cycle.

He is showing the web browser view, which is generated as HTML via XSL from the source XML.