What to Do About XML? Interest in XML (eXtensible Markup Language) is high and increasing all the time. It is hard to find a computer magazine that doesn’t have at least one article about XML. Meetings of technology and technical communications professionals need only announce that the topic will be XML to ensure a good turnout. The October 2000 Introduction to XML online class at Barnes and Noble University was completely filled--late arrivals had to be content to audit the class. XML books stay on the shelf in my local library about as long as the Harry Potter books. The DataDirect division of MERANT has for some years sought out new opportunities on the bleeding edge of technology. The company watches for new ideas from the Innovator group, and tries to have products ready for the earliest of Early Adopters (Rogers). This approach gave the company its current market leader position as a supplier of ODBC drivers, and explains our entry into emerging technologies such as OLE DB providers and JDBC drivers. The World Wide Web Consortium (W3C) estimates that by the year 2002, 75% of web document viewing will be through non-desktop devices like palm computers, televisions, and other non-“traditional” platforms. These platforms will require XML. Keeping up with technology, as usual, MERANT introduced an XML ODBC driver in November 1999, and an XML ADO provider in April 2000. Considering the intense industry interest in wireless technology, it would not be surprising if our documentation group has to support wireless access protocol (WAP) applications before long. These applications would be best delivered in XML. The Current Deliverable Situation We currently use FrameMaker 6.0 to produce documents in hard-copy and PDF; we use WebWorks Publisher to convert the files to HTML format. We will be switching to WebWorks Publisher to create the online help files as well in an effort to single-source our output, that is, produce all deliverables from one set of source files, instead of creating and maintaining multiple source files for different deliverables. We can create well-formed XML documents with FrameMaker, or create a WebWorks Publisher template to do a cleaner, more precise conversion. Instead of delivering HTML files, we could deliver XML files. This would be suitable for delivery to non-desktop devices. However, for our desktop audience, it is not clear what would be gained from the change. This plain-vanilla conversion would not add the content tags that allow a fine-tuned search by a web browser. Issues with XML Documents Ideally, XML documents should be structured, meaning that the documents are logically created, processed, and transformed, with each part of the document described in a defined and logical structure in a Document Type Definition (DTD)1. Lacking structure, an XML document is little more than a selection of paragraphs2. In the traditional editorial sense, our manuals are organized, they have a solid structure. However, they are not structured in the XML sense. To implement a fully-optimized conversion to XML would require a significant investment of time in analyzing the existing document base and tagging the files. Not only writer time would be involved. For the most effective analysis of the content, we would need to have developers assigned to define categories, identify requirements, and or at the very least, review our categorizations.3 In addition, although the newer releases of the primary browsers (Internet Explorer 5 and Mozilla5 and Netscape Navigator 4.75) support XML. Older browsers, and some of the less standard browsers do not. XML tools are not as sophisticated as we have come to expect for HTML or other online delivery. However, additional tools are being developed. Sarah O’Keefe of Scriptorium, an Adobe certified FrameMaker instructor, says that someone is developing a product “soon” which will be able to deduce a DTD from a FrameMaker document written in a non-structured format. That may be vaporware, but it has FrameMaker users excited. Another tool now shipping is W4F. W4F is a released product from PENN Database Research Group, which maps HTML to XML by generating wrappers as Java classes.4 Other issues may be waiting to pounce upon the unwary. For example, I saved this report as an XML file, creating a cascading style sheet in the process, and tried to upload the files to my personal web site at GeoCities. My file names included spaces, which are not allowed at GeoCities. I renamed the files, uploaded them again, and viewed my XML document. The document had no formatting: it was looking a cascading style sheet with a name that included spaces. It should have been an easy problem to fix, just open the XML document in the GeoCities text editor and change the name of the style sheet. The problem was that the GeoCities text editor could only open files with an HTM or HTML extension. I had to make the change in the text editor on my desktop, and upload the XML document a third time. This particular problem won’t be an issue for us and our manuals, because MERANT doesn’t use GeoCities to host its corporate web site. However, we may well encounter similarly unexpected problems when rolling out the XML version of our manuals. XHTML: a Reasonal Compromise A reasonable alternative to going directly to XML is to use XHTML 1.0, a new specification released on January 26, 2000 (Festa). XHTML is a reformulation of HTML 4.0 as an application of XML, which is to say that an XHTML document is an XML document that follows the HTML 4.01 Strict rules. XHTML documents can be displayed in most web browsers, as well as in non-desktop devices. And luckily, the new release of WebWorks Publisher supports XHTML, so our FrameMaker documents could be converted easily. Because XHTML documents are XML documents, they are extensible, meaning that new elements (tags) can be added.5 This means that even with a minimum amount of analysis, we could define elements that would make information easier for our users to find with a web browser. For example, we could define a System Requirements element for an installation manual. And as XML documents, they would be accessible for users with a non-desktop device. Conclusion Switching to XML now would be premature. XML support is not standard among the available browsers. Conversion tools are still evolving. However, with the increasing popularity of wireless and non-desktop devices, we cannot ignore XML. We can compromise by delivering our documents in XHTML: we can continue to compose in FrameMaker, and use Quadralay WebWorks Publisher to convert to XHTML 1.0. This allows us to continue to use our existing tools in our efforts to single-source our documents, making us more productive. Our documents will be in XML for the benefit of non-desktop devices, as well as in a format that can be viewed by most currently available browsers. Our documentation library is around 3000 pages of manuals, not counting online help. Clearly, the analysis involved in defining a structure for these documents will take time. A staged approach, one product library at a time, would spread the time burden. Using the lessons learned from the first product library, we should be able to make more rapid progress with the subsequent analysis, mapping, and processing of the remaining product libraries. Bibiliography Barnes and Noble University. "Introduction to XML." 25 September 2000. Barnes & Noble.com 08 October 2000 <http://www.barnesandnobleuniversity.com/Classroom/OfferingMenu/1,1473,25015_25001_,00.html> Bourret, Ronald. “Transforming relational databases into XML documents.” Updated September 2000. Home page. 02 October 2000 <http://www.rpbourret.com/xml/XMLAndD atabases.htm> Brennan , Elaine M. “Defining Attributes.” 05 October 2000. 05 October 2000. xml-doc@egroups.com. ClaBen, Michael. “XHTML 1.0: Where XML and HTML Meet.” 05 February 2000. internet.com Corp. 01 September 2000 <http://www.WebReference.com/xml/column6> Festa, Paul. “XHTML protocol get thumbs up Net group.” 26 January 2000. CNET News.com. 01 September 2000 <http://news.cnet.com/news/0-1005-200-1532441.html> Frentzen, Jeff. “Will you be ready when/if XML replaces HTML??” 05 January 2000. PCWeek online. 18 September 2000 <http://www.zdnet.com/eweek/opinion/0105/05jia. html> Lander, Richard. “XML: Time to Re-Tool.” 01 January 1998. University of Waterloo. 08 October 2000 < http://pdbeam.uwaterloo.ca/~rlander/XML/xml_trt.html> Martin, Didier. “Component-Based Page Layouts.” Feb. 16, 2000. XML.com. 25 September 2000 <http://www.xml.com/print/2000/02 /16/style/index.html> Martin, Didier. “Write Once, Publish Everywhere.” Aug. 16, 2000. XML.com 25 September 2000 <http://www.xml.com/pub/2000/ 08/16/didier/index.html> Richmond, Alan. “Introduction to XHTML, with eXamples.” 02 February 2000 Web Developer’s Virtual Library. 25 September 2000 <http://wdvl.com/Autho ring/Languages/XML/XHTML> Rogers, Everett M. Diffusion of Innovations. New York: Free Press, 1982. Tittel, Ed and Frank Boumphrey. XML for Dummies. Second Edition. Foster City, CA. IDG Worldwide Books, Inc. 2000 Wiggin, Peter. “XHMTL: Our last, best hope for clean code.” 16 July 1999. Webreview.com Cross-training for Web Teams. 15 September, 2000. <http://webreview.com/wr/pub/1999/07/16/feature/index.html> --------. “Ask the XML Pro.” Fall 2000. XML Magazine. 28 September 2000 <http://www.xmlmag.com/upload/free/features/xml/2000/04fal00/kc1_0004/kc1_0004.asp> --------. “Conversion Tools.” Updated 08 October 2000. XMLSOFTWARE. 08 October 2000. <http://xml software.com/convert> --------. “XML Is...for Content Providers.” 18 August 2000. Architag International Corporation. 03 October 2000. <http://architag.com/xmlu/xmlis/C ontentProvider.html> Lander, Richard. “XML: Time to Re-Tool.” <http://pdbeam.uwaterloo.ca/~rlander/xml_trt.html> -------. “Ask the XML Pro.” XML Magazine. Fall 2000. Available at <http://www.xmlmag.com/upload/free/features/xml/2000/04fal00/kc1_0004/kc1_0004.asp> Brennan , Elaine M. “Defining Attributes.” 05 October 2000. 05 October 2000. xml-doc@egroups.com. -------. “Conversion Tools at XML Software.” <http://xmlsoftware.com/convert/> Richmond, Alan. “Introduction to XHTML, with eXamples”. Web Developers Virtual Library, February 2, 2000. <http://wdvl.com/Authoring/Languages/XML/XHMTL/>