Extremism, 2006-08-08

Extremism: Day 1, Tuesday, 2008-08-08
David J. Birnbaum, djbpitt+xml@pitt.edu
The handsome gentleman to your left, who rules over the other denizens of the lobby fountain in the Best Western Europa Hotel, has been welcoming XML geeks to Extreme Markup every August since the conference first moved to this venue in 2004. He has probably forgiven us for the time that first year when, with the resigned permission of the hotel management (“are these people really adults?”), we took off our shoes (although not our neckties), waded in, and posed with him for a group picture (although perhaps we shouldn’t have been surprised when the management asked us not to do that again in 2005). This year there are between 100 and 120 of us, half of whom are speakers of one sort or another, and all of whom will avoid stepping into the fountain while en route to the Extreme Socializing in the Aqua Bar this evening.
The current 2006 conference marks the tenth year since the official announcement of XML at the Idealliance (then Graphic Communications Association) SGML ’96 conference in Boston (itself a tenth-anniversary conference entitled “Celebrating a Decade of SGML”), an occasion that provides an opportunity to reflect on the ways in which markup languages have developed since that time. With this goal in mind, over the four days of the Extreme Markup 2006 conference, this Extremism blog will identify memorable moments in the text of the presentations, in the general discussion, and in the corridors during coffee breaks. And because part of the pleasure of attending the Extreme Markup conference in Montreal is attending the Extreme Markup conference in Montreal, XML geeks who are also food geeks might want to visit the La Crême de l’Extrême postings under the “Restaurant” rubric at http://www.concretesyntax.com.
To this long-time Extreme (and pre-Extreme, since Extreme continues the work begun in many earlier markup-related conferences ) veteran, the 2006 conference feels different from its predecessors first and foremost because of the physical absence of Conference Chair Tommie Usdin of Mulberry Technologies, who is unfortunately unable to attend, although she is following the conference closely and with her usual eager interest from Maryland, and looking forward to returning in person next year. It was at a previous Extreme Markup conference that Tommie opined that “XML has made true all the lies we told about SGML,” an insight that I quote whenever I need to explain to the uninitiated all the wonderful things one can do with an XML document. That’s almost as often as I quote Tommie’s Extreme Markup 2004 keynote address, “When ‘It Doesn’t’ Matter means ‘It Matters,’” which argues persuasively that, among other things, when two ways of doing something are informationally equivalent, there are substantial advantages to permitting only one of them in circumstances that value interchange and interoperability. This may strike some as too obvious and self-evident an observation to bother sharing with a sophisticated Extreme audience, but for some of us, experience suggests otherwise.
Still much too young to vote, to drive, or to drink alcohol, XML may feel like new technology, but ten years of age is a long time in geek years, and XML has substantially overtaken its semi-retired SGML parent. The wonderful XML functionality that Tommie praised comes at a price, and although no Extremist would give that up, it is sobering to compare the proliferation of standards and drafts and work groups that begin with “X” on the W3C web site today with the elegance and simplicity of XML as promulgated in Boston ten years earlier. Anyone remember design goal #6: “XML documents should be human-legible and reasonably clear”? That goal can still be considered actual as long as we concentrate on modest textual document instances—but it is much less so when we think about three competing schema languages with four syntactic models, some of which use verbose XML, a programming language written in verbose XML (stay tuned for Sam Wilmott’s “Rethinking XSLT” on Wednesday), or any of the other wonderful new developments that make at least some XML documents—the ancillary ones we use to author and transform our XML, and that provide the functionality that constitutes the practical success of XML—less legible (at least to this human) and less clear than we once hoped would be possible.
Tuesday’s papers addressed the virtues and the costs of XML-related technologies, and ranged from forward-looking retrospectives by experienced industry veterans (e.g., Joseph Gangemi’s identification of SGML functionality that is missing from XML and that is missed by users in the publishing industry) to hot development areas (you can’t cross the street in Montreal this time of year without tripping over a Topic Map) to emerging technologies (e.g., Norm Walsh’s late-breaking report, fresh from the second face-to-face meeting of the W3C XML Processing Model Working Group, on a standard for expressing XML pipelines). Extreme Markup (official motto: “There is nothing so practical as a good theory”) has always welcomed both the abstract and the concrete. Those drawn to the former were well served by the latest installment of Allen Renear’s on-going examination of the FRBR properties of XML documents, while those at the other end of the spectrum appreciated Peter Flynn’s examination of human-system interaction in XML editing environments, which bases its theoretical reflections on specific user experiences with over two dozen structured-document editors.
Meaningful moments:

• Eduardo Gutentag gallantly stepped into the opening keynote slot originally allocated to Tommie and presented a markup technologies perspective on intellectual property in his “Extreme IPR.” The entertaining tour of silly patents (try to guess why you would need a toilet snorkel before looking it up) led to a survey of some of the absurdities and inanities of patent law and practice, and then to reflections on the inherent tension between patents (which claim ownership) and standards (which propitiate sharing). The situation is complicated by submarine patents; open source licenses cover copyright, but not patent, and while entrapment by developers is unlikely, nobody can predict what their heirs will do. Eduardo’s recommendation is that if you can’t beat them, you should join them, and that since the world permits interface patents, it’s best that the good guys own them. The Open Source Patent Commons, which just passed its first birthday, suggests possible directions for coordinated action.
• Readers of the XSLT developers’ list will recognize Dimitre Novatchev as the author of an XSLT script that affects an important aspect of information technology that is of vital interest almost everyone: sudoku puzzles (a hot topic of conversation at at least one table during the lunch break). In today’s “Higher-Order Functional Programming with XSLT 2.0 and FXSL,” Dimitre described an update of his Functional Programming Language for XSL (FXSL), which makes nearly all of the standard functions of XPath 2.0 and XSLT 2.0 available as first-class objects. This presentation struck some in the audience as providing elegant and simple (in some cases one-line) solutions to otherwise complex problems, but it provoked strong reservations from Ann Wrightson, who opened her response by asking bluntly “Are you really expecting people to use this?” and went on to caution against a complexity that risks breaking tools, challenging maintenance, and engendering performance problems because “not only is it very clever, but it looks like stuff that isn’t clever, and it risks performance issues with XSLT processors that do not expect this type of code.” (Ann also provided some contextualizing metadata: “I’m sorry; I’m being English and polemic.”) Dimitre replied that FXSL expressions are clear and compact, and are easier to understand than long text, and he suggested that one measure performance not only in the few milliseconds of execution time, but also in the hundreds of hours of developer time.
• James Mason reminded us that we have been discussing RDF and Topic Maps for years, and Steve Pepper alluded to the “shoot-out” between Eric Freese and Eric Miller at an earlier Extreme conference as part of his“Reusing Data across RDF and Topic Maps.” Many of this year’s Extremists are attending for the first time, and for the benefit of those who haven’t watched the history unfold, Steve outlined clearly the similarities and differences between the two approaches and then reported on the current state of the rapprochment that has been undertaken by the RDF/Topic Map Interoperabilitiy Task Force.
• Allen Renear’s “Is an XML Document a FRBR Manifestation or a FRBR Expression” answers its title question as “yes, no, neither, both.” Continuing a theme he introduced in his presentation at Extreme 2003, Allen concluded that FRBR expressions might best be considered not rigid types, but socio-linguistically contextualized roles, and from this perspective, XML documents can be seen as symbol sequences that can play different roles in different social circumstances. After all, as Ann reminded us, a particular XML document may represent, for example, both a work (an ancient Greek play) and an expression (a critical edition of that play). Allen’s presentation may have struck some newcomers as esoteric and without immediate practical application, but decisions about processing XML documents in the real world presuppose assumptions about what those documents are and what they represent, and reports like Allen’s, which exemplify the way Extreme welcomes and encourages presentations on the philosophy of markup, make us aware of those assumptions and help us learn to make them explicit.
• Joseph Gangiemi is an SGML veteran who built generalized markup systems for the publishing industry before most of us know what generalized markup was. In the first polemic of the conference, Joe observed that although XML, as the heir to SGML, ultimately emerged from the publishing world, its emphasis on simplifying parser development led it to favor applications that are well suited to web delivery, system-to-system communication, and non-textual data, but that create obstacles for the publishing industry. The omission from XML of some features of SGML that were created to facilitate otherwise expensive keyboard input (shorttag, minimization) could be compensated at the application level, so that, for example, we can no longer omit end tags, but we don’t have to type them because the editors insert them for us. But the omission of inclusion and exclusion exceptions from XML, on the one hand, and the omission of entity references from XML schema, on the other, create serious obstacles for the publishing market, and Joe argued that these features should be admitted by parsers because the cost of doing so is slight and the benefit substantial. Joe’s conclusion was that SGML will not advance, but it also will not become obsolete unless XML is enhanced to take its place in a way that also meets the needs of the publishing industry, its ultimate progenitor. Surprisingly, although Extreme reserves time for discussion of regular papers, it does not do this for polemics, the presentations that one might expect to generate the most lively and controversial exchanges, so those who might have wished to respond to Joe’s appeal (including those with substantial influence over the development of the XML standard and supporting software) had to do so out of earshot of the rest of the interested and engaged audience (although Joe’s nocturne later in the evening drew an impressive turnout, ultimately compensating for the lack of opportunity for general discussion at the polemic itself). Should what are currently monologic polemics be reconceived in future Extreme Markup conferences as debates with extended discussion opportunities?
• Peter Flynn’s “If XML Is So Easy, How Come It’s So Hard to Write?” notes that new authors who are advised to use XML typically don’t care about markup, but they do care about ease of use, appearance, and a familiar interface. A key issue is the obscurity of hidden markup; users who don’t want to see markup and think their editors should intuit their intent need to interact with markup even though they may not realize it. As one might have expected, Peter’s survey found that experts tended to recommend software that would be familiar for their client users (thus minimizing training and support), while for their own use they preferred such editors as XMetaL, Emacs, and Epic, and rank-and-file users almost unanimously demanded WYSIWYG, WYSIWYG, and more WYSIWYG. Ultimately, Peter’s report concluded that as far as editing XML is concerned, the editors all do pretty much the same job, but no single structured editor is immediately suitable for use by non-XML or LaTeX experts. In the following discussion, Debbie Lapeyre said that consultants typically customize editors for their clients, but in many situations, users operate without consultants, and seek software that will do what they want out of the box. John Cowan reminded us that at least the current generation of authors were not born using Word, and sometimes the correct solution is to upgrade not the software, but the author, citing as an example “At my company there used to be employees who could not use the telephone, but they no longer work here.”
• Norm Walsh’s presentation on XProc, the XML Processing Model Language, was truly late-breaking, reporting on facts that did not exist as recently as a week ago. There are already ways to, for example, pipe a document through an XInclude processor and then a validator, but the methods are often complex and they are poorly standardized, which complicates interchange. One of the pleasures of Extreme Markup is hearing up-to-date reports by the developers of the standards we all use, whether in official presentations or in informal conversations during lunch or the coffee breaks. Should Extreme build in and encourage a regular set of “What’s New in XML-Related Work Groups” component to the conference?

Odds and ends:

• Ever notice that at some conferences, some papers are more polemical than others? You’ll notice that at Extreme Markup we’re a cranky bunch, and such papers are plainly identified as “polemics” in the program. (From Debbie’s welcome address.)
• Think the Program Committee made a mistake when they rejected your paper? Nocturnes, complete with rooms and audiovisual support, are available for the asking every evening.
• You would think that information scientists would have remembered to indicate in the program which session is meeting in which room. Nah …
• If you attend a session in the Incognito Room, do you have to turn your badge over? (Not entirely a joke; the Incognito Room used to be the only meeting room in the hotel that did not have a sign identifying it.)
• Those who attended Peter Flynn’s presentation will probably never forget the importance of using a spell-checker that knows what a “filetype” is.

Kewl toys, games, and books in the Mulberry suite this year:

• Funrise Toy Corporation’s Gazillion Bubble Machine (see photo at right)
• Wham-o’s original Superball (not quite as super as I remember it being in my childhood, but it took less bounce to get over my head in those days)
• Play Visions’ Buzzing Stunt Magnets (note to self: keep these away from Debbie’s laptop next time)
• Jon Allen Paulos’s
  Mathematics and Humor
  
• Tom Peters’s
  Design