Balisage 2009 - GODDAGS and EARMARKS, Just Ducky And We Love It

Fabio Vitali both figuratively and literally gave an animated talk addressing the problem of overlapping markup and the problem of modeling documents as trees. The title of his presentation, Towards markup support for full GODDAGs and beyond: the EARMARK approach, does little to convey how entertaining he made the subject. Let's just say he didn't duck and run for cover.

The fact that Vitali used a song by my all-time favorite band, the Fab Four, certainly got my attention. And I love it! His case study was a karaoke application which he postulated poses interesting markup challenges. First, the selected song requires pronoun changes based on the gender of the singer. Lines are displayed twice for a one-line lookahead. Chord changes do not exactly match line changes. And the final challenge is embedded fun facts that popup at appropriate points in the song.

Vitali's paper discusses his approach to these challenges -- EARMARK (Extreme Annotational RDF Markup), an OWL ontology with RDF triples. See his Submitted Paper and also his EARMARK site. See also the earlier work by C. M. Sperberg-McQueen and Claus Huitfeldt, GODDAG: A Data Structure for Overlapping Hierarchies.

Balisage 2009 - Streamabilty of XProc Pipelines

Norm Walsh (Mark Logic) gave a talk on streamability of XProc pipelines. XProc lets users define a sequence of atomic operations to apply to a series of documents, using control structures similar to conditionals, iteration, and exception handlers. XProc: An XML Pipeline Language is presently a W3C Candidate Recommendation that is near and dear to Norm since he’s been working on it for awhile. He hinted it should become a Recommendation this fall or certainly by Christmas. As per W3C policy, there must be 2 implementations before a specification is finalized. One of those implementations is by Walsh himself, called XML Calabash which is built on Saxon 9.

Streaming would provide a sliding window in a single pass with output beginning before all input has been seen. Little in said about streaming in the spec, but it is clear it could improve end-to-end performance in certain situations and would be essential for processing documents larger than physical memory. Although there are no explicit requirements for steps to be streaming in the spec, implementations will add value by enabling this.

Norm indicated that certain XProc instructions such a p:count are streamable, wheras others such as p:exec, p:http-request, p:validate-with-relaxng, p:validate-with-schematron, p:validate-with-xml-schema, p:xquery, and p:xslt cannot be streamable. His paper discusses data he collected collected by XML Calabash between 21 Dec 2008 and 11 Jul 2009 representing more than 294,000 pipeline runs. (His implementation has an opt-out, phone home feature so he can collect certain usage data.) In his Submitted Paper, Walsh concluded:

The preliminary analysis performed when this paper was proposed suggested that less than half “real world” pipelines would benefit from a streaming implementation.
The data above seems to indicate that the benefits may be considerably larger than that. Although it is clear that there are pipelines for which streaming wouldn't offer significant advantages, it's equally clear that for essentially any set of pipelines of a given length, there are pipelines which would be almost entirely streamable.
Perhaps the most interesting aspect of this analysis is the fact that as pipeline runs grow longer, they appear to become more and more amenable to streaming. That is to say, it appears that a pipeline that runs to 300 steps is, on average, more likely to benefit from streaming than one that's only 100 steps long. We have not yet had a chance to investigate why this is the case.

Balisage 2009 - Beer and Demo

Q: What is the preferred accompaniment for demos at a technical conference?
A: Why, beer and free food, of course! (Save your divergent opinions about that statement, guys!)

On August 11th, Mark Logic was generous enough to provide great quantities of liquid and solid nourishment at the Brewtopia pub in Montreal. The demo format was simple: 5 minutes each to plugin and go. Over a dozen eager folks braved the cramped space and a hot room, not to mention an increasingly rowdy audience (funny how beer contributes to that). The contestants and the names of their demos follow:

• Micah Dubinko: Zero to App in 5 minutes
• Michael Sokolov: Bibilical Studies
• Bruce Bauman: Conceptual Models to XML Schema
• Josh Lubell: Quality of Design
• Mohamed Zergaoui: XML Prague and XProc Designer
• Uche Ogbuji: Freemix
• Markos Z(?): XQuery in the Browsercreain
• David Lee: One-Line Web Server
• Quinn Dombrowski: Visualizing Bulgarian Dialect Data
• Betty Harvey: Archival Description (NARA)
• Steve Newcomb: IEML Parser
• John Snelson: Higher Order Functions in XQuery 1.1

[Also, Vyacheslav Zholudev volunteered to demo Presentational OMDoc but unfortunately couldn't get a working laptop in the allotted time.]

Betty Harvey (Electronic Commerce Connection, Inc.) and Quinn Dombrowski received identical cheers twice in succession (as measured by the highly scientific decibel meter) so they were declared co-winners, splitting the cash prize. By sheer coincidence, they were the only two female demonstrators. Draw your own conclusions. Everyone who participated was awarded a Mark Logic t-shirt.

Thanks to Mark Logic, especially host Norm Walsh and his colleagues, for a fun and "educational" evening. Thanks also the Brewtopia wait staff who had to wend their way through the tightly packed crowd all night.

Balisage 2009 - Spicy XML Data Services Platform

According to Uche Ogbuji, Akara is an integration platform for data services over the web providing pipelines for managing and processing data in whatever form (one format to another). In his own words:

Akara is an open-source XML/Web mashup platform supporting XML processing in an environment of RESTful data services. It includes “Web triggers”, which build on REST architecture to support orchestration of Web events. This is a powerful system for integrating services and components across the Web in a declarative way, so that perhaps a Web request could access information from a service running on Amazon EC2 to analyze information gathered from social networks, run through a remote spam detector service. Akara is designed from ground up to support such rich interactions, using the latest conventions and standards of the Web 2.0 era. It's also designed for performance, modern processor conventions and architectures, and for ready integration with other tools and components.

I have to admit, Uche Ogbuji's talk entitled Akara - Spicy Bean Fritters and an XML Data Services Platform was difficult for me to follow. Probably to many in the audience who have closely followed his earlier 4Suite work, this was exactly the right degree of spiciness, but it gave me indigestion. The pace was very fast (partly because he thought he had less time then actually allotted) and the slides were replete with acronyms and terminology that he assumed everyone understood (which may be the case, but still...).

Akara is built upon a mature foundation, namely the 4Suite code base including a port of the test suite. Uche said Akara is being used in some production environments although he also mentioned it was technically alpha code.

See his (late-breaking news) Submitted Paper.

Balisage 2009 - Those Pesky Namespaces!

Liam Quin (W3C's XML Activity Lead) gave a highly spirited talk on the pros and cons on XML Namespaces, as well as several approaches to simplifying namespace specification. He mentioned solutions proposed by Tim Bray, Micah Dubinko, and Ian Hickson. Quin's own solution was to store namespace declarations in a special namespace file that is processed by XSLT and applied to the files that reference it so they can in turn be validated with normal namespace syntax.

Apparently Liam must be into multimedia presentations. In addition to sporting a very colorful hat, he read a passage from any old book about railroads and showed a video clip of a damsel in distress tied to the tracks. Unfortunately, he ran out of time before we learned the fate of the damsel.

See his Submitted Paper.

Balisage 2009 - XML in the Browser: The Next Decade

Alex Milowski (Appolux) reminded us of the earliest demo on XML in a browser -- Netscape’s 1999 XML book demo from XTech ’99 in which you could sort by author, title, or ISBN using a combination of XML, HTML, CSS and JavaScript. At that time, Netscape also had an IRS demo with a table of contents in a sidebar controlling which page is presented (a la JavaDocs). While this might seem like old hat to us in 2009, Milowski ran the demos in recent browsers. The book demo worked in Firefox 3.x, Safari, Android, and iPhone, but failed in Internet Explorer 6, 7, and even 8. The IRs demo was less successful across the board with the exception of Firefox.

He defined Intrinsic Vocabulary as any markup that a browser can natively process with some well-defined non-trivial semantic without the aid of additional constructs. HTML is an example, but XML is not. He is particularly interested in intrinsic support for HTML5, SVG, and MathML, so he created a Firefox extension called XML Application Launcher. After you install the add-on, you can view Alex’s Balisage paper directly as XML in the browser using the Balisage DocBook subset rendered with a popup table of contents with load-on-demand pages. I tried it tonight and it works just fine! on the Google code page, he wrote:

The main idea of this extension is that you can write your own applications, distribute them, and use this tool to launch them based on media type, XML namespace, content matching, or some combination of those three. Eventually the extension will have access to a registry of applications for XML vocabularies so that when an unknown type is encountered it can query for supporting applications.

Milowski concluded his talk with these points:

• We must have HTML5, SVG, and MathML.
• Embrace the idea of intrinsic vocabularies.
• Replicate the browser extension model.
• Support open-source and make it easy to use.
• Don’t wait for someone else to implement it.

See the Submitted Paper.

Balisage 2009 - Opening Remarks and Sponsors

The Balisage 2009 Conference Committee -- B. Tommie Usdin (chair), Deborah A. Lapeyre, James David Mason, Steven R. Newcomb, C. M. Sperberg-McQueen -- opened the 4-day XML conference. One of the well-received announcement was their determination to make Balisage conference proceedings persistent. Unlike some other XML conferences which shall remain nameless (but not blameless), you’ll always be able to find all papers in the series. An ISBN has been assigned to each volume (2 per year), the entire series has an ISSN, and each individual paper has its own DOI (digital object identifier). How cool! Thank you, Mulberry Technologies!

The co-chairs acknowledged the two main sponsors: Mark Logic and the FLWOR Foundation. The FLWOR Foundation is dedicated to providing middleware and clients to simplify the use of XQuery. They have 3 open source projects under an Apache license:Zorba - XQuery processor, XQuery 1.1, update facility, scripting and REST extensions; XQIB (XQuery in the browser) is a browser plugin for Internet Explorer which allows execution of client-side XQuery to navigate and update the DOM; and an Eclipse plugin (XQVT?).

Of course, we all know Mark Logic because they've given us: MarkLogic Server, a native XML database that implements XQuery for the CRUD functionality with full-text and structured search; MarkLogic Application Services which includes Application Builder provides an intuitive, browser-based user interface for creating applications without writing XQuery code; and MarkMail.org, a public email search site built using Mark Logic App Builder which currently archives over 40 million searchable emails.

And Mark Logic is now also known for sponsoring a “Beer and Demo” (more about that later).

And let’s not forget those cool ergonomic pens donated by Patrick Russo (sp?). Don’t confuse them with a tuning fork or wishbone ;-)

Balisage, Come to Me!

Kicking off the conference today (after the logistics, of course), James Mason surprised Tommie Usdin with a song about Balisage. Not sure of the official title (could be simply "Balisage"), but it was set to the tune of Bali Ha'i from South Pacific. As can be seen from the photo, conference chair Usdin was delightfully surprised. Nice job, James et al!

Check back here for the lyrics...

Balisage 2009 - Review and Summary: Processing XML Efficiently

Mike Kay summarized and reviewed the symposium and added his own observations. He was a little surprised that there was nothing about Binary XML and little about speeding up XSLT or XQuery. Kay made the following main points:

Performance against other objective:

• Saxon: first standard conformance, then usability, then performance
  
• performance is not always the most important consideration
  
• David Wheeler (invented the subroutine)- “Optimize the code that users actually write”.
  
• Should you optimize for Joe User or the expert?
  

Performance metrics:

• Response time / latency
  
• throughout
  
• resource cost
  
• scaleability
  
• Jim Robinson: “Good enough for us” -- your own requirements, not the world’s best solution
  

Performance methodology:

• measure
  
• understand - figure out where bottlenecks are and what can be done to improve it
  
• focus on critical components
  
• improve -but if not, don’t leave in the things that didn’t help
  
• repeat until ok - what counts as ok
  

Where are the bottlenecks?

• application-level glue
  
• parsing vs. query?
  
• validation?
  
• conversion to/from XML?
  
• serialization?
  
• query/transformation?
  
• too many phases in the pipeline?
  

What improvements can we expect?

• faster parsing has ben promised?
  
• faster and more scalable transformation
  
• optimized pipelines?
  
• smarter users?
  
• faster hardware
  

Balisage 2009 - Efficient Scripting

David Lee, Epocrates (with a writing assist from Norman Walsh, Mark Logic) discussed Efficient Scripting (Submitted Paper). Scripting is useful for splitting complex workflows into manageable tasks that are easier to debug and easier to develop than with programming languages; the piece part solutions can be glued together later. Lee devised 5 kinds of test cases (2 were baselines) for 4 scripting languages: DOS and bash (Cygwin implementation), xmlsh and xproc (calabash is Norm Walsh’s early XProc implementation). The 2 XML scripting languages performed significantly better than the other two. With approximately 600 XSLT and XQuery operations, xmlsh and xproc yielded 50-75 fold improvement. With 3,000 operations, there was a 150-fold improvement. I believe Lee said there was a 150-200-fold improvement when there were approximately 30,000 operations. It should be noted that the chosen test cases focus on areas where scripting performs worst.

Balisage 2009 - Symposium on Processing XML Efficiently

Michael Kay kicked off the all-day International Symposium on Processing XML Efficiently by explaining that the morning would focus mainly on bottom-up hardware-based solutions up to pipeline processing, whereas the afternoon would be more top-down from the application programmer’s viewpoint.

The first presentation was by Michael Leventhal & Eric Lemoine, LSI Corporation, and covered an XML chip in development through this decade. The tokenizer developed by Tarari was originally based on the DOM which didn’t yield vast improvements. By 2004, a new approach called RAX (Random Access XML, simultaneous XPath) proved highly acceleratable, with 40x faster performance than conventional DOM-based. By 2005-6, this work became XMachine, consisting of a Rava stack, XSLT processor, and XTM (XML Threat Management) for streaming large documents. LSI acquired Tarari in 2007. By 2008, the XML chip was selected for HP SAP accelerator, yielding 3x to 4x improvement for enterprise applications. XMachine 2.0 is the 2009 version. The key to this hardware solution is providing a software stack above the hardware, the goal being to accelerate the overall software applications, so XMachine offers of an optimized API. The main design decision for hardware acceleration, they contend, is throughput vs. latency. The field-programmable gate array (FPGA) was a game-changer for their work. LSI has obtained performance numbers of 4.9 Gbps for complex XML processing with documents as large as 4 GB. They believe their new design will accommodate 1 TB documents. According to their Submitted Paper:

Results show that the use of an XML co-processor can reduce CPU cycles per byte of XML processed by amounts ranging from a factor of 3 to a factor of 50 depending on the workload, while power consumption can be reduced by a factor of 7.

One of LSI’s competitor's, DataPower, presented next. David Maze discussed the IBM DataPower XML XG4 processor card, the goal of which is to accelerate throughput and providing parallel processing with a card that can be switched to different processors. Their main use case is SOAP traffic in, access control checking, threat checking, and ensuring output is safe for the consumer of the messages. IBM uses a post-processing engine (PPE) that performs XPath evaluation, schema validation, XML routing, and filtering in hardware (i.e., to reject dangerous XML). DataPower has achieved throughput (and schema validation) up to 2 Gbps with up to 64K simultaneous documents and 128K simultaneous XPath expressions with as little as 15W of power. The XG4 has three kinds of output: DOM-like tree for the SOAP header, a SAX-like stream for the SOAP body called TLA (sorry, it’s just another Three Letter Acronym to me ;-), and post-processed data (pass, fail, status). Their PPE approach enables application-specific optimizations and extensions for new standards. See their Submitted Paper (abstract only as of this blog).

Balisage 2009 Blog - Introduction

Welcome to the beginning of my blogs for Balisage Conference 2009 in beautiful centre-ville Montreal. Check back daily for highlights of this technical XML conference. For each formal presentation referenced in this blog, the keywords Submitted Paper will link to the relevant publication in the conference proceedings. According to the proceedings:

Balisage is a peer-reviewed conference designed to meet the needs of markup theoreticians and practitioners who are pushing the boundaries of the field. It's all about the markup: how to create it; what it means; hierarchies and overlap; modeling; taxonomies; transformation; query, searching, and retrieval; presentation and accessibility; making systems that make markup dance (or dance faster to a different tune in a smaller space) — in short, changing the world and the web through the power of markup.

See also the Extreme Markup Community site.

Photo: Hotel Europa, Second Floor Entrance to Conference (credit: Ken Sall, c'est moi)