The Last Balisage Conference
I’ve been chairing Balisage and its predecessor conferences for 30 years. In that time I have chaired 29 markup conferences. I really like occasions on which people who don’t generally interact get together to learn from each other. For many of our participants, markup is a supporting interest. They think of themselves as librarians or geographers or bankers or humanists or archivists or analysts of one thing or another. Most of the events they attend are focused on their primary interest. The tools they use, such as markup, are rarely discussed at these gatherings. Balisage is an event where people can show their markup-related achievements to people who will understand, appreciate, and sometimes make helpful suggestions about them.
I could open this, the last Balisage conference, with reminiscences. I could talk about the places we have met. Quirky conference spaces, fascist-style conference centers, buildings big enough that we all wished we had bicycles and small enough that there was no sound separation between us and the cheer-leader training on the other side of the air wall. Cities we learned to love and public transportation systems we learned to loath. I could talk about watching anxious first-time speakers become expert presenters. I could celebrate brilliant presentations and beautiful human moments. I could talk about games of werewolf and construction toys. I could tell tales of people behaving badly, because while we have had far less ugliness than many communities, we have had some unfortunate moments. I could have a wonderful, self-indulgent time, wallowing in the past. I would probably be the only person who enjoyed such a talk.
So, instead, I ask that you indulge your inner historian. Remember significant, funny, revelatory, or characteristic things about Balisage conferences or other aspects of our collective history. Write it. EDIT it. Edit it again, removing at least 1/3 of the words and taking your story to its essence. THEN post it in the Whova (conference portal) community discussion “Our History: Including Memories of Balisages Past”. Attach photos if you have them, AND if you have permission from everyone in the photo.
There is nothing so practical as a Good Theory
“There is nothing so practical as a Good Theory” has been the tag line for “Balisage: The Markup Conference” since it started in 2008. (Well, that’s not quite true. The conference committee has more than our fair share of opinionated editors. For a while our tag line alternated between “There is Nothing SO Practical as a Good Theory” and “There is Nothing AS Practical as a Good Theory”. We never published anything that said “There is Nothing So Theoretical as a Good Practice”, but it was suggested.)
With this as the guiding principle behind a conference about markup, Balisage has been a mix of case studies, updates on specifications, tutorials, discussions of best (and sometimes worst) practice, and conversations about how things should and/or do work. We have explored markup; tools used to create, manipulate and store marked-up content; markup-related information management principles; and formal and mathematical models relating to markup. Some of our content is solidly grounded in reality and the practices of daily production, some can best be described as imaginative. Each of these realms informs the other, often in unpredictable but rewarding ways.
Balisage has been a place where academia, industry, and government mix. We have been told people will only take us seriously if we get rid of the commercial representation. We have been told we will not grow unless we limit academic participation. We have been advised to reject government proposals because they lower the tone of the group. The mix has worked well for us.
All Conferences are Not Alike
What We Talk About at Balisage
For several years I spoke at an XML conference series that promised that attendees would take something out of each session that they would be able to use within the next month. Speakers were instructed to be sure that there were practical tips and implementable techniques in every talk. That was a popular series of events. While I don’t think every attendee took something immediately implementable from every session, I do think the emphasis on the useful created an atmosphere that many people appreciated.
Balisage has taken a very different approach. At Balisage we have talked about the logic (and sometimes the philosophy) behind our markup applications as well as sharing success stories, challenges, and experiments. We talk about high volume data processing with the same interest we talk about applications to manipulate what Josh Lubell called “Small Arcane Nontrivial Datasets” and some that manipulate datasets that are trivial by any measure, but also interesting. We expose people to ideas that most will never find useful, that some will find useful years later, and that may lead others to develop ideas of their own. We talk about things we are building and things we want to build. We talk about our successes and sometimes about our failures. We talk about currently popular technologies (anyone count the number of papers on the program over past couple of years about iXML?). And we talk about things that are off the beaten track, a bit weird, perhaps totally impossible. (There are a couple of talks on the program this year about things that I don’t think can possibly work and at least one that I think describes a spectacularly bad idea. We are going to talk about them because I may be wrong about the impracticality of these ideas, and even if they never come to anything, they may spark some interesting discussion.)
Balisage has been a place where people are allowed, or perhaps more accurately, encouraged, to talk about things “everyone” knows should not be discussed. This goes back a long way.
HTML
I remember a session at SGML ’94 about the audacious upstart called HTML. People were literally yelling about how much would lost if a “single DTD” model of the web (which is what HTML is) became widely used. How much damage to the community it would do if we allowed people to talk about HTML. HTML was an insidious effort to undermine quality markup and a direct attack on SGML. Also, it was an impossible effort, was bound to fail, and not worth the time to discuss. (There seemed to be little awareness of the cognitive dissonance between wanting to quash discussion of this danger to civilization and the certainly that it was totally unimportant.) We haven’t talked a lot about HTML at recent conferences, not because it failed, but because it has become ubiquitous.
Overlap
Balisage speakers make a lot gestures. Arms waving, fingers pointing to lines of code
and locations in graphs, and tiny spaces growing larger. But the most characteristic
gesture is what I call the overlap fingers.
Figure 1: Overlap Gesture Tommie Usdin making the “overlap gesture”
Extensible Markup Language (XML)
At SGML ’96, Jon Bosak, supported by a whole bunch of other SGMLers, introduced a subset of SGML intended to take the advantages of SGML to the web. This audacious proposal was greeted with a lot of skepticism and some hostility. People who had been working to promote SGML for years thought that any dilution of the single story that was the SGML standard would destroy any chance of SGML being widely adopted. There was some cautious optimism, and some real enthusiasm, but there was also a lot of defensive buzz. More than one person told me I should not have allowed that announcement at an SGML conference; encouraging that nonsense was stupid.
I don’t actually remember if these were the same people who insisted that the conference series be renamed “The SGML/XML conference” the following year. The year it was called “The XML Conference” banned all presentations about or discussion of SGML. (That was a step too far for me, I wouldn’t chair that conference.) At Balisage we are allowed to talk about SGML, and some of us will talk about SGML this year.
Topic Maps
Topic Maps seem to have gone out of style (actually, they never really went into style). We talked about Topic Maps at these conferences for years. Topic Map developers and users were never in the majority, but many of us watched as some very smart people worked out a specification and developed tools that used it. They had energy, intelligence, and creativity. There were people who thought Topic Maps would solve all information management problems. There were people who thought the whole notion was a waste of time. My opinion is that a really promising technology died of unclear messaging and internecine warfare. But, as you will hear later this week, there are still active Topic Map applications, and we will hear about one of them at Balisage this year.
AI and LLMs
It’s time to talk about the elephant in the room. In 2026 I am not getting phone calls advising me to ban discussion of SGML, or HTML, or Topic Maps, or even overlap. I am getting calls and emails decrying the use of AI and advising that discussion of it at markup conferences is irresponsible. People I know to be intelligent, well meaning, and reasonably analytical are conflating (and demonizing) all forms of AI and LLMs. There is name-calling and a lot of other unattractive behavior.
We have been hearing about the use of LLMs at Balisage for several years. Some of the papers have been thoughtful and well researched; some seem to have been thrown together in minutes and are totally absurd. I cringe when I hear someone who is clearly talking through their hat at Balisage. But that happens. With the best intentions we sometimes accept a presentation that turns out to be trash – a speaker who doesn’t do the preparation and bloviates instead of giving an informative presentation. I wish I could say this was limited to presentations on one topic, but it isn’t. Because we have had a speaker give a really bad presentation on a particular topic does not mean we should ban discussion of the topic.
Artificial Intelligence in many forms is with us whether we like it or not, and whether we recognize it or not. Large Language Models are with us whether we like them or not. Many of us are deeply suspicious of some of the ways in which Generative AI is being deployed and of some of the ways in which AI is being used in markup-based applications. I include myself in that contingent; these are tools that, it seems to me, can be used carefully and appropriately, but can also be used irresponsibly and harmfully. This does not, in my opinion, mean we should not learn about them, think about them, talk about them. On the contrary, this means that thoughtful discussion, while sometimes difficult, is important. We have been talking about AI, and LLMs specifically, at Balisage for several years and will continue this discussion this year.
Welcome to Balisage 2026
At Balisage this year we have the opportunity to hear about a lot of things that seemed to the conference committee to be markup-related. Some them are “how we did good at my place” stories, some are “what we are trying in my project” discussions, and some are “what I am thinking about this year” explorations. I expect that I, like you, I will find some of them a bit tedious, some completely incomprehensible, some irritating, and most fascinating and thought provoking. Let’s listen carefully, ask clarifying questions, and discuss thoughtfully and respectfully. And let’s have fun while we do it.
Welcome to Balisage.