A tale of two book writing journeys

An age ago, I coauthored a book. It’s taken me thirty years to write another.

In 1994, XML didn’t exist yet, the World Wide Web was super-nascent, and the identity and access management discipline was premodern. I was an expert in the arcane incantations required to design Standard Generalized Markup Language (SGML) schemas, known as document type definitions (DTDs).

Jeanne El Andaloussi and I became fast friends ever since we found ourselves on the same side of my first standards table. She and I both advocated for the Open Software Foundation to use a standard SGML schema in delivering its software documentation “source code” to licensees such as our two companies (DEC and Groupe Bull).

The design methodology we codeveloped, and insisted on using to herd the dozen or so “OSF DTD” stakeholder cats, became the fodder for our book Developing SGML DTDs: From Text to Model to Markup—published exactly 30 years ago, in print for ten amazing years, and now accessible only through the link above. Bless the Internet Archive!

I’m just wrapping up my second-ever book project, Mastering Digital Identity: From Risk to Revenue, out real soon now. It’s targeted at enterprise executives who need help understanding why identity is critically important and what to do about it. (Follow along at the link to learn more.)

It’s an understatement to say times have changed. A few stories to illustrate…

Speed and urgency

The book I privately nicknamed “DSDTD” was contracted to be written in a year, and it took 19 months. While we felt some time pressure, the tech world still moved at a relatively stately pace then—at least as compared to now. Deliciously, over that time, Jeanne and I undertook several coauthoring sessions back and forth between my Boston home and hers in Paris.

During that time, I joined ArborText, a premier SGML authoring and publishing platform, to develop a DTD consulting practice. When I discovered that most of my clients were barred from accessing the new “WWW” at work, I realized we needed to add a sidebar in the book explaining what it was, along with its SGML roots.

These days, any book treating a technical topic has to be produced as fast as possible to stay relevant. Many are developed in full public view, for example on GitHub, serialized for early readers, and live-updated. I didn’t take that approach, but did complete the first draft in about four months.

Publishing

In the 1985-1994 decade, techdoc editing and publishing were my specialty, and I became an early adopter of the draft version of SGML (ISO 8879:1986) while working in the ULTRIX group at DEC. A passage in book #2 relates the zeitgeist.

Back then, technical documentation was trapped on paper. If Boeing wanted to ship aircraft manuals, they literally shipped hundreds of pounds of paper. We were creating a way to “write once, publish everywhere.” The same content could be converted into a printed manual, a CD-ROM, or eventually a web page (although the web didn’t exist yet). I was designing information-capturing languages that machines could understand—protocols that allowed systems run by different companies to leverage the same digital information in a myriad of ways.

Jeanne and I wrote DSDTD in SGML, using the DocBook DTD (to which I was an active contributor) and both ArborText’s and SoftQuad’s SGML editing software, proving interop. Our publisher had to find a specialty contractor who could translate all that through FrameMaker+SGML into something printable. Ron Turner of Soph-Ware was a lifesaver and became a friend.

Mastering Digital Identity was born and mostly edited in GDoc, with Canva for graphics. Easy peasy!

Timewasters

I was excited to be working on Windows 3.11, with its brilliant color palette, to write DSDTD. When I needed a break, there was always Minesweeper. Along with Tetris (Tengen version), I could close my eyes and play whole Minesweeper games behind my eyelids.

Very soon, the world was introduced to infinitely scrollable reading and entertainment. My timewasting habits have never been the same since.

AI

Back then…surely you jest. It wasn’t an option.

This time around, I made a point of ensuring the writing was 100% human-generated. Not because I’m allergic to AI content, though honestly I’m not crazy about the faint whiff of caricature about most of it. In the current AI moment, this was just something I wanted to know.

Subject matter

This is where DSDTD and—what shall I nickname the new one? MDID?—share the greatest similarities.

I keep being drawn to projects where I have a chance to enable both individuals and businesses to dictate their digital destinies—whether it’s about preventing data from being locked into proprietary formats (SGML, XML, and beyond), or about controlling where one’s own data goes (UMA, HEART, and beyond). Digital identity is absolutely a central part of this story, with implications for both risk and revenue.

I hope you’ll join me on the Mastering Digital Identity journey still in progress.