Encouraging Tag Set Branching without Creating a Briar Patch

B. Tommie Usdin

Mulberry Technologies, Inc.

Abstract

Customizing a tag set can be an easy way to get the vocabulary you need. It can also be a a journey filled with dead ends, trap doors, and slowly-revealed and difficult to identify problems. Like many public tag sets, JATS (the Journal Article Tag Suite) was designed to be customized. Our original expectation was that individual users would customize it, and while a few have done that to good effect, we have found that the major customizations have been by groups of users. BITS (the Book Interchange Tag Suite), NISO STS (Standards Tag Suite), and Manuscript Exchange Common Approach (MECA) are widely adopted customizations of JATS.

When users customize a tag set they expect to be able to use the existing infrastructure associated with that tag set, making changes to accommodate the changes they made. They often expect to intermingle their new documents with documents tagged to the original tag set and perhaps with documents tagged to other customizations of the source tag set. They expect this to work gracefully, easily, seamlessly. Sometimes it does, but sometimes it does not!

The JATS Compatibility Meta-Model Description was developed to help people who customize JATS create tag sets to create models that will coexist peacefully with existing JATS documents and with documents tagged to other JATS customizations.

It seems unlikely that the particulars of the JATS Compatibility Model will apply to other tag sets, but the principles behind the Meta-Model might be useful to other groups thinking about ways to make their families of tag sets flexible and compatible.


Table of Contents

Customizing Tag Sets
A Little History (over simplified)
Generic Markup
Bespoke Vocabularies
Public Vocabularies
Need for Customization Soon Obvious
Vocabularies Build In Customization Tools
Symposium on Markup Vocabulary Customization
Debates about Value of Customization
Customization Mechanisms Vary
Customizations
Experience with JATS Customizations
Many Customizations by Groups of Users
User Expectations for a Customization
User Experience
JATS Compatibility Model
JATS Compatibility Model in 2 Parts
Design Principles
Respect the Semantics
Prevent Semantic Mismatch
Linking Direction
Section Model
Common Section Models
Subsetting
JATS: Subsetting Always Allowed
Compatibility Properties in JATS Guidelines
Don’t Munge Elements and Attributes
Section-like
Tagset-Specific Constructs
Alternatives
Display Alternatives Example
Whitespace Handling in XML Documents
Maintain Whitespace Type of Source Vocabulary
ID IDREF
Property Catalog in JATS Guidelines
Properties not Relevant to Compatibility (as defined)
Social Behaviour
Metadata vs Data
Framework or superstructure
Recursion
These Guidelines are not Universal
JATS users can Ignore Guidelines
Other Groups
Ignore and Find Yourself in a Briar Patch
Note: the Briar Patch in the Talk Title
Questions? Comments? Suggestions?