𝜏:torquens

Appearance

Token Management ​

A number of tools exist for token management. Many of them will support the features required to achieve the kinds of pattenrs described on this site.

However, setting them up manually is not to be advised. As most token management solutions do not support variable interpolation, some patterns will require a lot of duplication to implement.

As such, it is ideal to generate these tokens through scripting. Much of the information necessary is available within the structure of your existing semantic tokens.

What about my single source of truth? ​

It rather depends on what you mean by 'single', 'source', and, for that matter, 'truth'.

Many different parts of a design team may be responsible for different aspects of a library of tokens. Is the source of truth the output of these teams, or the documented (or undocumented) decisions that go into them? Which exact piece of information can you point to and say "this is it!".

It is better to think in terms of records of truth. This will come from multiple sources of truth, but coalesce into one definitive artefact. Against this, derivative libraries will be judged for their truthiness.

Maybe the real source of truth is the friends we made along the way.

cluster_0Sources of Truthcluster_1Record of Truthcluster_2Consumers of TruthdsnDesign PrimitivesColour ScalesBrandsSemanticsdsn->s:wpPrimitivesdsn->p:wtgsToken Generation ScriptsEncoding Design System rulestTorquenstgs->t:edeslDesign Librariest->desl:ndevlDevelopment Librariest->devls:e->tgss->tp->s

How you store your record of truth is immaterial—it could be a JSON serialisation of your tokens in a codebase, or a Figma variable collection (though ideally you would use a system with adequate change control measures).

That there are upstream tools that may determine the value of some of your design tokens does not detract in any way from your design tokens being the authoritative record of truth.

DTCG ​

The DTCG (Design Token Community Group) specification offers a huge improvement over previous token storage formats. The resolver module allows for the implementation of much of the concepts on this site.

However, it remains a poor authoring tool (no-one wants to sit down and write JSON), and still requires some duplication and denormalisation to describe torquens.

This should not be counted against it; maybe in the future it will have features like token interpolation for dynamic lookup, or mathematical operations.

Meanwhile, a DTCG collection of tokens can be considered the output of a process, as well as the input to others; it is a record of truth, of decisions made.