Evaluating

In what ways is the data-driven syntax better than a custom string syntax? In what ways is it worse?
In what ways is it better than a "fluent design" (using functions and macros)? In what ways is it worse?
How easy is it to programmatically compose parts of a program into a working whole?
How easy is it to read vs. write vs. modify?
Is there an "escape valve" to allow users to do something in a custom way?
How well would the syntax "scale" with a larger problem?
After part of the syntax is learned, how easy is the rest to guess?
Is there any possible ambiguity between components of a program?

Cognitive Dimensions of Notations

Below we've adapted several questions from Thomas Green's 'Cognitive Dimensions of Notations' a set of design principles for notations.

Viscosity

How easy is it to make changes?
Do small changes in one place require changes in many others?
Are there particular changes that are more difficult to make?

Diffuseness

How verbose is the syntax?
What sort of things might take more space to describe?
Would it be longer or shorter to do with a general purpose language?

Hard Mental Operations

How much 'cognitive load' does this syntax require of the user?
How many degrees-of-freedom are there in the syntax? If you had to write a parser, how complex would it be?
How complex is the mental-model of the underlying "machine"?

Error Proneness

To what extent does the syntax influence the likelihood of the user making a mistake?
What is the most likely mistake a user might make with the syntax?

Closeness of Mapping

How closely does the syntax correspond to the problem domain?
How much effort does it take to translate a problem to fit within the syntax?

Role Expressiveness

When reading the syntax, is it easy to tell what role each part of the syntax plays in the solution as a whole?
Are there some parts of the syntax more difficult to intepret?

Hidden Dependencies

If some parts of the syntax are related to others (ex. a change to one requires a change to the other), are those dependencies visible?

Progressive Evaluation

How easy is it to get feedback on an incomplete program? (ie. Is it possible to try out a partially-completed version of a program, and compose it in stages?)

Premature Commitment

In practice, does the syntax impose a certain order in how it is used?
Does the syntax force you to think ahead and make certain decisions in advance?

Consistency

Are there places in the syntax where some things ought to be similar but the syntax makes them different?

Abstraction Management

Does the syntax make it possible to encapsulate details through new abstractions?

Visibility and Juxtaposability

How easy is it to compare different programs in the syntax and spot differences?

More re: Cognitive Dimensions