Choose Boolean Attribute Text
The following general principles apply to the writing of Policy Modeling Boolean attributes.

A Policy Modeling Boolean attribute must include at a minimum a subject and verb. The subject is what or who the sentence is about. The verb tells us something about the subject. Most sentences also contain an object which is the thing the action is being performed on.
Examples of grammatical sentences are:
the payment was received (subject – verb)
the payment was not received (subject – verb - negative form)
the customer received a loyalty discount (subject – verb – object)
Note: There is an exception to this rule for Boolean attributes in Japanese and Korean for which Policy Modeling can handle verbless sentences.

The tense of a verb is used to indicate when the action took place.
Your goal attribute should usually be worded in the present tense as it describes the current state of affairs at the time the user interacts with the policy model. In general, every attribute below the goal should also be written in the present tense (or the present perfect tense), because the relevant issues are generally also being assessed as at the time the user interacts with the policy model.
For example:
the person is eligible for an award (present tense) if
the person has demonstrated exceptional conduct (present perfect tense)
the person has demonstrated exceptional conduct (present perfect tense) if
the person has been commended by peers (present perfect tense)
This principle applies regardless of the tense of the source material.
However one exception to this is where your policy model is being used to determine the current state of affairs based on past conduct or circumstances which are already completed at the time the user interacts with the policy model. In situations like this it makes sense to phrase the relevant attribute in the past tense. For example:
the person meets the three day requirement for victim compensation (present tense) if
the person reported the incident to police within three days of the incident occurring (past tense)

In English grammar we make a distinction between the speaker/s (I, we), the addressee (you), and the one/thing spoken about (he, she, it, they). This is known as person: first, second and third person, respectively. Boolean attributes should be written in the third person. (Note that there is a mechanism in Policy Modeling for switching attribute forms to second person for use in interviews.)
For example:
the person can go to the movies
the person has done a good job
Rather than:
I can go to the movies
you have done a good job

Some Boolean attributes can be difficult to negate and for this reason should be avoided.
Examples are attributes which use the conjunctions 'and' and 'or'. In these attributes ambiguity can result from the negation of the attribute as we do not necessarily know how the negation of the verb should affect each of the components. For instance, look at the attribute "the cat and the dog ate the man's dinner".
If this attribute is false, this could mean that:
- neither the cat or the dog ate the man's dinner, or
- the cat ate the man's dinner but the dog did not , or
- the dog ate the man's dinner but the cat did not
Given that there are three possible interpretations means that this attribute cannot be negated conclusively and should not be used.

In many instances, it may be tempting to word an attribute that could be split into two separate clauses as a single attribute.
However, if it is likely that part of the attribute is going to be used in other attributes, it is best to separate it into two attributes which each represent distinct concepts.
Note: This separation is essential when writing rules in Japanese and Korean.

In English, contractions are used in more informal styles of writing and speech and should not be used in Policy Modeling attributes.
For example, rather than "there's an application pending", you should write "there is an application pending".
It is particularly important to avoid the negative contraction "isn't" in attribute text, because Policy Modeling will not recognize an attribute using "isn't" as the negative form of the corresponding statement. For example, "the employee isn't entitled to travel allowance" would not be recognized as the negative form of "the employee is entitled to travel allowance", and the rules using these conditions would not be automatically linked. You should instead use "the employee is not entitled to travel allowance" if you need to use the negative form of the attribute in a rule.

Each Boolean attribute should be meaningful without reference to another. To do otherwise makes the policy model more difficult to develop, maintain and audit.
The following are examples of attributes which do not make sense in isolation:
- This section has been satisfied
- That discussion was recorded
- The person qualifies for the reasons above
- The latest of these two dates applies

The wording of the attribute should be as simple as possible while still retaining its full intended meaning.

If a Boolean attribute refers to a non-Boolean attribute for which substitution is desired, it needs to contain the exact text of the substitutable attribute, including the definite article (if the language uses articles). (See Personalize Attribute Text for information about this feature).
For example, if you have the non-Boolean attributes 'the student' and 'the educational institution', then a Boolean attribute for which substitution of the values of 'the student' and 'the educational institution' is desired needs to contain those exact strings of text. That is, it must contain 'the student' and the 'the educational institution' rather than simply 'student' and 'educational institution':
the student is enrolled at the educational institution
rather than:
student is enrolled at educational institution
Where 'the student' is Emilia, the educational institution is 'the University of Lisbon', and substitution has been enabled for both attributes in Policy Modeling (see Enabling substitution in Policy Modeling), the Boolean attribute 'the student is enrolled at the educational institution' will appear on screens, explanations and forms as:
Emilia is enrolled at the University of Lisbon
Note: There are language-specific considerations for choosing attribute text to facilitate substitution. For more information, see Special considerations affect substitution in some relevant supported languages.

If the attribute belongs to an entity, the exact text of the entity should be included in the attribute text to make it clear which entity it belongs to. For example, if you have an entity 'the child', then attributes which belong to that entity group should include the text "the child":
the child is happy
the child’s toy is educational
the birthdate of the child

A non-Boolean attribute can be replaced with the appropriate pronoun the second (and any subsequent times) the non-Boolean attribute is used in a Boolean attribute. For example, if we had a non-Boolean attribute 'the claimant' we could write a Boolean attribute 'the claimant owns the claimant's home' and then once we know the name and gender of the claimant this would be rendered as 'John owns his home'. This is preferable to hard-coding "his/her" or "their" in the attribute text.

Boolean attributes which refer to amounts should specify the unit of measurement to avoid any ambiguity. For example:
the person was 100 feet from the scene of the crime

When creating an attribute while writing a rule, if you want to have attribute text that includes "-" or "/" you need to make sure not to put spaces either side of the symbol. Spaces around these particular symbols denotes that the symbol is being used as a mathematical operator (for subtraction or division) and consequently two attributes will be created instead of one.
Note: There are language-specific considerations for choosing Boolean attribute text in Bahasa Indonesia, Bahasa Malaysia, Japanese and in Korean.