(1) THE APPARENT PRECISION OF THE DATA VERSUS ITS "SOFT" CHARACTER

* Informing the user about the soft nature of the data, before and during the dialogue (e.g., using relevant explanations).

* Using qualitative labels (which imply subjectivity) rather than numerical or quantitative ones (which imply accuracy).

(2) DEALING WITH WITHIN-OCCUPATIONAL VARIANCE

* Characterizing each occupation using a range of levels in each aspect instead of only one typical level.

(3) ELICITING ASPIRATIONS WHILE ENCOURAGING COMPROMISE

* Emphasizing to the user that these two stages are both necessary and complementary.

* Beginning by eliciting the optimal level, and then asking the user for additional acceptable levels on which s/he is willing to compromise.

(4) INCREASING OR DECREASING THE NUMBER OF ALTERNATIVES CONSIDERED

* The dialogue is flexible and tailors itself to the unique characteristics of the individual. That is, it decreases the number of alternatives if this number is relatively large, and expands it if it is too small.

(5) HOW SHOULD ABILITIES AND SKILLS BE TAKEN INTO ACCOUNT

* The system elicits preferences for utilizing abilities rather than eliciting self-estimates of abilities, on the assumption that people prefer to avoid areas in which their abilities are poor.

(6) TO RANK OR NOT TO RANK

* Presenting the user with a small set of alternatives without ranking them.

(7) SHOULD THERE BE A SINGLE LIST OR MULTIPLE LISTS OF "PROMISING" ALTERNATIVES ?

* Providing one main list which includes the set of alternatives identified as compatible with the user's preferences in those aspects he/she regards as most important.

* Suggesting another list which is based on a compensatory-model search, using all 43 aspects. In this second list, occupations which appeared in the first are highlighted.

(8) DEALING WITH AMBIGUITY, UNCERTAINTY AND FUZZINESS

* Guiding the user through the career decision making process by placing it in a systematic framework.

* Eliciting preferences using a range of acceptable levels in addition to the optimal one.

* Providing the relevant information (e.g., which occupations are not compatible with the user's preferences in each aspects).

(9) SOPHISTICATION VERSUS SIMPLICITY

* Translating the sophisticated notions into simplified but meaningful concepts that can be comprehended by the user (e.g., using the label "Almost suitable occupations" instead of "sensitivity analysis").

* Hiding sophistication "behind the screen" (e.g., weighting aspects by the individual's importance ratings in "Similar Occupations", without explicitly informing the user about such weighting).

(10) THE AMOUNT OF RELEVANT INFORMATION

* Facilitating the identification of a small set of "promising" alternatives on which the user can focus for further in-depth exploration.

* Providing only relevant information in order to avoid cognitive overload. For example, presenting the user with a short and concise printout (e.g., 6-8 pages) summarizing the dialogue.

(11) FLEXIBLE VERSUS CONSTRAINED DIALOGUE

* Beginning the interaction with a fixed order of modules and then allowing unconstrained exploration.

(12) PROVIDING TRULY INDIVIDUALIZED FEEDBACK

* Presenting the user with relevant feedback tailored to his/her responses, concerning, for example, important aspects, the need to compromise, the rate of elimination, and the desirability of using an additional search procedure to identify promising alternatives.