Building typologies is a good way for examining and discussing relevant indicators of network qualities. It allows to analyze combinations of indicators rather than indicators used separately, thus to compare networks on multiple dimensions rather than reducing them to a single quality. It also allows a generalization grounded on empirical data.

Typologies are involved in a tension between complexity and simplicity. They must be sufficiently complex and account for enough characteristics for the social scientists not to stray too far from the social reality, which is itself complex; but, ideally, they should also be simple enough to be applicable to larger numbers of personal networks stemming from various surveys.

Typologies may concern size, composition and structure of networks, but they often imply also sociological (or economical, historical, geographical or other) criteria for testing relevance and correlations.

Having stable typologies of personal networks based on structural criteria (among others) becomes more and more interesting given the increasing amount of available data generated by online activities. It may allow to process automatical classifications .

Different surveys and datasets, with different scales and from different disciplines may thus be involved in comparisons, cross-analysis and discussions about typologies.