Upon the long history of dislocation theory and dislocation imaging using transmission electron microscopy, improvement in the understanding of dislocations requires continuous experimental substantiation and calibration. In the meantime, emerging alloy design strategies are increasing the complexity of the materials, thereby challenging classical descriptions of dislocation behaviors. It is critical to integrate new information from experiment into modeling, which hinges on extracting quantitative and transferrable information of dislocations by developing microscopy techniques, including computational approaches. This talk presents recent advancement in 3D and in situ dislocation characterizations, facilitated by the automation and multi-modal capabilities in scanning electron microscopy. To demonstrate these approaches, a study of glissile dislocation junctions in deformed FCC Al is highlighted. The structure is determined by full characterization of Burgers vector, line direction and plane of all the dislocations comprising the junction. The experimentally determined structure is directly transferred to discrete dislocation dynamics (DDD) simulations to evaluate its evolution under stress. The analysis is further generalized over an ensemble of dislocations, revealing the importance of cross-slip on the formation of glissile junctions and the mobility of the resulting dislocation nodes.