Following the diversified topic of this conference, let me present a variety of comments — not all new, but resulting from a long string of stars drifting across the desk. The chase after visual orbital elements is not exactly a self-purpose but is aiming at further data, in particular, at good masses. The last published lists of high-quality mass determination represent the status of 20 years ago; but a compilation at this time would probably not last long as the progress promises to quicken.

We have 1000 positional (visual/photographic/speckle) orbits, among them about 700 acceptable in the range from fair to definitive. Yet less than 10% of them give good component masses. Most of them are outside the range of parallax measures with the requisite, high precision (that unfortunatly holds for the Hyades); some frustrate the parallax measurer by displaying wedge– and peanut–shaped images, and the more exciting cases of abnormal, non-main–sequence components often cannot get good mass ratios owing to large distances or long periods.