I'd like propose a visual aide for using the Translate tool.

Please add a "reticle/marker" for indicating the proposed displacement position/location of a translate of a selected point or centroid of the points selected.

Example: a user has in one view curves going vertically or radially. In another view, the user wishes to correct waviness in the heights of the endpoints. Tangent exists, but sometimes, it gets confusing or snaps around. It could be more straightforward for the user who is translating along an axis or path to:

-- select the point or points to be moved
-- positions the cursor start point and clicks
-- moves the cursor along the path as a precision placement path
-- watches the reticle for indication of tangence prior to releasing the mouse.

As an added enhancement, the user should be allowed to hit any one random key on the keyboard midstream withouth breaking the translate, and by holding any one random key, lock the translate motion to prevent trembling writs or twitching fingers from releaseing the translate to commit. This would save the user time when a surface is being reshaped, since some surfaces can cause an additional 45 seconds before SLT becomes ready for the next command.

A further enhancement would be to allow the user to select endpoints that form the arc to which endpoints are being translated/displaced. Then, the user can tell SLT to explicitly use the resulting arc as the termination, or it can smoothen out the virtual arc, then accept instruction to displace the endpoints to an averaged or weighted spacing proportional to the existing, associated curves (to avoid overbending any lines that look straight in one view but which sort of skew in another view, and when that shape is meant to be tweaked, not wholesale overdone), along the virtual arc.

This could save the user anywhere from 5 to 25 enpoints worth of translation. At some 40 seconds per translate, give or take additional time for erroneous translates, 13 minutes of time could be reduced to 2 minutes, assuming all the same 40-seconds processing per move/translate workes out to 2 minutes when more points are moved together adn relevant internal smoothing and work occurs.

This ultimately could be a step toward helping users smoothen an underlying surface without enduring any conversions that add unwanted control points to interpolared splines or turning them into other line types that might become splines of dozens or hundreds of control points.