Radiometry, i.e. measuring the power of electromagnetic radiationhitherto often referred to as "photometry"-is of fundamental importance in astronomy. We provide an overview of how to achieve a valid laboratory calibration of space telescopes and discuss ways to reliably extend this calibration to the spectroscopic telescope's performance in space. A lot of effort has been, and still is going into radiometric "calibration" of telescopes once they are in space; these methods use celestial primary and transfer standards and are based in part on stellar models. The history of the calibration of the Hubble Space Telescope serves as a platform to review these methods. However, we insist that a true calibration of spectroscopic space telescopes must directly be based on and traceable to laboratory standards, and thus be independent of the observations. This has recently become a well-supported aim, following the discovery of the acceleration of the cosmic expansion by use of type-Ia supernovae, and has led to plans for launching calibration rockets for the visible and infrared spectral range. This is timely, too, because an adequate exploitation of data from present space missions, such as Gaia, and from many current astronomical projects like Euclid and WFIRST demands higher radiometric accuracy than is generally available today. A survey of the calibration of instruments observing from the X-ray to the infrared spectral domains that include instrument-or mission-specific estimates of radiometric accuracies rounds off this review.