Recent (Feb 2005) good weather and subsequent use of RxW highlighted the issue of pointing on extended sources with small beams.

The fivepoint algorithm calculates a useful step-size for an extended source based on its size, and provided each non-centre point of the five catches the limb the centroid determination ought to provide decent pointing. Even if the telescope is badly mis-pointed initially and one beam or two sit squarely on the disk of a large planet or squarely on the sky, some iteration ought to pull the pointing in. These algorithms need to be told that the source is extended, and while this is the case for the planets, there is no such forwarded information for other sources in the pointing catalog. However, the centroid should still provide good pointing if the source is symmetrical (eg irc+10216). Some planets at some phases will not present themselves symmetrically in the submillimetre, and this will affect both SCUBA POINTINGs and heterodyne FIVEPOINTs:

• The phases at which we view Jupiter & Saturn from Earth are never far from full-face (10deg & 6deg from full, resp. at maximum), but their diameters reach 60arcsec, so the impact upon pointing could be measureable.
• Mars shows phases of 0.85 (phase angle = 46deg), but when only 10" in diameter. With differing emissivity from sunlit and unlit surfaces this could cause a difference between the submm centroid and the geometric centre, but I suspect it's rather little (i.e. of order 1" or less).
• Venus might be much worse, as the 3mm image shown here indicates; Venus reaches an arcminute in diameter.
• And Saturn's rings also provide asymmetry, as seen at 20 or 17 microns. The impact upon the centroid then depends upon the viewing angle, which varies slowly through the years.

Coordinates for the extended sources in our pointing catalog are often derived from work at other wavelengths. For those with particularly complex morphology these may not represent the local peaks at any other wavelengths, such as those used to make the POINTING or FIVEPOINT observations at JCMT. OMC1 and others in the "Compact HII Regions" section of the pointing catalog are likely to fall into this category.

So : Extended and/or asymmetric sources may be used for pointing, but with caution.