Operation of a target-search strategy requires a target list. Even if we knew precisely what types of stars were accompanied by planetary systems believed likely to provide favorable life sites, we would not know where to look for the bulk of these stars in the neighborhood of the Sun, that is, within a range of 103 light years (ly). Present catalogs out to 103 Iy are probably incomplete for F, G, and K dwarf stars by a factor of 103. For early M dwarf stars, the situation is worse. As long as we favor planets orbiting stable stars as the probable sites for intelligent life, so long will we be in essential need of a Whole Sky Catalog, or stellar census of stars down to the 14th or 15th magnitudes. Even then the catalog would be seriously incomplete for M dwarf stars.
A suitable stellar census should have the following properties:
1. A high degree of completeness out to at least 1000 ly. Because planets orbiting early M dwarf stars (M0-M4) are generally not excluded as possible life sites, the census should extend down to at least the 15th apparent visual magnitude. As a consequence, 25 x 106 or more stars must be identified.
2. In the main, each star should have its MK classification established to a satisfactory degree. Where feasible, more refined classification is highly desirable. In particular, it is desirable to provide estimates of stellar age.
3. Stellar position and radial distances of each star should be determined to the greatest precision consistent with completing the first edition of the census in about a decade.
4. The census should be cross referenced to previously existing catalogs.
5. Where known, additional information such as parallax, proper motions, duplicity, variability, etc., should be included. In fact, the initial census should, to the degree possible, be considered both as a summation of current stellar information and as first epoch observations for a massive improvement in our knowledge of the stars in our neighborhood. In the process, of course, similar information on an enormous number of stars at much greater distances will also be obtained.
It is the contention here, though yet unsubstantiated by thorough study, that such a census can be achieved at modest cost in about a decade following funding. The approach we suggest has the following major ingredients:
1. Digitize about 400 standard MK spectra. If necessary, take additional spectra at suitable dispersion.
2. Carry out a multivariate (or multifactor) analysis seeking the optimum achievable color systems using precision photographic photometry. Assess the error contributions as well as the capabilities of the system.
 3. Design a completely automated photographic photometric system, telescope to tape data bank. The design should avoid manual operations other than transporting developed plates or, preferably, sheet film, in proper containers, from the observatory to the automatic plate reading machines. Automatic standardization and calibration procedures should be provided. Separate small telescopes should be used to obtain extinction information. Moderate field telescopes of the order of 60 in. should be adequate. Above all, the design of the telescope, its controls, the auxiliaries, the photometric apparatus, and the domes should be considered to be a single integrated system design problem.
4. Install identical systems at optimum sites, one in the Northern Hemisphere and one in the Southern Hemisphere.
5. Let all photographic data be measured by computer controlled machines. The same computer can correlate, classify, and store the data, and assess the results in real time. That is, completion of the catalog should about coincide with completion of the necessary observations.
Such a stellar census would have much value to general astronomy. Discussions with some photometric and spectroscopic specialists have strongly supported the belief that such a system is feasible. In any event, laboratory trials at an early stage of the design study will surely clarify the situation.
Should photographic methods prove unsatisfactory, it will be necessary to develop an appropriate photoelectric sensor system. This is clearly possible, and should be explored as a technology requiring a special development effort.