Another problem that was making it difficult to automatically merge GCVS and survey catalogs has been recently solved: 'old' variable stars also had only very rough coordinates. The GCVS has been identified, on a star-by-star basis, with positional catalogs. Several thousand unpublished finding charts (L. Plaut, The Netherlands; M. Hazen, USA; and others) were used. Errors in coordinates up to ~10 degrees (!) were revealed and corrected. The total number of stars we were unable to identify, at least tentatively, is only 200, most of them from two or three sources of especially unreliable positional information. 38 GCVS stars had GCVS names with wrong constellations; they have been renamed in the recent Name List No.78 (just published in the IBVS).