'Recat' or 'recon'? Pursue a labour-intensive, costly, and time-consuming method (recataloguing) or utilise a relatively cheap, largely automatic procedure, admittedly imperfect and fraught with errors, which holds up a glimmer of hope for a complete online catalogue in the near future (retrospective conversion)? This is the dilemma a big library is bound to face when it determines to migrate its traditional paper-based catalogue - or a part thereof - to the net. The article presents the main aspects to be considered in choosing the appropriate method: given the criteria of cost, quality, and time, many libraries prefer to adopt the cheaper and automatic method, even if it is less perfect in terms of quality: retroconversion. A spe- cific case of this is the scanning and OCR of the catalogue cards themselves, XML-conversion of the generated records as well as the creation of a number of indices, which are keys for users' online search sessions. A successful project of large-scale retroconversion is discussed in detail. The ADAM Project - named after ALEPH's module for managing digital objects - was run by the Central University Library of laşi (Romania). About 1.3 million traditional catalogue cards have been digitised and integrated into the OPAC of the library during the six months of the project for a relatively low cost. [ABSTRACT FROM AUTHOR]