To meet the requirement of the common availability of the documents, after the fulfilment of the retrospective conversion of the catalogues in the various national and county libraries, now it is the time of this work in the smaller special collections as well. This paper'd like to compare two recent bigger projects of this work, both of them was an authomatic solution for retro conversion, but there were chosen two completely different ways. This two databases were migrated.

-Twenty thousand catalogue-card in the Library of Military History Budapest

-12 000 cards of the Libary of the Hungarian Dance Academy Budapest

Even the method of the process was completely different. In the first case, each electronic catalogue data was made from the data printed on the paper based catalog card itself.  In the oher case, the records were downloaded  from a shared catalogue system, retrieved by certain  data elements detected in the record of the scanned text of the catalogue-card, and these items were completed with the local data.   In the first case the segmentation of the data and data groups were made after the migartion of the data strings,  by the punctuation of the text, already in a Huntéka databasis. In the other case the scanning and text analyzing of the cards was done only to get some identifying units for the retrieval of the bibliographic records. The scanned and by punctuation segmented records have needed more operative work in the second part of the projekt, than the one built of downloaded records, especially becaues of the big amount of non identifiable authority records.

Why the two kinds of solutions were chosen?

What different problems were occured?

What were the costs of the two projects?

These questions are going to be answered in this paper.