Mathematically oriented texts are produced mainly by using TEX processing system. In the recent years, PDF has become de facto the standard for archiving, document presentation and exchange on Internet. The conversion from TEX to PDF format is the critical and most important point in obtaining scientific papers in PDF formats. We discuss in this paper some drawbacks of PDF files obtained in this way, particularly concerning indexing, copy/paste and OCR capabilities. We also discuss some shortcomings in uses of PDF files with e-book readers.