Sir

John Ewing argues in Correspondence (Nature 425, 559; 200310.1038/425559a) that, in the current reader-pays publishing system, researchers without a journal subscription can obtain an article by other means. This is probably true. However, article-sharing is an illegal and unreliable method of getting scientific information, because many commercial publishers own copyright as well as the rights to distribute the results. The magnitude of this copying activity is, as far as I know, unknown.

A few requests to “send me a PDF file of an article” will probably not hurt the current system. However, if this approach were to become more systematic, a PDF file (or other formats) could be 'shared' using databases and peer-to-peer networks just as any MP3 music file can be shared worldwide, to the great annoyance of the music industry. This would definitely hurt the current subscriber-pays system.

The open-access alternative is immune to such copying activities because the articles are available free of charge.