From: Erik Sandewall 
To: aiello@dis.uniroma1.it
Date: Tue, 28 Jan 1997 18:17:35 +0100

Dear Gigina,

I appreciate your comments, but wrt the first point I have the opposite opinion: I emphatically think that it should *not* be possible for an author to withdraw a work that he or she has published. This is for two reasons. First, a philosophical reason: I think that if something is published, it shall be referencable, and to be referancable is only meaningful if it remains available.

Secondly, and more pragmatically, the non-possibility of withdrawing a work has an important role for reducing the amount of non-well-thought-out papers. Consider: one of the important requirements on a scientist is to be able to show good judgement with respect to research work, bot other people's and one's own. Now, if you are in an academic promotion committe and you are faced with an applicant who has published (in my sense) a considerable number of articles, and very few of them got through the subsequent reviewing, then you will rightfully question the competence of this guy.

Therefore, researchers (and for graduate students, their advisors, for an analogous reason) will think twice before publishing things, and this is in their own interest.

The *present* publication system (review prior to publication) does not have this property, because the *reviewers* get to take the responsibility. As an author, you are fairly free to send off a lot of publication attempts: if they are accepted, they must have been good, and if they are rejected, then noone will know. It should be no surprise that reviewers get to be bombarded with subquality papers. The system I propose will *force* authors to take more responsibility. But this assumes that publications can not be withdrawn.

Re the second possibility, that authors will prefer to go to traditional journals after having put the paper in the FPA: yes, I don't think that would be a problem if it happens. Note, however, that some of the present journals may frown on such submissions on the grounds that the article has already been posted on the net using the magic phrase "published". It remains to see what position traditional journals will take on that issue.

Best regards,

Erik