We are reluctant to remove rapid responses from bmj.com because they are part of the BMJ’s record—and hence science’s record. We believe that there are few good grounds for tampering with it.

To date, over 88 500 rapid responses have been posted on our website since 1998, each with its own URL and each retrievable in an advanced search of the web site. We select the letters we publish exclusively from them.

To anyone who wants to send a rapid response, we say consider your rapid response to be an electronic letter to the editor. And what is a letter to the editor but part of the process of peer review dating back to before the birth of the BMJ, at least to the inception of the Royal Society? Thus rapid responses are integral to the scientific and historical record.

We do not delete rapid responses once we have posted them, just as we did not and do not cut out pages people are not happy with when they see them in print. This is not the way of publishing. Very early on in the electronic world we realised that rapid responses were published entities, living in the public domain as soon as one of us had pressed the button. We imagined that someone somewhere might have seen the response we had posted as soon as it became live so what would it look like if we started messing about with it or, worse, deleted it after publication? People might reasonably doubt our public integrity and professionalism.

We screen the content of rapid responses carefully before posting on our web site. We are liable for libel, and we will not post obscene rapid responses and those that might incite racial violence. Ad hominem attacks are never edifying so we try to avoid them while at the same time allowing some cut and thrust (colour). We do not edit to a style of declamation so that rapid responses remain immediate and idiosyncratic, the author’s voice loud and clear. At their best they are conversations showing how an idea is developed and knowledge advanced in a forum.

We adhere strictly to our policy of not deleting rapid responses once we have posted them. The only exception is when we’re told to do so by lawyers. On the very rare occasions when this has happened we’ve left an audit trail.

The commonest reason people ask us to remove their responses is not because they have changed their mind but because searches conducted by potential employers have returned responses with their unsavoury or embarrassing views. To this we reply, you were warned: the note on the rapid response submission page clearly states, “Once published, you will not have the right to remove or edit your response.”

Sharon Davies is the rapid responses editor, BMJ.