Comment Management Responsibility — a proposal

Just as Creative Commons cleverly emphasizes the rights of users (over what appear to be the overly restrictive rights of coyright), so should Comment Management Responsibility ("CommResp") emphasize the rights of the community members. O'Reilly pointed to the blogher Code of Conduct as an exemplary policy (note: I've been nominated as by the blogher co-founder as "bloghim"), but it focuses mostly on the prohibitions.

Here I clarify distinct parts of comment management:

  • Authentication: Credit-Card; Email-verified; Voice-Verified; None

  • Identity: Real Name; User-supplied username; Email address

  • Disclosure: to Anyone; to Members; to Publisher; to a Court

  • Screening: All posts; First posts; Hyperlinks beyond threshold

  • Maximum: Length; Hyperlinks

  • Prohibitions: Copyright violations; Libel; Defamatory; Abuse; Obscentiy; Privacy; Spam;

  • Redress: Disemvoweled; IP Address published; Purged ; Suspension of account

  • Closing: Never; Time elapsed; Discretion of publisher

  • Guarantees: Author's closing response; Subject's right of response

  • Editability: Always ; Time-limited ; Until Next Post ; Preview


Note that this doesn't make a value judgment on what the terms of service should dictate. It merely establishes the vocabulary for such terms. The standardization of these terms isn't just for publishers and readers; it's for makers of community management software to follow in order to enable the publishers to clearly assert their rules.


First, note that authentication is separate from identity (following the philosophy of the identity gang). That is, a person can write under a pseudonym while their true identity is of limited exposure. A publisher can promise never to look up an the authenticated identity unless compelled to by a court– and even then, they may claim journalistic shield protection.


Screening is often referred to as "moderation," which is confusing, since moderation often is executed ex post facto. Thus we declare screening to mean that posts are manually checked before being posted. One of the regular annoyances is that a number of none-registration blogs do not declare up front whether they use a hyperlink threshold for flagging posts for screening. Comment policies should make this clear.

It may be handy for a commenting system to set a maximum number of letters or words in a post; these can easily be enforced by the software code.


The act of purging an offending post is a seemingly obvious idea, but there are different ways of doing it. Should the post deleted, or merely blanked out? "Disemvowelment" is a clever moderation strategy that was promoted by BoingBoing several years ago. A moderator takes an offending post and removes the vowels from it so it becomes unintelligible to the casual reader. People who had read the original comment would recognize where it was in the thread. It's akin to the public stocks of colonial times, or primetime cop dramas of today– showing the guilty punished does its part to keep society on the straight-and-narrow.


These categories are most sympathetic to the rights of users making the comment posts.

The Closing category is simple: Are comments kept open, or are they closed after a period of time– or just at the discretion of the publisher?

The Editability category is so basic I forgot it in the first version of this document. Is a poster granted the right to edit a comment that they've posted? It can be always, or, more reasonably, it may be time limited (for, say, a few minutes), or at least until the next person posts. Perhaps the publisher just provides a preview functionality in lieu of ex post facto editing.

The Guarantees are more delicate. Before the publisher closes a comment thread on an article/post, is the author of that article expected to write a brief response to the responses? In What Lies in Conversations (February 2005), I likened the failure to do this (and starting the next post) as a betrayal of the blogs-as-conversations mantra.

Another guarantee is equal time– necessary in case a subject of an online post has been wronged and wishes to respond. The Delaware Supreme Court held in the 2005 Cahill case that "The internet provides a means of communication where a person wronged by statements of an anonymous poster can respond instantly, can respond to the allegedly defamatory statements on the same site or blog, and thus, can, almost contemporaneously, respond to the same audience that initially read the allegedly defamatory statements."

This point has been repeated by Susan Crawford, a professor at Cardozo Law School, and by Jeff Jarvis, director of the New Media Program at the City University of New York. The problem is that it assumes that the online publisher has established the code for this. Many newspapers, and some blogs, do not allow comments; other blogs and forums attract too many comments that the subject's response may be buried within them.


The present Posting Guidelines on this website could be coded simply E/U/P; M: 2000 chars; AO=>P; C:N (Email-verified, user-supplied username, known to publisher; maximum post length is 2000 characters; abusive/obscene posts will be purged; comments are never closed). By practice I tend to respond to all comments. I did give a "air time" for a subject's self-defense, but I made a mistake in not presenting his letter in full, and instead interspersed my reactions to his response (you decide).

Once a taxonomy can be agreed upon, the XML and representational icons would follow.