What is "scholarly publishing"?

In the comments to an earlier post, we have been having a conversation about what constitutes "scholarly publishing", especially with respect to publications on web-sites etc. (see the comments to this post). The issues, especially in relation to web-based publication, are clearly very important. I found the following characteristics helpful (from an article by Leah Halliday, 'Scholarly communication, scholarly publication and the status of emerging formats' in Information Research 6. 4 (July 2001); online at: http://informationr.net/ir/6-4/paper111.html).

A scholarly publication requires (essential (E); Highly desirable (HD); and preferable (P)):

Trustworthiness

• Publications should not be changed (HD).
• Different versions should be clearly identified (HD).
• To satisfy all potential interest, trustworthiness should be based on 'institutionalised' measures such as peer review rather than on personal knowledge (HD).
• Each publication should have at least one identifiable author (P).

Publicity

• The potential audience must be made aware that the publication exists (HD).
• The publication should have metadata containing a minimum set of information, preferably including information about all versions (P).

Accessibility: the document must be readily obtained by those who wish to use it.

• The author must intend that the publication be made publicly available in a durable form over the long term (E).
• The publication must be durably recorded on some medium (E).
• The publication must be reliably accessible and retrievable over time (E)
• There should be a commitment not to withdraw the publication (E).
• The publication must be publicly available, i.e. available to any member of the public on demand as of right, whether for payment of a fee or not. (E).
• The publication should have stable identifiers (HD).