Henrry Rodriguez, Kerstin Severinson Eklundh, (2005)
Conversation around documents: more than threading
Comments by ToddDavies
Document-centered discussion is a focus of our project (Deme), and I agree strongly with the idea expressed in your paper that discussion "around documents implies that the structure can be more complex."
A key to the ability of in-text commenting to elevate discussion may be that it can discourage paraphrasing and loose characterizations of what is being responded to. Paraphrasing is often a source of flame wars, as people misrepresent the words of others or accuse others of doing so. A comment inserted within text ought to be more likely to address a passage verbatim, provided that the original passage is seen by the comment author and by readers of the comment as they write or read the comment, respectively.
I can't tell from your paper exactly how comments attach to documents, and it would be helpful if you could post a link to your system so we can all look at it. It appears from your description that comments are simply attached to a whole document in chronological order. While document-centric discussion is often likely to be an improvement over unattached threads, and I understand the technical difficulty involved in annotating locations within an external web page, it seems to me that a big part of the potential for document-centered discussion is missing in this design, if I understand it correctly. I'm also unclear on the motivation for eliminating threading, which can, after all, be useful. You say it might lead to tangents, which is an interesting claim that I would like to see tested. I take the point of the workshop to be not that threading should be done away with, but that it should be supplemented with other online conversational techniques. Also, date sorting is a type of threading in my book, and can have similar effects.
I like the flexible email lists in Col-leccio as a way of choosing the most appropriate audience for comments, and I like your tools for visualizing discussion.
ToddDavies
In general I think there are some really good aspects to this work. For example, I think email notification is a very good idea (RSS feeds might be even better for solving the notification issue). And I very strongly agree that conversation around (and especially within) documents is can increase coherence and quality of conversation.
I have a real problem with the GFML though. All those checkboxes seem unnecessarily fine-grained. What are they for? It doesn't seem like you can arbitrarily add new names, so the only purpose seems to be keeping people out? What's the point of that? Talking "behind their backs"? If so, then a real concern is that people are NOT good at keeping private things private (cf Lynn's paper and discussion of locked posts). Also people will very quickly lose track of which subset of users is privy to which part of the conversation; it could get rather messy. I don't think the rationale of having it so people can uncheck themselves holds up, since it's easier to have a single checkbox that says "send to myself" or something similar. If you want finer-grained access control (which can make sense in some circumstances, I'm sure) then this could be a fun thing to talk about during the workshop.
ScottGolder
[Please insert additional comments in this section followed by your
WikiName]