At Future of the Book we recently hacked WordPress to allow comments to tie to specific paragraphs in the text and I think it turned out pretty well. Other people seem to think so [1][2][3]. In just a few sleepless nights we had fully functional comment system that worked well enough to launch Mitchell Stephens’ new project called “The Holy of Holies: On the Constituents of Emptiness”.
We are planning on releasing this tool as a WordPress theme, and that suggestion has been getting some attention, but we need to do some serious clean up so to make sure that things do not break with complex text formating (as it does now).
One of the more interesting problems we’ve spent lots of time discussing at the institute is what we like to call “holy grail of commenting.” This system will, theoretically, of course, allow text and discussions surrounding it to flow seamlessly in a common interface, while not taking away from the vanilla reading experience. But as it turns out, this is quite a complex issue.
This is due to technological constraints but also mental ones. In order to make reader contributions line up with the original text, you have to start breaking things up into boxes and columns and tabs — all legacies in one way or another of the printed page. In the electronic environment, we can do cool things like have boxes and columns overlap, or hide areas of text behind other areas, or make windows within windows that scroll up and down, but we’re still thinking in two dimensions. How many two-dimensional spaces can you pile up on one page before the whole thing collapses? How far can we stretch Word Press — how much hacking can it endure — until we rip open a worm hole that takes us to a reading/writing space where different rules apply?
One of the challenges with a system allows anyone to comment on anything is that the distinction between different types of comments is difficult to show, and often that is very important. But having arbitrary comment distinctions makes the user interface a lot more complex and inflexible.
For example, some comments, like author footnotes, don’t always lead to a threaded discussion, but such comments currently have the same weight as comments that ask questions. Also, comments that discuss specific words or phrases are at the level of someone who comments about the entire paragraph. So suddenly we are thinking about not only weaving together comments into text, but then also weaving the comments into themselves, and allowing a discussion to run parallel to text.
We’ve got some thinking to do.
We’re going to continue to develop this format and will be using it, or variations of it, for a number of projects in the near future. We’re also working on something that allows highly flexible line by line, even word-level, commentary.
This has actually turned out to be a tricky user interface problem. One of the current implementations of such a system is the GPLv3 comment system, which we love and works pretty well for finding active areas of discussion, but does not give much weight the ongoing discussion. The main text must be read in order to access comments. So the entry point into the discussion is limited to the text; the discussion is not allowed to develop on its own. Since there is no entry point from comments into text we’re looking at ways to solve this issue.
One of the ideas that I’ve been curious about, and hope to experiment with, is a text-zooming engine that will allow a user to read the text very “closely” and comment on specific text fragments or words, or for a user to read the text at a distance, and comment on higher level structures, such as sentences, paragraphs or entire pieces. Small hints would appear in the text to tell users that they might want to read more closely or father to see a new level of discussion and the comments would appear directly next to the text as the user zooms in and out. I think this would solve the problem of distinguishing comment types and discussions, but can potentially confuse the user and hide some discussions. We’ll see how this develops.
Mitch’s project was a good step forward from Gamer Theory, and now that we’ve taken this baby step forward we are getting into hairier issues.
sounds very cool! i was thinking about this post and i realized there’s another, related widget that might be helpful too: a way to hide/unhide detail in a post. so for example, i could write a post which says this:
When you click on the first “unhide”, you see more details, for example:
.. and you can imagine similar expansions for the other “unhide” links. Clearly, clicking “hide” above will then make the detail on the bookstore visit disappear.
Further, you could imagine nested hides; in the above example, I could include more detail about Freakonomics and what I thought of the book, etc.
It’s easy to see how this could be very useful. There are many blog posts i’ve seen where the main thrust is obscured by the copious evidence. You want to see the evidence, of course, but it’s easier to read the general idea first, and then go back for the evidence. Moreover, I frequently don’t have time for much more than a five sentence synopsis.
So, is something like that in the works? Or is it already here? Seems like a useful idea.
Am a bit late to this but it’s very very cool – I’d love to have something like this to offer edublog users and if I can be of any assistance in playing / testing then just let me know.
And incidentally, if I could advise anything I’d always suggest staying as close to WP as possible – makes for much greater take-up and the ability to leverage other important commenting tools like subscribe to comments / spam blockers and the evolving gravatar stuff… but you already knew that
Great work! Cheers, James