I’ve talked about personal comment aggregation before, and I want to discuss what I see as the problem(s) to be solved and what solutions I see. Please feel free to criticize my idea—it’s still nascent in my head.
Here are the problems as I see them:
A Personal Record of What, Where, and When By Whom
The who, from the problem’s perspective, is me or anyone like me who wants to have a record of what they’ve said, where, and why. Over on GFMorris.net, you’ll see a sidebar that links to places I inhabit on the Internet. That’s partially so you can see what I’ve said, where I’ve said it, and when I’ve said it.
What I’ve Said
I see knowing what I’ve said as an extension of the whole public face thing. I participate in discussions in a variety of venues, and if I’m going to be public about it, I want people to know what I’ve said. However, more importantly, I want to know what I said, because I’m forgetful. Knowing the “what” is extremely important to me.
Given that hypertext is largely a textual medium, the what is often going to be words. The what could also be images I submit [for instance, photos I post on Flickr], or it could be the one podcast where I am a part of the action [although I don’t have good links to that … I’ll get around to that at some point :)]. I’m leaving a breadcrumb trail, but it’s not just for me—it’s for anyone who’d want to know what I said.
Where I’ve Said It
I think that most of us adapt our message to the medium. For example, you will find that my posting style is different across my Weblogs to a degree, and you will also note a different voice on, say, The Rumor Forum than what you get on the USCHO Forum. For one, I spend a lot more time on The Rumor Forum; for another, we’re talking about wholly different things. Place is a part of the context.
When I’ve Said It
A foolish consistency is the hobgoblin of little minds, adored by little statemen and philosophers and divines. With consistency a great sould has simply nothing to do.
–Ralph Waldo Emerson, Essays: First Series. Self-Reliance
I’ve often said to those who’ve asked me in the past the writing and journaling personally online is writing one’s own narrative. As time passes, our views on things certainly can change—oh, sure, some things are going to remain the same, but many things will change by small or large degrees. If one read, say, geopolitical comments that I espoused on alt.books.tom-clancy in the fall of 1997—my freshman year at UAH; I was 19, full of piss and vinegar, quick-witted, fast of keyboard, and even faster to stick my foot in my mouth—you’d find my leanings a bit different then than they are today. I’m sure that this is true with things I’ve posted in Weblogs as well—I can name a handful off the top of my head, and could probably go bugnuts if I kept it going.
But in any regard, time metadata is important, for all the reasons that Web wonks have argued over the years—old entries likely have outdated discussions, etc. I like to think that words set farther back in stone are likely to be more weathered by the elements and less distinct—and sometimes, my delivery is vague and circumspect, and hell … I can’t remember who it was I was talking about but trying not to name.
A Record of What Others Said, Where They Said It, and When
Let me begin with an example: this entire entry is written in response to a comment made by Chris Meller. Mr. Meller writes:
So what features would your ideal system include? I mean, most blogging software out there today includes a seperate RSS feed for comments (WP includes a seperate one for each postâ€™s comments, Iâ€™m not sure about others). I have a tag setup in FeedLounge that is simply â€œmy commentsâ€, where I subscribe to all the feeds for comments on posts Iâ€™ve commented on.
Iâ€™d be open to the possibility of starting a new project in the near future to satisfy this need, but I really donâ€™t have a good grasp on what youâ€™re looking forâ€¦
Now, I could have written the Comment to End All Comments here, but that would have been painful to read, and frankly, difficult—comments are largely noise when compared to the entry [signal] that initiated them. That’s simply the general rule with comments, and on that score, it’s akin to James Madison taking notes at the Constitutional Convention and presenting the Virginia Compromise—after that point, discussion stopped being along the lines of “How do we fix the Articles of Confederation?” and more towards “What about Mr. Madison’s idea?”
Out of such sentiments was likely born the (fraternal) twin terrors of the ping, Trackback and Pingback. Both mechanisms sought to make a daisy chain of links, going from Entry A (Call) to Entry B (Response) through Comment C (TB/PB). Unfortunately, because this was designed to be as simple as possible, it proved to be easy to game by spammers, and many folks have given up on the mechanisms. [Not I! I Trackback myself regularly when I reference things I’ve said in the past. I believe in the medium!]
What’s great about these links-by-pings is that it turns what normally would be a one-way conversation—Entry B talking about Entry A—into a two-way conversation, because the very conversation-space of Entry A suddenly has a link to Entry B. Automating the process simply increased the likelihood of adoption by content producers. [Unfortunately, neither mechanism seemed to ever get enough traction with everyday users to advance beyond alpha geeks and early adopters into the mainstream. Alas.]
But before I lose your interest, reader, I hope you see my point—responses to content in one medium may not necessarily occur in that same medium. They could well be referenced in another medium—and that, folks, is the beauty of linking.
[Want an example? Okay: When I discussed homesteading our noÃ³sphere, I linked to a comment I’d left on an entry by Scott Sanders, which itself linked to a topic I’d started on The Rumor Forum about our membership drive. That’s an entry linking to a comment linking to a topic. Behold the power of links!]
The reason for wanting a response record is pretty simple: these mechanisms don’t go two ways very easily, and often, links have to be scraped. But the great thing is that links can be scraped, given enough time.
What Others Said in Response
The chief concern is, of course, what other folks said in response to something I had to say—“Whatchu talkin’ ’bout, Willis?” I may need to clarify my remarks, or someone may change my mind again. [Hello, Mr. Emerson!] Most of us who publish on the Web want feedback about what we say—that’s why you see all the comment forms out there, even if the spammers do try to mess with them.
Who Made the Response?
Oh, that bugaboo of identity again. I’ve bought into the link-as-identity meme—astute readers will note today that I leave GFMorris.net as the URL for comments that I leave around the Web almost exclusively. The idea is that, because I’ve made it a central node for pointers to things I’ve said, I can leverage that to claim the things that I’ve said. Note that this URL may be independent of the medium of the where the response is made—in fact, for me, it’s always going to be different, because I don’t produce content on GFMorris.net, I just link to it from there.
No matter the nature of how identity is determined—trust me, if you go read all the tracks I’ve left on my essay on user registration, you’ll see that I’ve had lots of thoughts about what identity means in the abstract and the specific. Identity can be established in many ways, and I argue that identity is doggone important.
Where Others Made Their Response
As I noted in my example above, one may switch media or places when responding. [Again, I’m responding to a comment Chris made on my Weblog here on my codelog—I’ve fundamentally changed place.] If Chris was tracking all references to the permalink to that comment, he could see this entry—regardless of whether or not I Trackback/Pingback the entry where Chris left the comment.
Also, people may simply reply in the comment space. Of course, as I noted in the original entry:
No one really seems interested in providing parent-child relationships between commentsâ€”oh, wait, Dunstan Orchard was interested in it, but heâ€™s quit Weblogging, and I canâ€™t even give a proper deep link to his intercomment links, because his CMS is throwing up MySQL errors by the dozens :sigh:â€”or, at least they havenâ€™t shown much interest.
Dunstan’s method was elegant, because it understood that there were often two or more parents per child comment. Linear threading simply isn’t how we humans think.
Again, where is important, because the breadcrumb trail needs to work in both directions.
When the Response Was Made
Follow any discussions about hot Web topics—Web 2.0, the hubbub about Atom a couple years ago, whether or not Trackback is dead, etc.—and you’ll see that opinions change. Just as I’ve noted that I change my mind about things, I’m always quite sure that others understand that they have the freedom to do the same and will exercise it from time to time. Time metadata, therefore, is quite appropriate in this context.
So, A Solution?
I thought that I’d have a solution by this point, but honestly, I don’t. At this point, my barebones solution of using del.icio.us to leave myself a breadcrumb trail has given me a very first approximation result to the first problem, although it’s very metadata weak: all it contains now is a link to a comment I’ve made, which really only gives me the personal where and the personal-who metadata I’m seeking—I’m relying on memory to know what I said [or simply clicking the link to refresh my memory], and the when may well be delayed if I have a user who doesn’t provide comment permalinks in their entry files and I have to do some sleuthing to find such matters [or, perhaps, if my comment’s held for moderation]. I reckon that I could take the time and copy and paste my comment into the notes section, and drop a timestamp in there, but honestly … that’s a hell of a lot of work for one link, and it really doesn’t scale very well.
I’m certainly open to suggestions. I think the personal stuff can be solved by providing permalinks [where available], because scraping can be done, but one runs into the problem of different comment presentation schema. Without a standard, you’ve got to build a lot of logic into the system to have the computer do the work for you. Again, maybe this is where a more robust standard for publishing and syndication like Atom will make this more of a reality.
But as it is, I’ve got no solution for the second part of the two-way conversation, the reponses to the comments made. I’m sure that the answer is search-and-scrape, but that takes CPU cycles as well.
However, because the problem exists without a solution at present doesn’t mean that I can’t take an hour on a Monday night to flesh out the problem as I see it, eh?