Alex wrote me an email while I was off at hockey—yep, another defeat of those pesky Beavers; thanks for the four points, bitches!—to let me know about Michael Arrington’s TechCrunch entry on CoComment, which seeks to help uses with t…

Using one centralized server seems like it’s still rife with issues; you’re probably better off with a decentralized network. Of course, this is all largely pie-in-the-sky stuff.

It started by doing some baseline collection of data (exactly the same way as a search engine but looking only for comments or other commentlike activity online – i.e. discussion forums perhaps). You might actually be able to reuse Alexa/Amazon’s index though I don’t know how well they capture comments.

THEN instead of (or at least partially instead of) continually polling for the data, what if there was a CLIENT based tool which UPLOADED to the central server pages with comments (with a probable restriction that it only does this for pages that are public).

The idea being to distribute and reuse the local cache of pages in place of the search engine like portion having to seperately poll sites for data.

There would likely still be some privacy concerns to watch out for (i.e. many sites might use a local cookie to display an indication of private messages or to reorder items by what you have/have not read etc) but that can probably be discarded by the index.

As well, it would probably be the case that certain frequently trafficed sites (Slashdot, Fark etc) would get submitted many times though each submission might contain some quantity of new elements but with many submissions would come the ability for the index to more readily discard what is personalized about a site vs. what is public and new content.

Just an idea – might also work for the feedlounge type application (if I understand feedlounge to be an rss reader) though since fewer people independantly request and cache RSS feeds seperately from their feedreader it might not be as useful).

Shannon

It’s like SETI@Home, only with feeds. And, you know, no green little aliens…

I wonder if this is a situation where distributing the load is the way to go. [I remember Scott Sanders bringing to me the idea of having FeedLounge client machines doing parsing and reporting the results to the centralized service. That's way cool and also impractical at this point.]

I fully admit that I’m thinking out loud here and am talking way out of my element. If you want to call me a space cadet at this point, I’m not going to argue; I did go to Space Camp twice, I live just down the road from it now, and I do work in manned spaceflight.

On the other hand, if you went with a centralized service, you’d have your scalability concerns, as well as privacy, security, it goes on and on. Let’s face it, if this service caught on, people would want to use it everywhere: random blogs on the net, their own blogs, their company blogs. That’s the potential for a lot of information, some of which may be private.

If someone were to take up this project, I think they’d need to take (as my boss always says) one bite of the elephant at a time. Start with a very del.icio.us-ish service that simply lets you “clip” and tag places you’ve commented (url, date, time, possibly comment content). Once that’s in place, they could start building on those ideas (ok, now we’ll check and see if that comment page has an RSS feed for future comments, so we can check back in the future). One step (or one bite of the elephant) at a time.

Maybe if you guys have any specific ideas of exactly how you’d want such a service to act and behave, it’d be a good idea to write them up: what data you want captured every time you comment, what you want the display to look like, that kind of thing. That way, you lower the barrier for entry into this project a tad by getting the first step of the development process started before there’s even a developer to pick it up.

Blah, I sooo wish there were more hours in the day. There are so many great ideas floating around out there that I really think I could run with, but I just don’t have the time to dedicate to them!

Getting as many people in on the discussion as possible would also help. That way you guys could toss ideas back and forth and help refine the plans for what you all really want out of a package (tool) or service, cutting down on the “oh, that just won’t work” phase of development when / if someone came along to do it.

I would be very interested in following this it is definitely something that I also would like to see. I leave many a comment in places such as this – and don’t always (okay rarely) remember to monitor the conversation for follow-up comments. I try to leave a post on my own blog(s) linking back to the comment, but don’t always so that is at best an inelegent solution

Del.icio.us is a okay solution but also less than ideal. I want something that will proactively discover where I’ve left comments in the past and what’s more, besides my comments will allow me to follow other people’s comments (from individual friends to “a-listers” like whenever Robert Scoble leaves a comment somewhere etc.)

Shannon

In writing the previous entry, I remembered that Alex had corrected me in the comments over on his side of the conversation fence. Since I’d left myself a breadcrumb trail, I was able to find it pretty doggone quickly. Awesomeness.
I do get u…

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 prob…

[No kickback needed.]

And really … I have a to-do on this and everything. I’m trying to fully frame what I’m looking for here. What you use and what I use are both partial solutions; I’m looking for a full statement of the problem so that a solution [or solutions] may be formulated.

I’m still working on an answer to Chris’s question, too.

I’d be open to the possibility of starting a new project in the near future to satisfy this need, but I really don’t have a good grasp on what you’re looking for…