Lists Home |
Date Index |
* Bullard, Claude L (Len) <firstname.lastname@example.org> [2005-08-12 11:42]:
> From: 'Alan Gutierrez' [mailto:email@example.com]
> > There's a problem with trolls that will turn comments off on
> > most high-traffic blogs, especially political blogs. Culling
> > comments can take up a lot of time for bloggers.
> It's a barnyard. The chicken fights dominate until the bulls
> go at it. Have you noticed the uptick in well-placed articles
> warning people that 'anything you blog can and will be used
> against you in the court of Human Resources'? Blogs scare
> the MBAs witless. Blogs out branding, they out attempts to
> use language to hide the truth, they out the con artists and
> the fakirs who can push a stock price into a bubble by gutting
> the employees's benefits without doing anything to get product
> out the door. The blogs scare the ownership society. They
> force them to confront the customers face on without the
> protection of their spin meisters and closed door meet-and-greets.
> They can co-opt blogs, but they can't change the essential
> openness. Vampires don't like bay windows on their crypts.
> > I've always felt that conversations ought to take place between
> > blogs, if there were applications that could piece the
> > conversations together.
> I agree with that. On the other hand, it seems to me you are
> describing a semantic aggregator.
A who? (You folks are, often times, hard for me to follow.)
> So I push back to the 'unstructured information' thread. We can
> make some progress here by staring a bit more deeply into systems
> that take raw or semi-structured data and categorize them for the
> sake of auto-linking.
I'm looking for simple things. My premise is that the newer
applications are going to be the product of human interaction
and human intervention.
Auto-linking? I'm interested in creating a way to have a link
carry more information, and have that information assigned,
quite manually, by the link author.
Then allow the genetics of search algorithms and user's gaming
there evolve (shh! I didn't say it) the structure of the links.
Or search may simply give way to referral. A powerful aspect of
W20 is the accountability of blog. It permits trust.
(We've always had listserv for these things. W20 is also return
to when newsgroups were troll free.)
It ties into your other thread, in that it took ten years to get
the web organized in such a way that is W20. (W20 is about
separate the presentation from the data. Your web site is a
separate interface from your RSS feeds. Creating permalinks, so
that links have lasting meaning.
Creating a richer link, simply creating the data structure,
ought to be enough to get the ball rolling. Trying to skip what
looks like a long and painful process of creating really:funky~tags.
> Note that in the UIMA architecture, those components are
> pluggable, so in theory, one could subscribe to a service where
> the last component is the categorizer by the description of that.
> I'm not sure that gets us beyond the Fox News Network model in
> blogs although they are very useful in databases where connecting
> the dots is the key to the mission.
> > A blog conversation protocol, then, in answer to your question
> > at the end of the post.
> Please describe that. It's an interesting idea but see above.
From a user's persepctive:
I want to move to Chicago. I will not blog about moving to
Chicago. After I've move to Chicago, I'll blog about living in
the Chicago. Not the same thing.
I'd like to people to take note of my desire to move to Chicago,
and tell me about the neighborhoods, the rents, the jobs, the L,
and other things.
If I could start a conversation on my blog, I could look through
my sub-blogosphere, and invite people to join in this
conversation, or put this conversation at a portal, and have it
tracked under a category that I assign, or join an existing
People could host conversations, if I'm perhaps, a Chicago
realitor. (Yes, it can be gamed. But people will leave the
conversation if it is.)
Thus, it's the idea of the dynamic creation and dissolution of
group blogs. An enclosure for the entries.
> > >The recognition of the "A-List" preceeds the creation of a market.
> > And just as markets have leaders, disruptors emerge to threaten
> > the A-list. "Burn down the mission, if you want to stay alive..."
> > Sure. The A-list will be relevant always. You can't have a
> > society without celebrity, and vice versa.
> If society is monolithic, true, but I see a world of different
> A-lists with occasional cross-referencing.
Exactly. We'll all have our own corner of the blogosphere.
> If you look at the political blogs, you don't see so much
> a conversation, but the same self-admiration.
I'm finding more blogs to be of the latter than the former. The
absence of comments on political blogs is because the trolls are
completely unacceptable, and they take over the comment section.
And there are windbags.
> On the other hand, I've stumbled into scientific blogs discussing
> say, quantum theory, and the arguments are a lot more formal and
There are nice places were good people congregate and share
ideas. There are a couple hyper-local blogs in my area that are
just excellent at discussing zoning and development, city
council elections, etc. I'd hold that up as an example of where
we're going in this discussion. I'd never get news about the
city council without http://annarborisoverrated.com/, since I'm
not going to read a cluttered, inky dead tree every day.
This is the space that interests me, not hyper-local
specifically, but sub-blogosphere. (Though hyper-local will
become a very powerful force in the U.S. to combat the isolation
of ex-urban sprawl.)
> So again, it n-furcates. That's fine. The utility of a semantic
> aggregator is one that can connect the dots among concepts that
> aren't labeled the same, but have hidden couplers. Outing hidden
> couplers is the prize. Then we learn and the web is doing what
> I've always believed such a system should do.
Okay, n-furcates. Let's agree on that. A semantic aggregator. I
see it now.
Blogs to me are nodes in a graph structure. Duh. But far more so
than web pages of old. They are OOP objects, rather than
procedural data structures. They include methods as well as data.
Create a few clever WordPress/MovableType plugins, get people to
adopt them via the blessed network effect, and you can have an
graph of documents with objects that run their methods in their
Maybe you can feed those objects XSLT/XQuery/XPath.
Can your semantic aggregator is built from this data structure?
I call this searching in the small. Rather than searching a via
an search engine, simply point the, uh, semantic aggregator
(right?) at a feed you find interesting, or an article you find
It builds a new feed from the links in that feed. The user could
read through blogs in a feed reader, and get summaries and
rankings of the links off of each entry, and decided whether or
not to follow the links. (Problem for those of us with 100+ feeds.)
> > Most of what exists today addresses the needs of the A-List,
> > however. Blogging software itself makes sense for someone who is
> > trafficked, but doesn't make as much sense for a programmer like
> > myself, who blogs only when time permits.
> I don't understand. Which features address the steep curve vs the
> long tail?
> > A personal CMS, makes more sense. With feeds, of course, but
> > also with versioning, for the essayists among us.
> Ok. Please expand.
It's the approach.
I'm not going to read through blogs, quote, remix, and comment
on everything I see. I'm better of cutting code.
When I do write about something, it's doesn't lend itself to an
article. I'd like to keep notes about my programming in my blog,
but it ends up reading like a CVS log. Fixed an off by one
error, renamed a class, checked in this, checked in that...
If I'm working with, say, Java NIO. I'd rather create an essay
about Java NIO. I'd update it with what I learn, and create a
definitive work that captures what Alan Gutierrez knows about
Java NIO, rather than try and search disparate blog entries for
notes, blurbs, and links.
It won't say anything that the documentation doesn't say. It
will be a record of my experience, however, and easier for me
to absrob when the sponge is dry.
Of course, this definitive work is available to protential
employers as well.
The long tail aspect is that my experience with Java NIO may be
in the recesses of my mind, but it's still there, and even if I
don't need it, someone else may want me to recall and employ it.
> > > There are going to be new protocols for the construction of
> > > edges in the network, that are richer than tags and URIs, and
> > > the meaning of the edges will not be entrusted to a ranking
> > > algorithm.
> >> That's the music market. You may not like what emerges here, but selah.
> > Lost me.
> The music market is based on bribery for airtime. There are subtle ways
> to achieve that and few if any ways to avoid it.
> > > Rather, it will be negotiated by the nodes, the individuals, who
> > > will agree on a meaning of the link.
> So we are back to 'good for whom'?
> > If it is important enough, it will be packaged, marketed and gamed.
> >> Yes, yes, yes. It's all good.
> It's all traditional, that's for sure.
The system can be gamed, but I think the system will be
I don't think that the threat of manipulation is such that this
architecture is a threat to society.
If you want to get into it, let's fork the thread, social
implications apart from different from application and use.
Please, since it's simply too difficult to do all this quoting.
> > > I'd like to enhance exsiting blogging software so it can act as
> > > a node in a graph data structure, that can be accessed
> > > programatically, and see what new applications can be created.
> > That's what it is now. What programs are you going to add?
> > I'd add an XML/REST interface to blogs for searching that blog.
> Ok. That's good.
> > I'd also add a XML/REST vote interface, that would allow anyone
> > to rank an article on that blog.
> Ok. Somewhat like the Microsoft voting on their technical pages?
Thinking about that link data structure. One that describes a
link that exists in the feed. A scale of 1 to 10 is a starter.
Might wish to put tags on a blog entry and invite people to rank
acording to to those tags. Maybe just make it yes or no. Less
complexity, less opportunity for mischef.
> > I'd let to go wild, address the problem of gaming later, (or
> > maybe after I get some sleep).
> It's a difficult problem much like trying to keep voting exit
> surveys from influencing voters in other time zones. Get some
As far as anti-gaming, it's in the auditing.
Perhaps, I only allow people I know to rank me, for better or
best. I'm imagining a system that follows the principle of, if
you don't have something good to say, don't say anything at all.
I'll let you rank my pages, until you start giving me rankings I
don't like. Then I'll prevent you from ranking, maybe even
delete your rankings.
But if I do this to enough people, I have no rankings, and I
simply don't exist.
In fact, I can go ahead an change your rankings, but a third
party will XML/REST your blog for a list of the rankings you've
given me, and get wise to my games, and ignore those rankings.
Positive social networks.
> It's Friday; we can turn back into humans tonight. :-)
That's good to hear. I didn't know what I was going to do with
all this fur.
Alan Gutierrez - firstname.lastname@example.org