Re: bachelor thesis about social web annotation
Mushon | Shual.com
hi Jakob,
[I am continuing this
thread on the ShiftSpace discussion list, as more people might have
feedback for you there and we would love for your ideas to be shared by
the rest of the community, please register to
the list here]
we're very interested in your research and would be happy to assist.
Dan & I have just graduated so you can have the benefit of reading our thesis papers. (mine is online, Dan might upload his soon as well)
You can also check our thesis presentations: Mine & Dan's
apart from that there is some literature, not much of it academic on the subject of web annotations.
You might want to browse through my delicious tag 'metaweb' or 'shiftspace'.
Some things worth reading are:
http://www.wired.com/techbiz/media/news/2001/04/42803
http://www2003.org/cdrom/papers/refereed/p831/p831-dill.html
http://www.loosewireblog.com/2006/03/a_directory_of_.html (directory)
http://ideant.typepad.com/ideant/2005/05/facilitating_th.html
http://www.paulperry.net/notes/annotations.asp (old stuff)
http://portal.acm.org/citation.cfm?id=875984&dl=acm&coll=&CFID=15151515&CFTOKEN=6184618 (more old stuff)
We prefer speaking of metaweb rather than web annotation when we speak about ShiftSpace, as we see more potential to the system than just annotation or social annotations.
For example we see Greasemonkey, the system we're currently using as a part of our architecture as a metaweb platform (which is simply not inherently social in function).
We would love to discuss with you the option for practical collaboration with ShiftSpace too, so feel free to ping us on that as well (we would love to hear what you have in mind).
good luck with your research, feel free to contact us personally or the list with any question or idea you might have and want our feedback on,
cheers,
| Mushon Zer-Aviv | |
 |
Shual.com - design studio |
 |
ShiftSpace.org - an opensource layer above any website |
|
+
|
1-646-283-6057 |
Jakob Hilden wrote:
Hi guys,
I'm a student at the University of Duisburg-Essen in Germany and I'm currently working on my bachelor thesis about social web annotation. I came across your site and think it is a really great approach.
Since you are university students also I though there could maybe be some ways to cooperate.
Right now I'm still in my research phase and I'm not yet sure what the exact topic of my thesis is going to be. But it is most definitely going to be about social web annotation. So far I'm working on the introduction where I will simply present former and current annotation systems.
When I originally had the idea I thought of developing my own prototype annotation system as my thesis. But since a whole bunch of systems have been popping up everywhere lately, I kind of discarded that idea.
Now I'm thinking about maybe benchmarking a couple of systems (with a focus on user-centeredness), but I could also imagine developing an extension for another annotation system (> shiftspace?)
Well, what do you think? Could you imagine cooperating in some way?
As I said, right now I'm still researching. Do you by chance know of any good recent literature about social web annotation?
You can find at least some of the articles I have been reading here:
http://www.citeulike.org/user/jhilden
Hope to hear from you,
--Jakob Hilden
Johan Sundström
> hi Jakob,
>
> [I am continuing this thread on the ShiftSpace discussion list,
> as more people might have feedback for you there and we
> would love for your ideas to be shared by the rest of the
> community, please register to the list here]
Good idea. Welcome to the § community, Jakob!
I'd suggest thinking of ShiftSpace as not just the annotation system
that it presently implements, but as client side framework for
delivering the functionality and UI for any annotation system's
content (for instance that of Hoodwink'd), using whatever back-end
they implement.
> we're very interested in your research and would be happy to assist.
Speaking as a loosely involved Hoodwink'd contributor/developer, I'd
warmly welcome you to help out porting that client layer to live atop
ShiftSpace too. It's something I've been planning on getting to at
work (for testing out §, though for now put on ice until a § license
is finalized), but I'd be happy to assist if you'd get to it first.
--
/ Johan Sundström, http://ecmanaut.blogspot.com/
jakob
thanks for the warm welcome. I'm sure we will be able to work well
together.
@Mushon: I read your thesis and was really thrilled. Some really good
writing and a lot of great thinking. It's by far the best thing I
have read yet on this topic.
@Johan: Hoodwink'd sounds really interesting, but i fear porting it to
ShiftSpace wouldn't be a sufficient topic for a thesis project.
As I said, I still haven't found the exact topic I want to focus on
for my thesis. So, for my own orientation and maybe for somebody else
to give feedback, I'm just gonna list some of the things i find
potentially interesting so far:
* How can people make PRACTICAL (daily) use out of a system like
ShiftSpace? What would the decisive factors to influence this be (>
interface, usability, ...)?
* How ARE people using a system like ShiftSpace? [How much data is
there already available? Enough for some research?]
* Can a system like ShiftSpace become a mainstream application? (e.g.
see some of the comments at the techcrunch article
http://www.techcrunch.com/2007/04/10/5-ways-to-mark-up-the-web/#comments
for this topic)
* Can a system like ShiftSpace have a 'real democratic effect' on the
internet/society? Could users/communities/poeple/humanity really
benefit from such a system, or is it just pure theory?
... just some thoughts. I hope I will be able to decide on something
soon. Feedback is very appreciated.
Another question would be, whether there already is a concept for the
community/social network part of ShiftSpace?
I'm really interested in that field and could also imagine doing
something in that direction. I just developed a social software
prototype application in Ruby on Rails (with users, profiles, friends,
groups, ...) for a university project. Maybe this could be reused for
ShiftSpace in some way?
Greetings, --Jakob
Johan Sundström
> @Johan: Hoodwink'd sounds really interesting, but i fear porting it to
> ShiftSpace wouldn't be a sufficient topic for a thesis project.
I assumed your mentioning developing your own prototype annotation
system or an extension for another annotation system was just a part
of a means to some end for some larger scope project that you were
noodling on.
Anyway, I liked your mentioning user centricity; few people do that
with a scope of the entire web, today; there are few players in the
software and content delivery communities today that transcend the
fragmented mini-puddles imposed by the underlying structure of the
web's (every site doing its own thing in its own way). ShiftSpace
would, and Hoodwink'd does, letting users learn interface constants
that look and work the same way, pan web.
> As I said, I still haven't found the exact topic I want to focus on
> for my thesis. So, for my own orientation and maybe for somebody else
> to give feedback, I'm just gonna list some of the things i find
> potentially interesting so far:
>
> * How can people make PRACTICAL (daily) use out of a system like
> ShiftSpace? What would the decisive factors to influence this be (>
> interface, usability, ...)?
I think you ought to try using one and become part of some social
network sharing a com channel atop the web, to get this into
perspective. It can be more or less the same idea as communities and
social sites like Facebook or MySpace, but with the whole web as your
social space, and powerful aggregate views to track conversations of
interest to you, wherever they happen, in context, or from the
laid-back comfort of your feed reader.
> * Can a system like ShiftSpace become a mainstream application?
Naturally. Without serious efforts on findability, usability and end
user value? Hardly.
> * Can a system like ShiftSpace have a 'real democratic effect' on the
> internet/society? Could users/communities/poeple/humanity really
> benefit from such a system, or is it just pure theory?
Possibly, but we'll probably see penetration in really small niches at
first, and once/if it takes off, there will likely be a strong lobby
towards outlawing or otherwise taking legal action against the
invasion atop domains formerly under the sole control/jurisdiction of
site owners, if I were to guess (based on the content industry's other
legal moves in recent years, once some amount of control is starting
to slip out of their hands).
> ... just some thoughts. I hope I will be able to decide on something
> soon. Feedback is very appreciated.
>
> Another question would be, whether there already is a concept for the
> community/social network part of ShiftSpace?
There is one in Hoodwink'd, anyway; I believe some of it may have
already been touched on-list.
> I'm really interested in that field and could also imagine doing
> something in that direction. I just developed a social software
> prototype application in Ruby on Rails (with users, profiles, friends,
> groups, ...) for a university project. Maybe this could be reused for
> ShiftSpace in some way?
Then you might be glad to hear that Hoodwink'd's server side back-end
is running on Camping, a Ruby microframework imitating the guts of
Rails but without the memory/cpu overhead/footprint.
Mushon | Shual.com
@Jakob:
First I'm flattered you enjoyed my writing (not being a native English speaker/writer I'm always surprised it makes sense to people, you might have picked that little foreigner flavor in my writing, including the disproportionate use of brackets)
The social network aspect of ShiftSpace is an open task for you to lead if you wish. We always knew we will include a social network but we always had other things to take care of first. As you could see in my paper I have given that some attention but it is definitely open to be explored. Personally I see Hoodwink'd as the *most* interesting case of metaweb social network and I would definitely start my research there. Join it (and if you don't know how, we would tell you the trick, off-list...) and I think you will find a very interesting social network with dynamics like non-other.
ShiftSpace is currently using PHP & SQLite for the backend, but we try to keep the backend very very simple (as you can see in our SVN repository) and do most of the magic with JS on the frontend. I think you would easily find your way there. The way I see social software development it should be a custom made thing. I think you would probably have a lot of experience and design approaches to port from the previous social network you designed, but you would probably do the right thing by starting fresh on this one (it's more fun, anyway). In general, I'm saying, while social networks are starting to become a boring repetitive design convention, developing a metaweb social network can allow for tons of innovation, and again, fun. We would love to share that fun with you, or have you share that fun with us.
@Johan:
I love it: anonymity
> niche > popularity > public enemy (as deemed by private
sectors)
sounds like the theme for a Hollywood production, and I thought you're the European in the group...
I would add:
> get the girl > live happily ever after
anyway, looks like an exciting plot, I'm a believer in happy endings...
Your remarks about
usability and user centricity are very true, and are a very important
principle for us. For example, I believe almost every metaweb app I've
tried (and I tried a lot of them) has failed on being too intrusive.
This is what led us to the interface we use now which we worked really
hard to appear only when you need it, yet have you remember it exist...
I personally am a great believer that this would rise and fall on user
centricity, as it is we're asking quite a lot from the user: "Take us
with you anywhere you go".
In general,
I would love for this to evolve and we will definitely try to help you help us (or just help you in general) in any way we can,
cheers,
| Mushon Zer-Aviv | |
 |
Shual.com - design studio |
 |
ShiftSpace.org - an opensource layer above any website |
|
+
|
1-646-283-6057 |
Johan Sundström
> I love it: anonymity > niche > popularity > public enemy (as
> deemed by private sectors)
Heh. :-)
> sounds like the theme for a Hollywood production, and I thought
> you're the European in the group...
Correct (for the time being, anyway; it's likely subject to change).
Though I guess I'd retain European descent even in California, should
I eventually get US visa. ;-)
> I would add:
>
> > get the girl > live happily ever after
I wouldn't hold my breath for that, though stranger things have
happened (future girl friends being initially sighted in web server
referrer logs, and the like :-).
> Your remarks about usability and user centricity are very true,
> and are a very important principle for us. For example, I believe
> almost every metaweb app I've tried (and I tried a lot of them)
> has failed on being too intrusive. This is what led us to the
> interface we use now which we worked really hard to appear
> only when you need it, yet have you remember it exist...
Also the reason I got § on my radar in the first place, incidentally,
and bothered finding out that there are good merits to the code base
as well.
> I personally am a great believer that this would rise and fall on
> user centricity, as it is we're asking quite a lot from the user:
> "Take us with you anywhere you go".
Yep. A "don't be evil" clause somewhere in project values is sort of
mandatory for success in the field, I'd believe. Hoodwink'd made a
move a few months ago on that very subject, announcing its purging and
not keeping any future traffic logs for the click trails of the
Hoodwink'd user base (which by technical necessity are bestowed on the
back-end server), so even legal action / court orders to Hoodwink'd
maintainers could not retroactively breech the personal integrity on
the web of any of its users and deliver testament of their whereabouts
on the web if they were ordered to, simply because the data would not
exist.
A good thing I'm seeing for § is that it would likely get trust
management in the hands of the user, allowing her to discern which
shiftspace layer providers she trusts to get data on her behaviour
patterns on the web, and simply opt out of the others. The privacy
concerns of metaweb applications are very real, and have not been the
focus of much discussion in wider circles so far, as far as I know,
though they definitely should be, both for things like §, and indeed
many user scripts as well. I think metaweb users should be informed of
the implications of the software and tools they use, and am fairly
certain they will be. But we shouldn't force the knowledge on them, at
the expense of usability, easy adoption and pretty user interfaces.
(And were I to decide, we'd keep them way out of reach of any lawyer
speak too. Legalese is a toxic byproduct of the American way of life,
I think. ;-)
Jakob Hilden
to do :(
I just wrote a really long message through the web interface of
GoogleGroups and when I wanted to send it, I was signed off and the
message gone. DAMN! But well, so I'm gonna write it again per email.
First of all, I would really love to take on the task to build a social
network for ShiftSpace. I think ShiftSpace is a great tool, I think
social networks are a extremely interesting field, especially in
connection with the metaweb, and I'm very sure that the mentors of my
thesis are gonna think the same (although I will still have to ask them).
So that you know where I'm coming from I'm gonna give you some more
information about my thesis. We have to write a ~40 page scientific
paper for our bachelor degree in the field of human-computer
interaction. I got interested in social web annotation around last
summer so I have known for quite a while, that this was going to be my
broad topic. But I need a more focused "research question" the thesis
should also contain some form of practical part. I think something like
"Social Networking on the Metaweb -- Design and implementation of a
social networking functionality for ShiftSpce" could be something
appropriate.
The social networking features that have been described in Mushon's
thesis and in the screencast on the website, look all very good to me.
I especially like the filter mechanism with the slide control (me <-->
world). I think that's going to be really helpful for the users. Now we
only need a social network to make this feature work.
The most important question that comes to my mind is: How and where is
this social network going to be build? As Mushon wrote, the architecture
of ShiftSpace so far is a slim backend with a real fat client. But I'm
not sure if the console would be the right place to view and build a
social network? I think you will probably need some more space (>
regular website) to create a real "social awareness" of the network and
it's users. So should the console (to view and created shifts) maybe be
kind of separated from the social networking functionality? So it's not
going to get overloaded.
Another thing I have been thinking about is user profiles. Detailed
user profiles are a very central component of social networks and they
are usually designed to encourage people to reveal as much information
about themselves as possible. But ShiftSpace has the policy to get as
little information from the user as possible. How is that gonna go together?
And now to some more practical stuff: How am I gonna get started coding
for ShiftSpace? Can you set up an account for me for the repository?
Should start with "installing" an own version of ShiftSpace on a local
webserver? Or what would you suggest?
@Mushon: No, I really liked your paper. It was really good to read,
which is not to common with papers dealing with computer science topics.
About Hoodwink'd: I installed that and I'm currently exploring it. It
seems really cool and I'm sure there are going to be plenty things to be
adopted in ShiftSpace in terms of social networking functionality.
Greetings, --Jakob
Clay Shirky
> this social network going to be build?
The phrase thats been running around my head lately is "Every URL is
a latent community", which is to say that the set of people looking
at any given URL might also have something to talk to each other about.
In practice, as with most social things, failure would be the rule.
Most latent communities wouldn't become active, even offered a low-
cost way to do so -- Google is too generic, Danni.com is too private,
livejournal already offers the network internal cohesion, and so on.
These difficulties would cover well into the 99th percentile of cases.
But 0.01% of URLs is still in the millions, and latent groups around
pages hosting parkour videos, Egyptian democracy activism, Andean
travel guides etc might be happy to interact.
> As Mushon wrote, the architecture
> of ShiftSpace so far is a slim backend with a real fat client. But
> I'm
> not sure if the console would be the right place to view and build a
> social network?
View, probably. Build, no.
I think I may differ from Dan, who is more committed to
decentralization than I, but by observation the building and
traversal of large-scale directed graphs works better on managed
servers than on distributed clients. Everything from LJ's friends
page to interestingness on Flickr (to say nothing of the MySpace and
Google algorithms) models and traverses the networks on big iron.
The one get-out-of-jail-free card I can see in all this is that the
N^2 problem is only a problem when N is large, and in the clusters of
a social network, N is not large most of the time. So as long as your
search horizon is limited to FOAF (2 degrees of separation) instead
of F(OAF)^4 (6 degrees of separation), you might be able to
parallelize some of the graph traversal on the client. You'd miss
real-time updates and large views, but if decentralization is
philosophically important, that may be a workable strategy.
> Another thing I have been thinking about is user profiles. Detailed
> user profiles are a very central component of social networks and they
> are usually designed to encourage people to reveal as much information
> about themselves as possible. But ShiftSpace has the policy to get as
> little information from the user as possible. How is that gonna go
> together?
Maybe the policy for a social network on SS should be about getting
as little *required* information, and putting as much control of it
in the users hands as possible?
-clay
Dan Phiffer
Sorry to get in so late on this thread. Very exciting stuff
happening! Jakob -- welcome, glad to have you with us!
More written inline below...
On Jun 5, 2007, at 12:44 PM, Clay Shirky wrote:
>
>> The most important question that comes to my mind is: How and
>> where is
>> this social network going to be build?
>
> The phrase thats been running around my head lately is "Every URL is
> a latent community", which is to say that the set of people looking
> at any given URL might also have something to talk to each other
> about.
>
> In practice, as with most social things, failure would be the rule.
> Most latent communities wouldn't become active, even offered a low-
> cost way to do so -- Google is too generic, Danni.com is too private,
> livejournal already offers the network internal cohesion, and so on.
> These difficulties would cover well into the 99th percentile of cases.
>
> But 0.01% of URLs is still in the millions, and latent groups around
> pages hosting parkour videos, Egyptian democracy activism, Andean
> travel guides etc might be happy to interact.
Yeah, actually I just discovered that Lisa Cho (another recent ITP
grad) is also a regular visitor to The Morning News (http://
themorningnews.org/) because of a note she left there. It was one of
the few times I've been surprised by a shift.
The interactions I'd have enjoyed in this case would have been:
1. Indicate to Lisa that I've found her note
2. See what other "URL groups" she's also a part of
3. Attach discussion threads to the "headlines" section of the site
4. Subscribe to subsequent activities at that URL
>
>> As Mushon wrote, the architecture
>> of ShiftSpace so far is a slim backend with a real fat client. But
>> I'm
>> not sure if the console would be the right place to view and build a
>> social network?
>
> View, probably. Build, no.
>
> I think I may differ from Dan, who is more committed to
> decentralization than I, but by observation the building and
> traversal of large-scale directed graphs works better on managed
> servers than on distributed clients. Everything from LJ's friends
> page to interestingness on Flickr (to say nothing of the MySpace and
> Google algorithms) models and traverses the networks on big iron.
>
> The one get-out-of-jail-free card I can see in all this is that the
> N^2 problem is only a problem when N is large, and in the clusters of
> a social network, N is not large most of the time. So as long as your
> search horizon is limited to FOAF (2 degrees of separation) instead
> of F(OAF)^4 (6 degrees of separation), you might be able to
> parallelize some of the graph traversal on the client. You'd miss
> real-time updates and large views, but if decentralization is
> philosophically important, that may be a workable strategy.
With the "me <-> world" slider mock-up, we only go as far as FOAF ( +
everybody). Which I guess you're saying could be implemented within
the client. I see "more robust privacy-centered architecture" as
philosophically important, and decentralization as an extension of
that if it's the best option for achieving it.
As Johan was describing in the case of Hoodwink'd & their server
logs, I don't like bearing the risk of data being harmfully
exploited, especially since we're not seeking to earn money from it
(as in the case of Google). This trust management option seems like a
good one to me. Maybe api.shiftspace.org is just one server that
users could choose to opt-in to.
>
>
>> Another thing I have been thinking about is user profiles. Detailed
>> user profiles are a very central component of social networks and
>> they
>> are usually designed to encourage people to reveal as much
>> information
>> about themselves as possible. But ShiftSpace has the policy to get as
>> little information from the user as possible. How is that gonna go
>> together?
>
> Maybe the policy for a social network on SS should be about getting
> as little *required* information, and putting as much control of it
> in the users hands as possible?
I agree -- profiles should require as little required as possible.
One thing I think we should start requiring is email addresses. I
don't think there's a better way of retrieving forgotten passwords
for users.
> And now to some more practical stuff: How am I gonna get started
> coding
> for ShiftSpace? Can you set up an account for me for the repository?
> Should start with "installing" an own version of ShiftSpace on a local
> webserver? Or what would you suggest?
I'll send you a ShiftSpace svn login offlist. Take a look at the
readme document (http://shiftspace.org/code/trunk/README) which
should explain how to get a local server running. We haven't really
codified submission policies yet, but in general try not to commit if
things are in a broken state. If you plan on refactoring something,
just do it in the "branches" section until it's relatively stable.
I'm sure there are more details I'm forgetting, but for now, welcome!
-Dan
Johan Sundström
> Hi, sorry that it took me so long to respond, but there is just so much
> to do :(
(3 days later: :-) It happens to us all; don't feel bad about it. :-)
> First of all, I would really love to take on the task to build a social
> network for ShiftSpace.
I think this is how a technician thinks of the task, equating crafting
functionality to forging a social network. That's the *least* bit of
work, and it's an on-going project that needs maintenance, love and
improvements throughout the entire lifetime of the social network.
Building a social network is about people, not technology.
Being in for the long haul (i e planning on sticking to the community
for years to come), building one from scratch *could* be worthwhile,
despite repeating what others have done over and over again, as I
believe Mushon & c:o did and aim for. It's a little like fatherhood;
producing offspring is the least concern, raising and nurturing them
into secure, mature and successful adults is the real challenge, and a
long term endeavour.
If not, i e the project is a thesis project with a start, an end and a
paper result, I'd suggest directing the energy between those points in
time towards picking up an already forged network, boosting it and
improving it as you go along, building on top of both its present
strengths, user base and pool of experience and history of success.
It's of course not me doing your choices, but I'd grab Hoodwink'd in
an instant. :-)
(My own position is that unless one has some pretty smashing idea and
an inexhaustible source of promotion energy, coming up with a new
social network is about as well spent time as trying to launch a
competitor to e-mail. Not impossible. Many opportunities to do better
than there is, sure. But the world would be better off with e-mail
being incrementally better than it is.)
> So that you know where I'm coming from I'm gonna give you some more
> information about my thesis. We have to write a ~40 page scientific
> paper for our bachelor degree in the field of human-computer
> interaction.
Cool. With that core focus, taking focus off technical backend issues
and working on the user end instead with real users ought to be a good
thing too.
> I got interested in social web annotation around last summer so I have
> known for quite a while, that this was going to be my broad topic. But I
> need a more focused "research question" the thesis should also contain
> some form of practical part. I think something like "Social Networking on
> the Metaweb -- Design and implementation of a social networking
> functionality for ShiftSpce" could be something appropriate.
That sounds a lot more like developer-computer interaction to me.
Shouldn't human-computer interaction questions be more along the lines
of "the metaweb - dissecting elements forming social networks
orthogonal to the web, on top of the web"?
> The most important question that comes to my mind is: How and where is
> this social network going to be build? As Mushon wrote, the architecture
> of ShiftSpace so far is a slim backend with a real fat client. But I'm
> not sure if the console would be the right place to view and build a
> social network? I think you will probably need some more space (>
> regular website) to create a real "social awareness" of the network and
> it's users. So should the console (to view and created shifts) maybe be
> kind of separated from the social networking functionality? So it's not
> going to get overloaded.
Hoodwink'd has a central aggregator site
(http://hoodwink.d/onslaught/), where you can see lots of stats of
what is happening now in the entire commentary network, who is saying
what where, most recent threads, sites most commented, recent users,
and so on and so forth. You can do searches, subscribe to RSS feeds
for the hits they generate now and in the future, and search on cool
things like written by me (or some other user / users), new post to a
comment thread where I (or perhaps someone else) have posted, and
much, much more, so users feel very connected and so you can drop a
question wherever on the web, and, in case someone finds it and
responds to it, you get alerted in your feed reader, if you subscribe
to such a feed (I used to, when I was more active). An example feed
listing posts where I have posted, but not by me:
http://hoodwink.d/onslaught/search.xml?q=anywho:ecmanaut+!who:ecmanaut
It's a *really* powerful improvement to the web, and it uses its own
framework to great effect, i e using itself to provide commentary
functionality to all of its own pages; your profile page (where only
your alias, date of creation and some stats on your activity, plus
user comments, show up), and a separate discussion board site, that is
just one url per subject, all serving blank pages with some CSS
(http://wasteland.hobix.com/), forum too built using Hoodwink'd's
client side default functionality.
So yes, having a whole site, beside the console, helps immensely. But
would you have to build it yourself to do your study? Or rather, would
you have to build anything non-client-side to fulfill the project?
> Another thing I have been thinking about is user profiles. Detailed
> user profiles are a very central component of social networks and they
> are usually designed to encourage people to reveal as much information
> about themselves as possible. But ShiftSpace has the policy to get as
> little information from the user as possible. How is that gonna go together?
Sounds like a non-problem to me, but I'm Hoodwink-biased too.
http://hoodwink.d/ecmanaut is good enough for me, listing above
details. There is the option of decorating your messages with a custom
background image too, for those wanting to personalize their style a
bit wherever they comment, too.
> About Hoodwink'd: I installed that and I'm currently exploring it.
Great! Hope this was useful.
> It seems really cool and I'm sure there are going to be plenty things to
> be adopted in ShiftSpace in terms of social networking functionality.
Still doesn't feel appetizing giving Hoodwink'd ShiftSpace leverage
instead of copying Hoodwink'd features to ShiftSpace (or a
new-from-scratch ShiftSpace layer)?
Clay Shirky
> an inexhaustible source of promotion energy, coming up with a new
> social network is about as well spent time as trying to launch a
> competitor to e-mail. Not impossible. Many opportunities to do better
> than there is, sure. But the world would be better off with e-mail
> being incrementally better than it is.)
I agree with your point that viewing this as a thesis, rather than as
a project of which the thesis is merely the first step. I also agree
that trying to displace MySpace et all is a mugs game.
However, its possible to build technological support for social
networks without competing with Myspace. The reasons Friendster et
fils work is that humans form social networks quickly and well. The
fact that Facebook has a network in the tens of millions doesn't
obviate, say, Meetup, which creates real world social networks, or
HowardForums or Gaia, which create networks tied to narrower affinities.
What Facebook can't do is say when two of its users are looking at
the same page. Now it may be possible, given their newly opened APIs,
to fuse SS plus FB in a way that makes that possible, but the basic
intuition of a human network forming around an annotation tool isn't
an FB replacement so much as an alternate way of taking advantage of
human talents and desires for conversation.
-clay
Johan Sundström
> I agree with your point that viewing this as a thesis, rather than as
> a project of which the thesis is merely the first step. I also agree
> that trying to displace MySpace et all is a mugs game.
MySpace/Facebook vs brand new shiftspace layer is an almost as unfair
comparison as my e-mail example. :-) The Hoodwink'd vs from-scratch
comparison is less apples and oranges, since both are metaweb and
share scope.
But I'll try to stop pushing the let's-adopt-Hoodwink'd argument now;
I've been much, much too wordy already. I'm mostly waving that flag to
see if someone else would like to do it, or if I should do it myself
some productive late night session. It's probably an interesting
exercise melding them, that will shed some light on modifications that
would be useful to add to § to accommodate another metaweb system it
was not explicitly designed for from.
A side thought: maybe ShiftSpace ought to listen in to the § key too,
for keyboards that have one. The Swedish keyboard layout, for some
inexplicable reason, has that symbol mapped to the key to the left of
1, right below escape. It's got the onKeyPress keyCode 168, same as
its unicode code point (according to
http://mochikit.com/examples/key_events/).
Jakob Hilden
sorry I couldn't join the inner circle meeting yesterday, but I wasn't
home.
> I think this is how a technician thinks of the task, equating crafting
> functionality to forging a social network. That's the *least* bit of
> work, and it's an on-going project that needs maintenance, love and
> improvements throughout the entire lifetime of the social network.
> Building a social network is about people, not technology.
You are absolutely right. Thanks for pointing this out. I have the same
position on that subject but I guess I tend to not communicate it too
well or even forget it a little bit.
I'm aware that a social network is created by people and not technology
and that it is a long process that will take much more time and work
than can be done in my thesis project. But I think there is a lot of
interesting things that can be done concerning ShiftSpace and social
networking. And I think ShiftSpace would benefit from it in any case. I
know, that at the end of my thesis project there won't be anything close
to a complete social network, but I hope there will be some
functionality that should enable users to build one and there should
have been a lot of valuable thoughts spent on that issue.
> If not, i e the project is a thesis project with a start, an end and a
> paper result, I'd suggest directing the energy between those points in
> time towards picking up an already forged network, boosting it and
> improving it as you go along, building on top of both its present
> strengths, user base and pool of experience and history of success.
> It's of course not me doing your choices, but I'd grab Hoodwink'd in
> an instant. :-)
I agree with you, that if I would be looking for a good sized thesis
project with a definite end I would go for Hoodwink'd immediately. I
think it's a cool peace of software and an interesting community, but
honestly its secret & exclusive character doesn't make me want to
contribute to it too much. I prefer the open and inclusive philosophy
that seems to be behind ShiftSpace much more. And if it will continue to
develop in this same way, I'm sure I will stick with the SS community
long beyond my thesis project.
I really liked Clay's remark that "Every URL is a latent community".
They should most probably be the center around which a ShiftSpace social
network arranged. Maybe in some cases URL could be a little to specific
and one should get with the whole domain, but that will be figured out.
I also want to point out that this should not be about creating any type
of competitor to existing networks (MySpace, FB). It's just about doing
whatever needs to be done to make ShiftSpace more useful. And I think
that would definitely be some social functionality, to find the shifts
that you might be looking for.
Right now all my ideas of how this could work are kind of fragmented and
I need to sit down and organize my thoughts into some global strategy.
Other than that I have checked out the code and started a local
installation. I will be traveling for the coming week but I will
certainly get back to you once I have it all set up next week.
Greetings, --Jakob
Clay Shirky
> They should most probably be the center around which a ShiftSpace
> social
> network arranged. Maybe in some cases URL could be a little to
> specific
> and one should get with the whole domain, but that will be figured
> out.
I doubt there is a way to handle this programmatically, because this
has to do with semantic overloading of unique identifiers. This is an
interesting problem, and probably one only the users can figure out.
If you like something of Anil Dash's, you may be interested in most
things at the domain dashes.com. If you like something of Tila
Tequila's, you will probably not like everything at the domain
MySpace.com.
> I also want to point out that this should not be about creating any
> type
> of competitor to existing networks (MySpace, FB). It's just about
> doing
> whatever needs to be done to make ShiftSpace more useful. And I think
> that would definitely be some social functionality, to find the shifts
> that you might be looking for.
It always starts that way -- the obvious use of collaborative
filtering, back to Ringo at the Media Lab, is "use people to find
interesting items", but over the long haul, these systems tend
towards a reversal -- "use the items to find interesting people."
This is what happened to Firefly (the commercial version of Ringo.)
My bet is that the finding function will be the early use, but that
communities of practice will be the long-term benefit (assuming, as
usual, that you can overcome the cold-start problem in the first place.)
-clay
Johan Sundström
> > I think this is how a technician thinks of the task, equating crafting
> > functionality to forging a social network. That's the *least* bit of
> > work, and it's an on-going project that needs maintenance, love and
> > improvements throughout the entire lifetime of the social network.
> > Building a social network is about people, not technology.
>
> You are absolutely right. Thanks for pointing this out. I have the same
> position on that subject but I guess I tend to not communicate it too
> well or even forget it a little bit.
Your thoughts on Hoodwink'd not being very open (as in setting a high
opt-in threshold enshrouded in technical mysticism) are probably about
as right too. :-) I'll get off that case and prod at it myself when I
find the time (and probably be glad I did, as it forces me to pick up
on all relevant aspects of ShiftSpace core while at it :-).
At this point in time, we are probably better served with a wealth of
different initiatives trying to make fruitful use of ShiftSpace core,
than aiming for transporting a live community with a working client on
to a less time-tested unknown one. :-) Yesterday's meeting suggests
there are a few, and that we'll see rather different-scope projects
trying it out.