Inkwell: Authors and Artists

    
personally, I think tentative is appropriate.  Some things do work in
some social contexts and not others, after all. 

(Don't get me started on Yahoo and Flickr.  Yahoo is smart in many
ways, but the interactions with the Flickr user experience have mostly
seemed ham-handed, from out in user-land. Changing the password system
and adding "Yahoo" to the logo felt like a nervous dog lifting a leg,
from some intelligent fraction of the oldskool community point of view
in the community there.)
  

    
I'm probably going to steer clear of apologias for Yahoo! and say it's
complicated, the constituents are many, and any kind of change in an
ongoing social environment is going to frustrate some of the people
involved.

Trying to keep up, I still need to talk about

   * how UI patterns and social patterns bear on each other

   * the differences between the social impacts of "vote to promote"
versus thumbs up/thumbs down? and we observed these effects?

   * the Social Mania card game and why we designed it

   * to what extent, in this "viral" climate, might the design of
sites be accidental to success (and example of follower models playing
out differently in different eras)

   * what I think about social media marketing and how well is it
served by today's discussions

   * how this current collection of patterns may fall short of what's
needed, what's been done in the past, or what was expected (or
promised) - Ari, I may ask Erin to drop by to talk about registration
and sign-up, as she took the lead on those patterns and may be able to
field your questions more effectively, though I'm happy to respond
myself as well. In the meantime, by all means please also tell me some
of the other patterns you found disappointing or wrong! I'll be making
notes on the wiki and trying to address that sort of feedback as part
of the "unbook" process.

   * What testing we do with patterns.

   * and Isn't Design of Sites great? (This one I can answer. Yes it
is. We have read it, do cite it, and are aware that it also uses a
pattern approach. One thing is that it's contents, like Tidwell's, are
only teased online, so the deeper linking from our wiki goes elsewhere.
Still, I cited DoS in our recent Accordion beta pattern in the Yahoo!
Library.
 
   * and we can still talk more about how monetization and making
sustainable businesses fits in and whether that should be viewed as
part of the pattern landscape related to designing environments for
social experience ("social design" and "social design patterns" for
short). and why banner ads may or may not be a good business model for
twitter. and how the valuation of such a company is done. those latter
two points i could mumble through but there are experts on such things
who could do a better job of it.

   * and some thoughts on tentative and incremental (dare I say
agile?) iteration, avoiding potemkin villages and being sure to pave
the cowpath and let your users teach you what the interface is really
good for...

(may need a brief moratorium on new questions on this lazy birthday
weekend of mine)
  

    
happy birthday!  maybe say something about prioritizing social patterns
when RL becomes more entertaining.

the cover of the new yorker this week has goblins ringing doorbells while
their parents stand nearby, busy on their iphones.
  

    
Happy birthday!

And, yes, the "Potemkin Village" anti-pattern is both true and witty. (In 
general, I loved the anti-patterns in the book--it was the patterns that 
felt incomplete.)
  

    
Happy birthday. Nice recap list above -- it gives a clue as to how
rich the area of inquiry and the book is.  
  

    
so (deep breath), first of all UI patterns and social design patterns
definitely affect each other. Many of the social patterns in our
collection are in fact UI patterns, or at least compounds of UI
patterns. The key difference as I see it is that social patterns reach
*deeper* into the core of an application than UI patterns tend to do.

This is not to say that UI and interaction design takes place entirely
on the surface, on the presentation layer. Slick, rich, Ajax-style
interactions for example require that data be loaded and made available
through the DOM, etc., so we have be careful about isolating any part
of a working application.

Nonetheless, the social patterns *especially* cannot be managed as a
surface, display, pixel-specifying, button-placement matter. In the
simplest sense, you can't add people or objects to your application
simply by adding UI widgets: they need to be reflected in the data
model.

At Yahoo we lump the patterns all together in the same library. This
may not scale well over time but we do it for the reasons that Gail
hinted at, that UI patterns and social patterns implicate each other
and need to call out to each other at times.
  

    
As for vote to promote versus thumbs up / thumbs down, we derived our
patterns from plenty of observation both on the Yahoo! network and
across the rest of the web. 

Bryce Glass, the shadow "third author" of this book, wrote the
original Vote to Promote pattern and the internal docs about thumbs
up/down that drove the pattern in our book.

THe primary difference between them is that voting up is a more
straighforward popularity-contest method of surfacing preferred
content. By contrast, vote up/down can bring out the possibility of
competitive gaming (voting down competing objects) and introduces a
sense of ambiguity: is something with a score of 10 the results of 11
upvotes and 1 downvote or 2000 upvotes and 1990 down votes?

(You'll notice that most such examples of this patterns, notably Digg,
Reddit, and Buzz do include a vote-down feature but tend to deprecate
it, styling it as a plain text link instead of a big juicy button, or
allowing it in the drop-in javascript but not requiring it, etc. - in
fact, Reddit uses a hybrid: the button dropped on publisher pages tends
to favor the vote-to-promote function, but the listings and
particularly the comment voting on the Reddit site itself foster much
more of a culture of voting up and voting down in competition.)
  

    
Well, if sites used TiVo-style collaborative filtering instead of
a single global popularity metric, then the metric would not
be subject to gaming.  It's not even that much harder to
implement.
  

    
slipped by Jef...

The story behind the card game is that Erin and I have been trying
throughout this project to encapsulate the information in multiple form
factors. We had the book and the website (and of course the Yahoo!
pattern library, where these patterns are gradually showing up as they
become blessed by Yahoo!'s somewhat more particular process) but we
knew that encyclopedic books and a dense website inviting group
collaboration are not the ideal channels for *everyone* to learn abut a
topic.

At Interaction 09 earlier this year we saw a UX designer from
Salesforce.com show some of what he called "postcard patterns" which
were basically less rigorous/HCI-esque patterns, designed to be taken
in at a quick glance, making good use of an illustration, calling out
key points, and keeping it brief - more designerly, less writerly.
These postcard patterns also double as good powerpoint-type slides for
talks, so we started thinking that as we were already rolling out our
speaking schtick to promote these ideas, that we should think about
coming up with shorhand versions of the patterns suitable for a
flashcard or projected slide.

Meanwhile, we had pitched along with the book idea that we would make
a poster to help visualize the landscape and perhaps a deck of cards.
The deck began in our minds as something more like flashcards or
oblique strategy cards: a stack of charismatic objects, fun to
manipulate and thumb through that could function as educational aids.

Then, as we started iterating the cards and figuring out what to put
on them, and continued thinking a lot about social design and game
design and how they interrelate, I think we both arrived at the same
insight: that we should try to design a card game you can play with the
cards, that would help people learn the dynamics and tensions at play
across the pattern collection and related ideas. This ended up
generating cards with content not drawn from the book (anti-social
circumstances and lucky events to spice up the game dynamics).

Matt Leacock, who I've mentioned before, is a longtime colleague of
Erin's, the first curator of the Yahoo! library, and an award- winning
game designer. He happened to have already spent some time in the past
imagine a game to model software design, so he applied some of these
ideas and met with Erin to sketch out a possible game play concept.

We started prototyping the cards and the game, did some play-testing
is small groups with an alpha set, and then printed up a first beta set
of cards, debuting it at the IDEA conference in Toronto in September
(where people mostly really loved it, despite some of the rough edges
to do with gameplay that we hadn't perfected yet).

I also used it in a workshop in Sydney and gave the game to the
attendees as a take-home. Each play of the game has taught us more
about what works, what needs tweaking, and - perhaps most important -
how to explain and teach the game rules.

There's now a beta 2 of the game available for purchase:
<https://www.thegamecrafter.com/games/social-mania---designing-social-interface
s---the-game>
and we'll be playing that as an activity at Interaction 10 in Savannah
next February. We're still trying to figure out the best way to
manufacture and distribute the game. We don't see it so much as a
commercial venture as something that supports the overall project and
should at least break even.

I've certainly learned a lot about game design so from that
perspective I'm glad we stretched ourselves (it seemed crazy at times
to give ourselves additional work on top of writing a book, putting to
gether a website, preparing talks for multiple conferences, etc), and I
hope to do some more game design in the future.
  

    
Jef, you're probably right. I don't think we really addressed the
algorithmic aspect of things like collaborative filtering but that
belongs in there as an option for sure and perhaps it's the answer to
all the complexity of voting systems. As I delved into voting I
discovered (to my chagrin) that there are scores of voting models each
tuned to optimize some aspect (things like instant run off, etc.) and
all of which also seem to have some weak spots, sort of along the lines
of Goedel incompleteness, but I'm honestly out of my depth there.

I would recommend the forthcoming Building Web Reputation Systems from
Randy Farmer and Bryce Glass (whom I've mentioned earlier and who
wrote most of the reputation patterns in our collection). It goes to
the heart of how to model and tune systems of object and people
reputation in a social site.
  

    
arriving at work now, got a skype call at 9 and have to hit the gym
and tackle some work but will check in at my lunch hour to keep working
through the backlog and to at least begin mulling over any new
questions that might come in, in the meantime...
  

    
Scott's question was, I think, particularly astute. All of this
cataloguing of interface ideas and the tradeoffs between different ways
of displaying people and objects and providing methods for interacting
with and around them may play a secondary role to other factors such
as luck, popularity, timing, and business factors unrelated to user
experience.

That being said, I think at least some of Twitter's success (thus far)
beyond embracing radical constraints, has been its use of "openness"
patterns, in particular providing an API and support for (unpaid)
third-party developers to create applications that extend Twitter's
reach and capabilities and provide users with more sophisticated or
nuanced interfaces.

One issue we sometimes run into with our pattern catalogue is that
some people are completists and view our materials as a giant checklist
and somehow get the impressions that we think an application should
exhibit all of these patterns, which is the opposite of the truth.
Generally speaking, I think fewer features are better, especially when
an application is new. In fact, one theory is that you should launch
with the minimal number of features that still provides some actual
value (and perhaps a promise or suggestion of where it can go).

After that, I think as much attention needs to be paid to what people
are actually using, what's getting traction and what isn't, and what
users seem to be (or are explicitly) asking for. Until there's some
kind of a "hit" or traction, you're kind of just casting about in the
dark, but the moment you have any liveliness at all within your app,
it's time to really let go of the idea that you are going to totally
control things or that some a priori vision you had is about to come
true. Instead, it's better to extend what's working on rethink what
isn't.
  

    
the next couple of questions in the backlog are hairy, with
subquestions embedded, so I'm going to take a little more time to mull
them and try to add more replies later today or this evening.
  

    
I want to toss in one more, which I thought of when reading your
comment about limiting features. You've been evolving this catalog of
patterns within a company that's constantly evolving new interfaces and
interactions, so I assume that your library is getting some use. Can
you describe how developers typically interact with you, Erin, and the
pattern library to determine the right feature set for a particular
application or site?
  

    
got it, Jon. I'll tackle that one in a bit, maybe in the context of
addressing some of Ari's concerns about the extent to which Yahoo!
itself adheres to some of the patterns we recommend or delineate in our
book.

in the mean time, I need to take on "social media" and "social media
marketing."

so first of all I want to note that a lot of the feedback and interest
I've always gotten from this project has been not so much from
practitioners (designers, developers, product strategists,
entrepreneurs) as from "lay people" who are enmeshed in social media
and are building up their own mental models of what's going on and who
want to share anecdotes and figure out why they like what they like and
why they hate what they hate and where it's going and how it could be
better.

I love these conversations, even if they never get anywhere near
helping a designer decide where to reflect reputation achievements on a
user card. They inform the whole process because they reflect behavior
as well as motivation and aspirations (and frustrations) and as
interaction designers we work with that stuff the way a smith works
with metal (and dross).

In fact we sometime distinguish "social behavior patterns" from
"social design patterns" (insert any contemporary synonym for ux or
design in the latter). The former behind what people actually tend to
do or want to do and the latter being more about the UI strategies and
elements that respond to and channel those wishes.

(gonna break this one up a bit)
  

    
next, an aside on terminology. we address this a bit in the book. we
are dealing with some buzzwords here. first of all there's "social."
when I wrote my last book (The Power of Many) the subtitle used the
term "the living web" which was derived from Mark Bernstein's essay in
A List Apart <http://www.alistapart.com/articles/writeliving/>. 

I recall sitting at lunch with my editor on Lakeshore in Oakland
trying to find the perfect phrase for the subtitle to capture what is
now commonly referred to as the "social web." So that was a near miss I
suppose.

Social is the thing today. It's the new black. A lot of the people who
come to our workshops and will probably buy our book are practicing
interface makers who have been challenged (by a boss or client or by
themselves) to "add social" to what they're doing. In the past the
challenge may have been to "add ajax" (or rich interaction) or to "add 
RSS" or to "add frames" even.

Any words stretched to cover too much ground will become thin and
social is at risk of that. There's the whole "everything's social"
observation, as well as "the web's always been social," etc. These
things are at least partly true. I defined earlier in this conversation
how we distinguish social interfaces from what I sometimes call
"solipsistic" user interfaces.

The term "social media" has also been stretched almost to the breaking
point. Frankly, most of the time it's just a euphemism for "the social
web." Sometimes it refers to contexts such as Flickr, where people are
being social and exchanging media (in this case photos and videos).
Other times the word media is being used more in the sense of mass
media and then "social media" refers more to filtering and following
information through your friends rather than by reading USA Today and
watching CNN (the television network, not the collection of Twitter
feeds).

Finally, there is the idea of "social media marketing" which at one
end of the spectrum means replacing old-school PR, marketing, and even
advertising with a cluetrainful, humane, person-centered, permission
based engagement one on one with customers through "social media" (in
this case really meaning blogs, twitter, and other social
applications). At the other rend of the spectrum it really just means a
new form of spam (or at best a kind of cottage industry of village
explainers and local social media "gurus").

Somewhere in the middle there, perhaps a bit closer to the
super-enlightened end of that spectrum, is a real of *using* social
interfaces to explore new modes of commercial and customer-centered
relationships and even *building* social interfaces specifically
designed to foster cluetrain-esque conversational marketing, monitoring
of discussions, and engagement between people representing brands and
people using products.
  

    
We do address that last idea a bit in our chapter called "Social Media
Junkies" unite. Mostly we talk about interfaces for filtering media
socially, for tracking topics and conversations, and for taking social
factors into account to enrich searches.

There's probably a whole 'nother book that could be written pushing
those ideas further and filling in some of the gaps.

And then there are book, such as Tara Hunt's "Whuffie Factor," that
are centered on that idea (as well as perhaps some of Seth Godin's
better stuff).

Meanwhile, the UX community is wrestling with these ideas. I've seen
Chris Fahey, who leads Behavior Design, writing in public about how UX
agencies need to get better at some of the services traditional
advertising agencies excel at (such as, well, dealing with
advertisements for one thing, and monetization in general as part of
the design process - as Brian probed about earlier)
<http://www.graphpaper.com/2008/08-07_ux-of-a-salesman> and .

I've also seen Peter Merholz, who co-leads Adaptive Path blog about
the question "What does the user experience field have to say about
social media?" <http://www.peterme.com/?p=768> suggesting something
similar: that customers want more than just a website rebuilt or a blog
or twitter or facebook account integrated into a homepage. They want a
social media strategy, they want campaigns, they want success metrics
they can understand and use to make decisions about where to allocate
their (yes) media buys, and so on.
 
I've been meaning to write a blog follow up to Peter's post since it
showed up. It's one of the to-do's that's been bouncing down my list
ever since (don't tell Merlin Mann) because I think it hits a really
interesting nexus, it tracks with the topics that generate the most
fervor when I'm speaking in public, and maybe also because a commenter
on that blog singles out another one of our patterns ("Manage Project")
as toothless, which is probably a reasonable critique but which puts
me in that same defensive frame of mind I'm hoping *not* to answer from
when I get to Ari's question(s).

I may not need to write that blog post, though, since the comments on
Peter's post (and the links from it to Chris's blog and to Josh
Porter's post <http://bokardo.com/archives/the-agency-problem/>)
provide pretty solid coverage of the key ideas. There's very little
posturing in that thread and the commenter names read like a who's who
of UX and social design.

Jon, did I get anywhere near to answering your question? I'm going to
back and re-read it now, and then probably take a "work break" before
continuing here.
  

    
ok, the nut graf says, "Is there an understanding implicit in the
patterns overall about how marketing and business should be conducted
in the evolving online social context? Best practices for social media
marketing? Was business use much of a consideration in pulling the
patterns together?"

I think I'd have to say that we do not make a clear statement or
present an agenda for how marketing and business should be conducted.
We touch on it. We tell people to "talk like a person" and so on. There
are myriad assumptions about the sort of web we think we should be
trying to build that underpin our recommendations and caveats, but I
don't think that quite hits what you're asking about.

I definitely know we didn't present best practices for social media
marketing. Again I'd refer to people to Tara's book and her Horse Pig
Cow blog for a treasure trove of insights about that (and there are
other "big names" in social media marketing who have their own
platforms and whose best advice can probably be sifted, particularly if
you use your friends to filter it :').

The last part, about whether business use was much of a consideration,
I'd say it was *a* consideration and that, again perhaps more
implicitly than explicitly, it's always a factor in what I spend my
time on at Yahoo!, but I don't feel that it was a primary rubric or
structuring tool for the book.

ok, gonna do some other stuff now and then try to answer the two or
three facets I read in Ari's questions.
  

    
(I wanted to say in regard to Flickr and Yahoo! that I was being
grumbly, not expecting an actual answer. Sorry about that!  

In reality, all the branding issues about changing the doorway and the
name of the tribe of users, if you will, for commercial reasons, are
sticky and some just have to be dealt with in compromise by users and
companies alike.  So no need to reply to that.  Overall, I love Flickr
and am very glad Yahoo! is allowing the communities there to scale and
persist, in itself not a simple proposition.)  
  

    
thanks, Gail. Ari (and you) raised legitimate question about
registration and sign-in and the challenges of federated accounts and
so on, which I will tackle soon in my copious spare time.
  

    
Here are a couple of things in the "Forums" Pattern that I would have 
developed as anti-patterns (p.293)

* Consider pre seeding topics for discussion on sites that have a specific 
context

In fact, this example forms the core of the "Potemkin Village" 
antipattern. 

Turn the page, and we get to:

* Indicate when a topic is new

This is, again, contra-indicated and would form the basis for an 
anti-pattern. "New" is a funky concept, but one of the harbingers of bad 
forum software is when threads are tagged as "new". 

- It is almost never clear what "new" means
- It is extremely important that the concept of "new" always mean "new to 
the user"--it may be new since her/his last login; it may be a thread that 
he or she has never read, but which now has recent posts. What 
connection to "newness" and the partipant means depends on the 
forum, its community, number of participants, and purpose. All I would say 
as an absolute is that it must be personalized, and can never refer to the 
system, in general.

That's the anti-pattern, at any rate, and there is a lot that could be 
written about how/why "new" as a system announcement inhibits 
participation and connection.

[This doesn't mean that participants shouldn't be able to see "hot" or 
"most recent posts"--just that the concept "new" doesn't work at the 
system level.]

On the pattern side for forums, rather than
* Allow users to follow discussions via RSS

the pattern piece you are feeling for would be something more general: 

+ Provide ways for a participant to "subscribe" to specific threads via 
RSS or email so that she/he is notified when/if there is further activity. 
This should generally be the default setting when someone starts a topic 
or posts to an existing topic/thread.

And, similarly,
* Present messages in a threaded format....

were it to reflect what we have known for the last decade or so, would 
read something like

+ Present the ability to view topic contents in a threaded, summary, or 
linear form, as best suits the needs of the participant....

That could lead to some discussion about where the various forms work best 
(long-term discussions map better to well-style linear, say; help topics 
and short discussions often work better as threads; both make different 
sense when (as is now common) each post can have a title reflecting either 
the current thread/topic or a new branch)....

But, I think the problem with "Forums" is that there is a lot a lot to 
cover--it's not a pattern, per se, but a collection of several interaction 
methods, not all of which fit together, but all of which might be needed 
depending, as I said above, on the community, number of participants, 
purpose of the forum.

They also relate strongly to mailing lists (and in some cases, people 
can/should be able to follow and/or participate in forums using a mailing 
list interface, through email clients).

Right?
  

    
Anyway, I don't know that TDOS says about what they call "Message Boards" 
in 2e, but in the first edition, they cover many of these subjects. It's 
not like reading Amy Jo Kim's book, and what I just posted goes beyond the 
TDOS patterns on the subject, but TDOS does, on this subject, provide an 
excellent nexus of useful information and related patterns. I just felt 
that this was an area (okay, it happens to be one I think a lot about, and 
have for a couple of decades) where DSI feels very partial and very 
tentative - but presents the information as though complete.
  

    
thanks for the further examples. Erin, btw, is set up now with a Well
account but I may need to walk her through the engaged interface. She
was also the lead author on the groups material (we collaborated on
everything but divvied up the initial authoring) so she may also want
to weigh in on these ones as well. 

I'll address your questions above along with the earlier ones wrapped
around the registration context. Should be able to get to it later this
evening.
  

    
Waiting to see the responses to Ari, meanwhile considering Ari's point
that the forum patterns seem partial and tentative. I get the sense as
I discuss social technologies (which is pretty constantly) that forums
are considered, by some, to be old technology, something we've gone
beyond. The platforms du jour are more about social connection than
shared relationship, and it feels like there's less opportunity to form
community than to form crowds. It's late and my mind's not quite
forming the connection to the patterns discussion, but I'm interested
in considering this question: how do we focus more on building and
sustaining high-value reelationships and collaborative communities vs
"social networks" and drive-by conversations? 
  

    
<jonl> brings up an important point. I do think that while we were 
kvelling over tagging and friending and commenting, "forums" went out of 
fashion for a while. They weren't the "new" (to abuse a term I have 
accused ya'll of abusing) thing.

But as I look at my efforts here at JWA, and what I am seeing elsewhere, 
it is slowly sinking in that there are some types of communication for 
which forums are critical. Sadly, current forum software often makes FIDO 
boards look sophisticated--it's like the knowledge fell into a black hole 
and has to be rediscovered.