Gail Williams (gail) Sat 31 Oct 09 15:19
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.)
Christian Crumlish (xian) Sat 31 Oct 09 18:05
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)
bill braasch (bbraasch) Sat 31 Oct 09 18:42
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.
Ari Davidow (ari) Sun 1 Nov 09 10:20
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.)
Gail (gail) Mon 2 Nov 09 00:15
Happy birthday. Nice recap list above -- it gives a clue as to how rich the area of inquiry and the book is.
Christian Crumlish (xian) Mon 2 Nov 09 08:27
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.
Christian Crumlish (xian) Mon 2 Nov 09 08:32
Jef Poskanzer (jef) Mon 2 Nov 09 08:38
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.
Christian Crumlish (xian) Mon 2 Nov 09 08:48
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.
Christian Crumlish (xian) Mon 2 Nov 09 08:52
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.
Christian Crumlish (xian) Mon 2 Nov 09 08:54
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...
Christian Crumlish (xian) Mon 2 Nov 09 13:45
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.
Christian Crumlish (xian) Mon 2 Nov 09 13:46
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.
Jon Lebkowsky (jonl) Mon 2 Nov 09 15:12
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?
Christian Crumlish (xian) Tue 3 Nov 09 13:00
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)
Christian Crumlish (xian) Tue 3 Nov 09 13:10
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.
Christian Crumlish (xian) Tue 3 Nov 09 13:25
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.
Christian Crumlish (xian) Tue 3 Nov 09 13:30
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.
Gail (gail) Tue 3 Nov 09 14:37
(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.)
Christian Crumlish (xian) Tue 3 Nov 09 15:17
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.
Ari Davidow (ari) Wed 4 Nov 09 17:21
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?
Ari Davidow (ari) Wed 4 Nov 09 17:25
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.
Christian Crumlish (xian) Wed 4 Nov 09 18:43
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.
Jon Lebkowsky (jonl) Wed 4 Nov 09 23:13
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?
Ari Davidow (ari) Thu 5 Nov 09 06:22
<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.
Members: Enter the conference to participate