Well it’s been an interesting week at the O’Reilly Emerging Technology Conference. And now it’s time to wrap things up.

My summary of the experience is that I’m glad that I got to be around people as excited and curious as I am about the development of technology, accessibility of information, and usability of interfaces. The theme was the Attention Economy, and while some of the presentations didn’t really fit, most of them had good insights that I will be taking with me.

I think the real theme was about how to deal with the information overload most of us are experiencing now, especially since RSS feeds came on the scene and exponentially widened the river of information we swim against every day. Attention is definitely the most valuable resource we have. Focusing it by removing the noise from the information stream is the next immediate task. As Linda Stone said, we need to develop applications that are designed to improve the quality of our life, not just make it easier for us to do more.

I took a lot of notes this week, many of which will never be polished enough to make it to this blog, but here are some thoughts from various speakers that I jotted down during the seminar that resonated with me for some reason or another. For the things that I managed to write down exactly as I heard them, I’ll enclose in quotes, otherwise they’re paraphrased ideas from the speakers I attributed.

“What our job is, fundamentally, is getting into the heads of strangers.”

Jesse James Garrett

“Being digital is more important than being electronic.”

Alan Turing, 1947 - quoted by George Dyson

“The item is the thing, not the feed.”
“Complexity breeds consolidation; simplicity doesn’t.”

Eric Lunt

“No moment offline ever goes unpunished.”

Rael Dornfest

If the software doesn’t make it possible for users to defend themselves from the moderators then something is wrong.
We are encoding the ability for freedom of speech and freedom of expression in our tools.

Clay Shirky

“A URL is a promise to update itself in the future.”

Jon Udell

“We are motivated by a desire to be a live node on the network.”

Linda Stone

How to become an expert:

1. Write code
2. Make mistakes
3. Get yelled at
4. Fix code
5. Write about what you fixed

“Write when you’re filled with wonder; waiting till you’re an expert is too late.”
“Burnout is what happens to people who don’t know when to quit.”

Mark Pilgrim

FUSE stands for Find, Use, Share, Expand… Good thing we didn’t go with Find, Use and Connect with Knowledge.

Bradley Horowitz | on Social Media @ Yahoo!

Lists are the cockroaches of design… they’ve been around forever and they’ll never go away.
“Gray on gray is a really bad idea.”
Cool is always easier than good.

Scott Burkun

Notes from the last day of the ETech conference:

Carsten Bormann - PANIC mode

Persistence of Data with AJAX in Networks with Intermediate Connectivity.
This was an interesting presentation in offering considerations that need to be addressed when building AJAX websites that are used in situations where there is not continual connectivity. And the reality is that all network dependant applications should be built along these guidelines because they should be tolerant when connectivity is unexpectedly cut off, or even just extremely slow.

• Make the “backup” path the normal path
• To handle asynchronous updates and concurrency conflicts:
  • Use generation numbers to track/figure out what data is the most recent
  • Use both server based and local based generation numbers to compare against
  • Conflict resolution is up to the application (according to business rules? allow the user to decide? merge or automerge changes?)
• There will be more limitations on storage for keeping multiple versions of data, or adding the conflict resolution metadata

Storing data in cookies means all changes are saved with every request. This is good because no special code needed to send updates to server, but bad because you’re potentially wasting bandwidth sending the same data over requests.

Steve Yan - Web Apps without the Web

Steve’s presentation was an introduction to the open source libraries offered by his company TrimPath that aid in AJAX development and in particular TrimPath Junction which offers a way to handle data storage locally using pure Javascript.

Traditionally you have only 3 storage options for data locally on the client:

1. The Dom:The Dom tree stays intact during File->Save As so it can be used to persist your data.
2. Flash storage: You can save data up to several megabytes using a Flash VM but it requires Flash 8 and pops up a warning at a certian capacity.
3. IE persistence: There are some facilities built into IE now, but have been vastly ignored because they’re not available in Firefox and other popular browser alternatives (in much the same way that XMLHTTP has been ignored)

You must approach storing data in disconnected apps differently than constantly connected apps:

• Track deltas in data
• Do change requests, not changes
• Changes are inserts only (versions of data)
• Unique id generation

Trimpath Junction is an MVC Framework for Javascript designed as WORA (write once run anywhere) and able to run with more than one storage provider.

Trimpath uses TQL which is a Javascript, function based substitution for SQL to manage data. It also uses Javascript templates which are enabled using the With statement, which allows you to change the scope of a call during runtime, and so switch whether you are going against the network or a local store.

Steve was just getting into showing us how Trimpath could be used when he ran out of time. I will be looking into it more though because I found TQL and the Javascript Templates intriguing.

Bradley Horowitz - Social Media at Yahoo!

Bradley spoke about some of the history of how Yahoo! got to where it is today and how their focus has morphed since Google appeared and they’ve aquired social sites like Flickr and Del.icio.us.

Yahoo!’s motto now is to “Enable people to find, use, share and expand (FUSE) all human knowledge”.

The most interesting part of the presentation was their perspective on Flickr. Directly from the slides:

User Generated content

Content not licensed from providers such as Corbis or Getty, but rather contributed by by users

User Organized content

Content is tagged, described, organized, discovered, etc., not by editors but by the users themselves

User Distributed content

Flickr acheieved disturibution across the internet, not through “business deals” per se, but rather throught the Flickr community which distributed Flickr content on 3rd party blogs.

User Developed Functionality

Flickr exposed apis (php, pear, etc) that allowed the community of developers to build against the Flickr platform

Flickr is an…
“Entire ecosystem created by less that ten employees aided by millions in the Flickr community.”

I think mainly this is the lesson learned: If you want to have a successful social site you need to encorporate this level of user participation. Or, at least you should use Flickr as a model if you want to get bought by Yahoo!

Some nice presentations this morning. Here’s some of the stuff I found interesting.

Jon Udell - Attention Focusing Stratagies

To get yourself or your content more attention, there are four strategies

1. heads, decks and leads: make sure that your content is acurately represented by it’s title.
2. active context: provide date, source and subject information in your titles (such as for blog posts) to give the user an idea of what they will find if they click on the link to read the entire article. Clicking is a big investment and they will be more willing to make it if you give them enough information to make a good decision. Some examples he showed as included blog post titles in the format “Post Title | Blog Title | Subject | Date”. Subject is important because it gives the user an idea of relavency and Date gives currency against similar posts.
3. canonical names
4. multimedia story telling

An interesting concept that he proposed is that a URL is a promise to update itself in the future. If it points to a list of items (such as a del.icio.us archive), the list can continue to grow. Sharing your links with people instead of full copy of text or lists is more of a dynamic exchange.

And he demonstrated something I hadn’t seen before: you can bookmark soundbites! Very cool.

Clay Shirky - Patterns for Moderation

Clay’s talk was an analysis of different ways to control how groups participate with authoring and moderating content on collaborative sites.

First was Slashdot. Clay emphasized several times that it is set up to defend readers from the writers. Slashdot doesn’t have “users” in a general sense. They also have a very dedicated small group of moderators who work tirelessly to ensure that the content that appears on the site is worth reading.

There is a problem with group contribution sites, in the “Tragedy of the Commons” where people will try to get attention by being disruptive. Slashdot handles this through design choices and moderation:

1. Move comments to a separate page
2. Treat readers and writers differently
3. Let users rate posts
4. Implement defensive defaults

This can give the moderators a lot of power which could also be abused. They implement another pattern to answer the question “Who will guard the guardians?”:

1. Treat users and members differently
2. Measure good behavior
3. Enlist Commited members
4. judges can’t post

Another example is Bronze Beta, which is a very stripped down and simple group blog. Users are not required to be logged in to post, and the posts are not threaded, but the lack of complexity is intentional and the focus is not on posting articles but on commenting and group participation. The pattern follows:

1. Don’t have complex features (KISS) for posting or reading
2. Make comments central
3. Make Login optional (passwords are only used if you want to protect your “identity”)

Clay’s final example was Writely.com, which takes yet another approach

1. Login is at a group level (so participation is context specific)
2. Participation is by invitation only (gatekeeping)
3. Participation is time limited (project focused)

Clay made a good point: by deciding what moderation patterns to use, we are encoding the ability (or lack thereof) for freedom of speech and freedom of expression in our tools. We should be conscious of these decisions and their effects. If these tools do not make it possible for the users to defend themselves from the moderators then something is wrong.

There’s more information about these ideas on Clay’s wiki or you can join a yahoo group to participate in discussing them.

Kevin Lynch - Flash and AJAX

Kevin perfomed some pretty cool demos showing how to use Flash/Flex to augment or supplant the capabilities of AJAX. Flash movies can remove some of the limitations of using pure AJAX such as small storage (cookies), cross domain processing, lack of direct access to web protocols, etc.

He listed several free tools for development:

• Flash Player
• Flex Framework (you may code in your editor of choice)
• Flex/Ajax Bridge

There’s not much else I can really say other than I am definitely curious about it after seeing the demo and will be shortly checking out the development labs that he mentioned.

Brian Dear - Eventful/EVDB

This presentation was a lot more intresting than I thought it would be.
Brian mentioned two services that I didn’t know they had:

• There is a tool that allows you to “import” your iTunes lists into Eventful and it will build a list for you of all the events related to the artists you listen to.
• There is an AIM bot called Eventfulbot you can ask about events
• You can create a prospective search that will notify you as events are listed that meet your criteria (such as, “tell me whenever Beck comes to Phoenix”).

The most interesting part of the presentation was about the new featureset Eventful released today called Eventful Demand. It is a way for you to make a wish for an event, request it, promote it, and hopefully (if you got enough interest and it’s actually possible) have the wish fulfilled. Once you make a demand request, if you get enough people interested, the performer will be contacted. You can promote your demand through email and blog badges. It’s a cool idea, though it may be hard on a general blog to get enough people in your physical area to vote for your demand and have it reach critical mass.

It’s an exciting idea though, to be able to let performers know where they’re really appreciated and it potentially takes away some of the power from the big bad record companies which is always a good thing.

Hans Peter - Plum

Omigosh this software looks cool. It may just be what I’ve been looking for for the last couple of years or so. Basically it is a storage place for everything, links (and looks like potentially copies of the page), photos, email, addresses as well as a way to reorganize and share that information with other people. It has an open API to develop your own types of data to collect and a floating toolbar widget (called a Plummer) that sits on your desktop and lets you collect information from all your PC applications. I was extremely impressed with the demo and found Hans after the break and literally begged him for an invite to the beta. So far this is the most exciting thing I’ve seen here at ETech, mainly because it seems to solve a personal problem I’ve been struggling with for so long. I can’t wait to try it and I will write up my review. I REALLY want to try Plum!!!

Joel Spolsky (Joel on Software) - 2006 Report Card

Joel gave an entertaining review of several popular “Blue Chip” products and websites. He listed some psycological criteria to judge your product as successful.

1. Does it make people happy
2. Is it (positively) involving emotionally
3. Aesthetics are important (you should obsess over them)

Joel had one of the best quotes so far in the conference while talking about why SUVs make you feel safe because they are full of soft rounded curves but where they fall short of the obvious psychological message implied there: “You can’t put breasts on a sports utility vehicle.”

Yay!

Update: We’re at break so I got to go pick up my prize: two Griffin iPod products. PowerJolt, a cigarette lighter car charger and an iTrip for the iPod mini. So now I have to go get a mini I guess! We just got the iPod Video for Christmas. Always an excuse to spend more money .

Yes, I know it’s Wednesday but I didn’t have my laptop with me on Monday and then left the notebook that I recorded interesting stuff in on Monday in the hotel yesterday. So here we go, a day late. I’ll do this as a series as I get them in.

Designing the Next Generation of Web Apps

(Jesse James Garrett and Jeffery Veen)

Jesse and Jeff spent this session dissecting Jesse’s Elements graph from the bottom up and then from the top down again.

Here’s the bullet points that I picked up:

• There are two ways of approaching the web:
  • Web as information - the web as a giant newspaper, or as a collection of hyperlinks
  • Web as application - the web as a giant database with input/output mechanisms

  These approaches are both valid, and should be considered during all steps in designing applications for the web. Becoming too rooted in one mindset or the other can cause problems in the design.

• There is a developing change in perspective for web designers. It used to be that designing a site was all about complete control of the user’s experience, to the level of pixel positioning. Now successful designs treat “Users as Peers”. Users need to have control over their experience within boundaries set by the scope of what the application is intended to do. Users should feel as if they have creative control in organizing how information is presented to them, at some level in themeing, colors and fonts, but also even in site organization. Designers are creating a container for the user to have an experience in.
• Trust: it takes just 1/20^th of a second for a site to have made an impression on us, either positive or negative. From that point, we are biased and all experience with the site is colored by that impression. So aesthetics are important. Even the most well designed informationally architected site may not be successful if it is not appealing visually.

  Jeff refered to this as “Emotional Design”. If a site is displeasing, then people will not be willing to invest the effort to learn how to use it, but in studies, they’ve noticed that people will keep trying at sites they find visually engaging even if they are not getting the results they want.

  Jeff said that there have been studies that show that we tend to anthropomorphize web sites (or anthing that we’re exposed to that has interactivity). It’s a different psychology from designing products or art. Just as you don’t want to spend time with a person that treats you badly, you don’t want to stay in an abusive relationship with a website.

• Jesse gave his modified version of what AJAX is. He said that his original article (where he coined the term) was not intended for a techinical audience that would take it so literally.
  Really, there are only two things that AJAX is, regardless of what language or protocol is used:
  1. An asynchronous interaction model
  2. Using browser-native technologies
• Web 2.0: What’s changed is not that we have new technologies (AJAX, RSS, etc) to work with, but that we are now enabled by those technologies to re-explore interactional design.
• Jesse: “What our job is, fundamentally, is getting into the heads of strangers.”
• Usability is about training users to use your design as quickly as possible because you don’t have their attention for long (1/20th of a second). This means that you have to be conscious of reusing standards, which is difficult as stadards for Web 2.0 are still being defined. When is the cost of innovation (training users to a new paradigm) worth breaking with convention? This is subjective, but the best answer is “when it makes sense!”
• Usability testing should be used for “checking our thinking” verses as a tool for discovering where innovations need to be made.
• Recoverability is extremely important for todays web designs. Users should not be penalized for exploring. People are cautious of clicking a link because the cost of clicking a wrong link (especially on a mobile device) is often very high, resulting in loss of time or data or both. Making users comfortable that they can’t make an unrecoverable error is crucial. Users need to be allowed to “play” without consequences.

Here are the slides, but it’s a big file so be forwarned!

Besides Ray Ozzie’s Live Clipboard, here are some other things I found interesting at this morning’s presentations.

Amazon’s Mechanical Turk

Felipe Cabrara spoke about MT and some of the applications that it can be and is being used for. There is a programmatic API that developers can use to send out jobs to Mechanical Turk and collect the results.
Here’s the sample pseudo code:

read (photo);
photoContainsHuman = callMechanicalTurk(photo);
if (photoContainsHuman == TRUE){
    acceptPhoto;
}
else {
    rejectPhoto;
}

I hadn’t realized that it was that “simple” before.

The Nathan McFarland of CastingWords.com spoke about how their business was basically built around “outsourcing” to MT and gave the conference goers a url to try 30 minutes of transcription for free.

Previously I wasn’t sure how Mechancial Turk could be used for things that required specialized knowledge, but there is a way you can specify criteria for the people who will be performing your “HITs”. Nathan cited examples of college chemistry students working on a transcription of chemistry lectures for him and so earning money while also educating themselves.

Here’s a list of some of the types of tasks Mechanical Turk can be applied to according to Felipe:

• Identification/categorization
• Keyword generation
• Polling
• Transcription/Translation
• Spelling, grammar checks
• Market research
• Photo Editing/Retouching

Sxip and Sxore

Dick Hardt presented Sxip and the concept of “Identity 2.0″. Sxip is a major sponsor of Etech and I’ve seen the name all over the place here but I wasn’t really very interested until this session. Sxip provides a way for you to store your “persona” information, the various things that define you in your different roles (work, home, hobbies, etc), as well as credentials information on a “homesite”. The homesite can be a site that runs the Sxip software, or there’s a way to use a Firefox extension to make your PC your homesite (and therefore “own” your data). Authentication is done by sites that are Sxip membersites (have the Sxip software installed) and you control what of your personal data is shared. It’s an interesting concept that avoids some of the pitfalls of predecessors like Microsoft’s Passport.

Sxip has an implementation for bloggers in it’s sister site Sxore.com. Signing up and using Sxore is a way to control the quality of comments posted on your blog by both eliminating comment spam and allowing frequent commenters to build trusted identities. It’s an interesting concept and they offer a Wordpress plugin for the functionality (which I will be evaluating soon).

Second Life

A representative from Second Life (who’s name I believe is Cory Ondrejka) had 15 minutes to give us a tour but it was pretty fascinating. I wasn’t prepared to be impressed because TAD has taken a look at Second Life before and didn’t seem to find it very interesting. But after the demo today, I’m definitely curious to try it out myself.

Cory described the way that the users of the game have surprised the developers with the things that they have created (using the tools provided by the game) and also described the unique economy that has been created in both the virtual world of the game and the real world we live in. It’s not a typical MMORPG in that there are no levels, tasks, quests or goals. There’s no subscription model for paying to participate in the game. You pay to make the things you create and the property you aquire “permanent”.

Some facts he shared that I found interesting:

• 35% of adults (over the age of 18) spend more time in virtual worlds than working at a job
• Women and older users demonstrate greater creativity than 18 year old boys
• The median age of a user is 36
• There are more than 100 classes a week held in the game on how to participate in the game (how to craft, buy real estate, etc).
• 17 universities are teaching universitie classes in Second Life
• Tringo, a puzzle game along the lines of Bejeweled and Tetris, was developed by a user in the game and will soon be licensed to Sony as a game for the GameBoy. The Second Life user who created it will receive all royalties.
• A user has made more than 150K US dollars by buying and selling real estate in the game
• Make magazine will soon be creating interactive tutorials in the game so that you can see how projects should be built and interact with the steps

Overall, Second Life is something I think I’m going to have to take a second look at.

Root

The purpose behind Root was presented by Seth Goldstien. Unfortunately it was another 15 minute presentation that just scratched the surface so I am still not 100% convinced I want to track my attention data and store it somewhere, but he brought up some interesting concepts to think about by coining a new (to me) acronym:

PPA

Promise to Pay Attention

PPAs are valuable because if you default on them your reputation suffers. For example: if you don’t show up when you commiteed to watching your kid’s soccer game, you’ve lost some of your reputation as a good parent.

PPAs are a precious commodity because they involve the most scarce and perishable factor of all: your time. Since they are so valuable, according to Seth, they can also be “pooled, securitized and traded”, which means, in other words, you can make money off them if handled properly. Root is setting itself up to be a way to monetize your PPAs.

Linda Stone on Continuous Partial Attention

This speech was something that I (and pretty much everyone else in the room) identified with on a personal level. Linda talked about the information overload that we all experience and how we have tried to cope by trying to pay partial attention to everything but rarely manage to give our full attention to any one thing. We have embraced technologies that allow us to be constantly connected to an ever incoming stream of information and it is drowning us. We don’t feel in control of our lives because we must continually be scanning our information inputs so that we won’t miss any opportunities. And we are in a “constant state of artificial crisis” because of it.

Linda mentioned that in the mid 60s to mid 80s there was a cultural change to a focus from the group to the individual. It became all about “me, me, me” and what’s best for “me”. After that, and up to the current time, we’ve been going through another shift, into building a network, and becoming connected to other people through that network. Though Linda didn’t say it, it seems to me that these connections are all in the virtual space, and that, in the majority of cases, they’re relatively shallow. One example she mentioned was someone being proud of their 3000 friends on Frienster. She said that our current situation means that our actions are mainly “motivated by a desire to be a live node on the network”. The stress of maintaining your connection to the internet can be overwhelming, but cutting yourself off from it is equally as stressful because you might be missing something really or artificially important.

Linda said that the mantra of new software that came out used to be “ease of use”, but with today’s problems all new software should focus on how it “improves the quality of life”. I think on top of this is a need for us to connect in a deeper and more fundamental way with the people in our networks as well. Hopefully some relief is on the way.

More notes tomorrow!

Ray Ozzie (MS CTO) gave an interesting presentation this morning about incorporating the functionality of the standard clipboard into the web. Deemed “Live Clipboard” he showed us examples of how it could be used to transfer structured content on websites between websites and between websites and Windows applications.

The interface itself is a modified text box which looks like a button. If there is structured data on the web page, the user can click the button to copy it to the clipboard. The first example was copying event data from Eventful to a sample website. The “pasted” data showed up already formatted into fields as it was on the Eventful site. It is stored as a modifed version of CFText in the clipboard so that it can be created in the same structure that it was copied from. Transfering the copied event to Outlook was just a matter of pasting it into a new event.

Ray then showed how he could paste profile data from his MySpace blog to his profile on Facebook as a “live link” that would be automatically updated if changes were made to the MySpace profile.

Finally he gave some interesting examples with Flickr: One-click copying and then pasting an image directly into Windows explorer as well as pasting an entire Flickr feed packaged automatically in a subfolder. A special file was created in the feed folder which would allow the feed to be continuously updated as new items are added.

While definitely a cool idea, Ray admitted it will take wide adoption to make it really useful to most people. It requires both the inclusion of a script on participating sites and the use of specialized microformats that the data will be wrapped in.

Ray said Internet Explorer 7 will have some of this functionality already built in when it is released. He has more details about the concept and the microformats necessary to support it on his blog post.

This could be extremely helpful for synchronizing information between sites that you have a presence on, for creating automatic downloads of files as they become available (using the updating RSS feed to Explorer) and just to make it easier to maintain your personal data both online and offline. I can see potential, if there’s enough momentum to actually implement it on a critical mass of websites.