Here’ a nice pic of results from a tab usage survey from Cognitive Daily:
There are some great comments on the post as well.
David Munger summarizes “Only 16.7 percent of respondents said they had more than 10 tabs open. Three-quarters of readers had from 2 to 10 tabs, and most of those had from 2 to 4 tabs open.” The results make sense in a 7+-2 kind of way.
I was impressed with the methodology, given the low cost of online surveying. By asking the question “how many tabs do you have open now?” the survey approaches an ethnological method, while also asking “how many tabs do you usually have open?” The authors used a creative way to get responder age — by recognition of celebrities. This showed that open tabs tend to decrease with age. The Mozilla Spectator project would be a great way to get this type of distribution information, but it’s been pretty quiet on that front lately.
Reactions to the (incorrect) notion that Mozilla might deploy a Google Toolbar style clickstream reporting system were appropriately highly negative. In fact, Spectator was carefully designed to never log actual URLs, but only tab creation and session trails, indicating sequences of page loads and the distribution of sessions across tabs. Pure server-side metrics can’t offer this kind of insight into technographics of internet use or, closer to home, insights for Firefox product design.
The proposed final 2010 data goals focus on data portability, which is certainly a great place to start. Still, the results of this survey remind me of the potential of Spectator to help design systems, like tabs, to manage the pervasiveness of the browser in our day to day lives.
In the meantime, hats off to Dave Munger @ Cognitive Daily & Research Blogging for a useful methodology and interesting survey. Update:I’ve requested raw data for more digging.
I’ve been having fun with CSS Transforms (supported by both Webkit/Safari/iPhone and Firefox) as well as the Webkit/iPhone only transform animations. The resulting framerates on the iPhone are hard to achieve any other way. I’ve even built an iPhone app called iBlipper with this tech.
This is not a major hack, but scratches an itch I’ve had for years to be able to code typographical animation with HTML and an augmented set of attributes — in this case: effect, duration, and start. Check out the iBlipper blog for the demo (plus video) that originated this code, a happy holidays medley.
The application I use the most on the iPhone is Stanza, a book reader with integrated browse+download from public repositories of CC licensed works as well as out-of-copyright classics (notably Feedbooks & Project Gutenberg).
Among the newer reads available freely is the mind twisting Accelerando. It starts out simply enough in a world with not just reality augmentation, but integrated cognitive amplification. From there, it gets really wacky. Highly recommended.
Of course, I’ve also been reading the few remaining undigested Cory Doctorow titles and enjoyed Rudy Rucker’s Post Singular.
The iPhone isn’t the perfect reading device — I find my natural head on pillow and corresponding hold of the iPhone tricks the accelerometer requiring a slight posture adjustment. It’s better than almost any other device of it’s size however, due to the high DPI of the iPhone screen.
Back in the mid to late 90s, a slew of business was done porting “green screen” applications to web apps. These old text based menu UIs had a steep learning curve but supported high levels of mastery. A user with 5 years experience in a system could execute a half dozen keypresses and move through a set of menu choices in an order of magnitude less time than a novice user. The port of these applications to the web dramatically increased the learnability, friendliness, and approachability of the systems, but had disastrous effects on overall productivity for experienced users.
With Oracle and SAP being early adopters of the Ubiquity command line interface (CLI) as an opportunity to extend web applications, I’m confident that the CLI is actually a frontier for enterprise web applications.
We’ve released an initial command set at VersionOne using our REST api and the jQuery functions in Ubiquity. We have noun types for our core vocabulary. This has a bit of a wrinkle as the product can be configured for Scrum, DSDM, or the Rational RUP meets agile AUP. Our find and checkmail commands have rich previews re-using our iconography. GoTo has a custom nountype to auto-complete to page names. I’ve named commands with the “v1″ prefix, pulling us out of contention with most other commands.
I’m just starting to scratch the surface of what it means to design a good command line vocabulary. Equally critical is working with the Ubiquity team to improve the overall UX. I submitted a patch for some of the changes I recommended in the last post and have teamed up with a Mozilla intern doing tests on Ubiquity.
The work on pie-menus for the ubiquity context menu is very cool. I’ve been a fan of this since before I did the first marking menu Mozilla implementation in ‘01, but I actually place more value in selection independent operations, perhaps using the current page or automated extraction of information from a page. The context menu (of whatever) form is great for selection based operations, but in our enterprise use case, it seems more valuable to craft dynamic nouns. Text selection, and subsequent deep menu navigation, is always going to have a higher raw cost in time and effort than a well crafted command line with fluent keyboard driven auto-complete and selection from suggestions.
With the upcoming release of a Ubiquity (Firefox CLI addon) command set for the VersionOne software application, I’ve had the opportunity to take 3 users into the usability/eye-tracking lab to try out ubiquity and our command set. Alas, Ubiquity escaped capture by my screen recording software, so no pretty pictures!
I’ll summarize my observations here on Ubiquity proper with the caveat that in none of the 3 cases did I allow the user to go through the tutorial or engage deeply with the documentation. Note, only one user really wanted to!
In the spirit of Rapid Iterative Testing & Evaluation (the RITE methodology), I’m using my personal judgment to triage the issues to the ones that I think are likely to happen to many users. In theory, one could bring another dozen people into the lab in order to have more statistical confidence in the quantity of affected users, but the agile approach for an early stage software product like Ubiquity is to work on fixing these issues immediately rather than dedicate more time to assessing severity.
Observations
The about:ubiquity start page is a weak quick start
While the goal of the about:ubiquity page is not exclusively to support new users, the display of this page after an install makes this a core use case. If it fails for this, none of the other info will make a difference! The tutorial is quite good, but requiring a tutorial before demonstrating value is a aggressive technique.
For two of three users, I had them start on this page. Both required a good bit of effort to determine how to make Ubiquity happen. Neither user was able to map from “Change Hot Key” to the information they needed: “the key sequence I need to use is shown here”.
Recommendations: Rewrite headings of Users and Developers to “Using Ubiquity” or “For Users”. Provide a short statement of user value, “A command line for your browser”. Edit the labeling of the hot key input box to read: Invoke Ubiquity by pressing: [ input box ] Click in box to change your hotkey beneath the box.
Update: This change has been made and checked in! (sans intro value statement)
Hitting Enter is Irresistible
For all 3 users involved in the test, it took explicit instruction to get them to pause long enough to see the preview pane. While there’s definite value in Ubiquity as a quick way to jump to a page, the value of the inline preview is likely to be under appreciated by new users.
Recommendations: This one is tougher, but expanding the the preview pane on recognition of a preview enabled command or offering user feedback that data is being fetched would be helpful. I would argue this should be core to Ubiquity and not require each author to implement.
Inconsistent Keyboard Support Guarantees User Error
While this one is clearly on the Ubiquity roadmap, the fact that command and noun suggestions support up/down arrow navigation while previews don’t have a way to invoke the standard actions via arrow nav creates guaranteed user frustration. The worst is when the user first tries to keyboard nav for result previews and thinks no keyboard support is offered at all.
Tab completion of commands was not discovered by two of the three participants. In one especially painful scenario, the user tried to copy the command from the preview list, made a typo, and lost the model for the command name.
Recommmendations: Tab completion is a standard characteristic of command line interfaces and likely not something that would affect users with strong experience using other CLIs. This one may have to be one of the goodies a new user picks up from the docs and tutorial.
Wrapup
Despite these challenges, all 3 users walked away thinking Ubiquity was a valuable offering. Showing them functions on text selections would likely increase that estimation.
We’ll be releasing our Ubiquity command set for VersionOne shortly. Special thanks to Agile software development luminary Jeff Sutherland for participating in the test (pictured to the right).
It’s been a quiet few months on SurfMind as I transition to a UX role at VersionOne, a vendor of software process management tools for Agile methods. I’m still working on the Scrutinizer vision simulator and teaching with the Stompers, but I’m excited to be focusing on enterprise application design.
Software project management involves a wide range of technical experience in it’s user base. To offer power users some additional ways to optimize their use of the toolset, I’ve been working on a Ubiquity vocabulary to navigate, create, and find.
I shared some of my experience at BarCamp Atlanta last weekend.
A new release (0.1.2) is due any moment now with serious enhancements to both the developer experience (e.g. the improved command editor) and the user experience (e.g. better multi word support, order independence in modifiers).
I’ve been imagining a command line for working with places data, given the Labs Ubiquity project. I wrapped up the chats in a zooming presentation, via beta service ZuiPrezi, that walks through a chat with Dietrich on the nitty gritty. Click through for an impressive zooming UI experience with some unique use of rotation.
Mozilla Labs has issued a Call for Participation for “Ideas, Mockups or Prototypes”. Alas, the post seemed to follow Techcrunch’s coverage and thus the one example versus the CFP is getting the buzz on techmeme.
Adaptive Path steps up to the plate in the open source design arena with a compelling video. This is fully rendered scenario, from actors & setting to full screen capture of a rich interaction.
In this excerpt, we see a user working through a view of her browsing history by time to help a friend locate something.
In the second excerpt, an automated content clustering algorithm is conveyed in the organization of items on the x/y axis, where z is time.
Can Open Source UX Beat Commercial Alternatives?
It’s hardly fair to lump Firefox into the average open source UX project experience, given funding and a dedicated UX team and research lab, but MPT’s list of problems with open source usability stem from a long history of involvement in the Mozilla project. This effort by Mozilla labs seems to address several of Matthew’s top issues.
Numbers 3 & 5 are obviously addressed just by the CFP and Mozilla Lab’s existence.
3. Design suggestions often aren’t invited or welcomed.
5. Coding before design
Note, I’m really impressed increased upfront design that the Mozilla team is putting forth in the latest iterations, thanks to MikeB & a stellar UX team. In contrast to the original Firefox UI which was (skillfully) developer driven.
7. Chasing tail-lights.
While there is some really outstanding prior art along all of the dimensions shown in this video, it’s awesome to have a high profile open source project setting the bar against fierce, and highly business constrained, efforts by Apple (Safari) and Microsoft (IE).
The video included in the CFP that isn’t excerpted here, Aza Raskin’s mobile (Z)UI concept, is the most unique in terms of lack of historical precedent and sheer user interface engineering in combining zooming/panning with UI control access. That’s especially welcome, but there’s a lot of accumulated wisdom in the prior art on hypermedia that has yet to be brought to consumers. The ZuiPrezi team seems to share that same conclusion.
This blog’s archives are a veritable run-on sentence on the need and opportunities to enrich browser history. One that I haven’t dug up in a while is the MSR prototype “Data Mountain” which combined zIndex stacks of thumbnails with a topographical landscape model and innovated on the gestures for re-organizing. I had a history based RSS mashup running called “blog mountain” back in ‘04 before I joined the devil
12. Design is high-bandwidth, the Net is low-bandwidth.
There may be a paradigm here for high fidelity concept videos, especially to garner the contributions of top design firms.
Yes, as much as I appreciate the videos, the technology-hacker/designer in me wants to find a satisfying incremental gain using the “zone of promixal development” within the current Mozilla technology.
It’s also challenging to pull out the real defining points from a video mockup. The Adaptive Path Aurora video packs a lot in for example. Collaboration, visual revisitation support, and some serious machine learning and statistical data mining tech.
Getting to Prototype Implementations
Functioning software in the history visualization space is much more feasible in Firefox 3 than ever before. The new places datastore and easy screen capture may create performance bottlenecks for a aspiring developer, but much of this is realizable at the prototype stage.
Screenshots, however, are not enough. We need to be able to extract assets, text or image or whatever, from browsing history to create better memory cues and representations. Imagine being shown the button images that you clicked on as well as the thumbnail. That would better differentiate the site you considered buying from versus the one you did buy from in a historical browsing session. I’ve filed a bug to add cache access support to the FUEL.js browser library.
One of the CFP commenters offers up this set of sketches of a trails implementation. I’m a big fan of the trails notion, and it address collaboration as well as revisitation support. The real magic is not just in the authoring, though Dgray@Xplane captures some of the illusive extracting and subsequent mashup that is missing from common authoring tools. A playback component is needed, ideally along with a standards based representation format, ala microformats?
Designing richly interactive UI controls for the iPhone and the web is an interesting challenge and much more feasible now that touch events are exposed in Safari for javascript hackery. Check outsomeexamples.
On the iphone, you don’t havemousemove. Not much sooner than I had inked this draft, I discovered TLRobinson’s library that ports multitouch events to mouse events. This is one part of the equation, providing “down level” experience for user input in multi-touch designs. For the reverse, touch events can proxy mousedown and up.
With improved javascript perfomance, and toolsets like Aptana, looks like web apps for the iphone are going to be fun. Alas, Canvas is lame compared to SVG compound DOM for direct manipulation UI.
Following up on my first generation video on how to podcast with the iPhone, here’s a look at the 2.0 appstore options for recording, with an eye to podcasts.
The iPhone SDK doesn’t apparently make it easy to record and transfer audio files, so to make a long story short, jailbreaking and then SSH’ing to pull your podcasts is still the only way you’ll get full fluency with big files.
Transfer limited recorders
A bunch of apps haven’t figured out how to get their files off the iphone:
YouNote (free): email is coming soon, but this is picture, text, and audio note taker & manager. It’s likely too clunky for on the go recording, requiring immediate naming of audio recordings. Nifty free app nontheless with note geotagging.
SpeakEasy Video Recorder: audio levels, playback scrubber, unique features include resumption after a call. Attach a pic. Organize in categories.
Of course, the no-transfer restriction is likely easily solved by Jailbreaking and accessing the iPhone filesystem directly.
Recorders with Email Transfer
Some contenders that do have the ability to transfer files include:
Note 2 Self $4.99: interesting customization and unique controls (trigger recording by moving phone to ear). Email compressed or uncompressed.
Audio Recorder $2.99: Add to a recording, rename, email to self or contact, audio levels, playback scrubber. One of the slicker looking apps in the category. CAF core audio format, making email to contact somewhat wacky.
Hybrid web/iphone/operation system applications
Evernote: Evernote gives 40mb of space per month free, with a subscription yielding 500mb at $5/mon. Audio files are auto-synched to the server in .wav format. While there’s no download button, the file is easy to find and there’s an easy greasemonkey fix. Note, no longer windows only! There’s a Leopard client now but audio files are stored in a fairly obfuscated library folder in a sequence of .reco files.
Reqall: An integrated note consolidation web app like Evernote with the unique feature of transcribing audio.
So for short clips, typically under 10mb, you can get away with one of the contenders and email. With most email servers blocking attachments over 10mb, for serious interviews at a conference or routine meeting recordings, this solution is inadequate. Alternatively, an Evernote subscription might work for moderate monthly volume.
The right way to do this is synch recordings with iTunes in a special category.Given privacy concerns & legal restrictions, it makes sense to try and prevent the surreptious recording of audio calls, but hiding the music library from 3rd party apps is a user experience crippling profit inspired outcome.
I haven’t pwn’ed my freshly replaced iPhone 2.0 yet (old one had a broken screen), but is sure is tempting.
I’m a bit fan of tag clouds, yet sense there’s much more to be done. Marti Hearst’s recent work on tag clouds challenges their usefulness as a true visualization or UI, while other acknowledging benefits. More on that some other time…
Here’s a sampling of the best / most-representative stuff I found in a quick run through of the search startup SearchMe’s results… displayed in their new Stacks widget.
About 80% of them are from a half dozen queries I ran. The other 20% (mainly research articles) I had to manually add. The widget is quite slick, though I wish there were a few more items onscreen at one time. I’ve already asked the SearchMe folks when the API will be available for creating “stacks”. I’d love to hook this up to my delicious feed or build a widget to auto-create a stack for all of the links in a blog post.
Rich document previews for the web are here! I can’t help but wish for more — but most of the best ideas of HCI over the last 20 years are yet to be deployed to the masses. Hats off to the folks at SearchMe for the acumen and perserverance to get this feature to market.
It’s been a long time since my hyperactive extension building days of ‘01-02. My first extension to make out of basic prototyping stage in 6 years is now available, timed to coincide with the new Firefox 3: StomperNet Ranker.
The tool features:
An interactive SVG-based visualization of result overlap in the top 20 results for Google, Yahoo, and Live Search
“Brushing” feedback highlighting same domain results and displaying a info panel
Click to preview the result page
Tabs for the full result views from each engine
Access to the 7 search markets
It also integrates the Scrutinizer browser, which simulates foveal and peripheral vision for design inspection and more informative observation of use, with a “Scrutinize This Page” option.
We used the Komodo IDE (howto pdf) to build it. It makes getting started pretty simple and I roped in two members of our StomperNet dev team who had no experience with extension development more easily because of it.
Parsing search results with the DOM is a breeze compared to using regular expressions, no matter how good a PERL hacker you are. The tool also features interactive SVG, sharing CSS between the XUL and SVG layers. I’ve long been a proponent of SVG in Firefox, notably in my MS Thesis on menu mousing behavior. It’s an under appreciated part of the Firefox UI toolkit.
Note: Updated from original post for improved accuracy.
Made a bit of progress tonight in using places to assess my personal success and use of the big 3 engines. I’m not quite ready to share the code, but using Firefox places sqlite database and the google visualization API I came up with the following:
Sessions
Clicks
Engine
ClicksPerSession
705
2340
.www.google.com
3.32
39
38
.search.live.com
0.97
62
54
.search.yahoo.com
0.87
One of the subleties of Google’s brand is the intense degree of trust users place in their results. This not only leads to a perceived advantage in quality, but also to a greater level of trust. This results in users returning to Google for another click as opposed to spending more time on the site they land on or one it links to.
My sql for this is:
select count(distinct s.session) as N, count(distinct s2.from_visit) as clicks, rev_host from moz_historyvisits s, moz_places p, moz_historyvisits s2 where s.place_id = p.id and s2.from_visit = s.id and( rev_host like ‘%moc.elgoog.www%’ or rev_host like ‘%moc.hcraes.nsm%’ or rev_host like ‘%moc.evil.hcraes%’ or rev_host = ‘moc.oohay.hcraes.’) and (p.url like ‘%q=%’ or p.url like ‘%p=%’) group by rev_host
If we exclude queries with no clicks, we get
Sessions
Clicks
Engine
ClicksPerSession
606
2340
.www.google.com
3.86
25
38
.search.live.com
1.52
40
54
.search.yahoo.com
1.35
My sql for this is:
select count(distinct s.session) as N, count(distinct s2.from_visit) as clicks, rev_host from moz_historyvisits s join moz_places p ON s.place_id = p.id LEFT OUTER JOIN moz_historyvisits s2 ON s2.from_visit = s.id where ( rev_host like ‘%moc.elgoog.www%’ or rev_host like ‘%moc.hcraes.nsm%’ or rev_host like ‘%moc.evil.hcraes%’ or rev_host = ‘moc.oohay.hcraes.’) and (p.url like ‘%q=%’ or p.url like ‘%p=%’) group by rev_host
So session abandonment is:
Google
Live
Yahoo
14%
36%
36%
Another critical stat is the frequency of refinement activities. Looking at sequences, I see 36% of sessions involving query to query chains on Google, 8% on live, and and 40% on Yahoo (sql).
While my search behavior is rather abnormal, both through extreme geekitude and SEO activities, my goal with this work is to find a way for many users to contribute this data to form an independent analysis of search experience.
It’s been a long time since the Mozilla 1.0 party in Boston, but I’m happy to say that the Mozilla manifesto animations I did back in 2001 still work today, even on the iPhone. The beauty of standards… one and two.
I’ve been excitedabout the new history and bookmarks system, Places, in Firefox 3 for a long while now. With the release impending, and some solid usage behind me, I can now look at some key metrics that the industry has been tracking in isolated research activities.
What percent of browser activity is spent on re-accessing information & sites versus consuming new experiences? A WWW’06 study by Harald Weinrich, et al. paper updated these stats, showing across studies 61% of pages visited were repeats in ‘94, 58% in ‘96, and only 45.6% in their 2006 study.
The nature of the effort executed in getting to the location is also predictive. This too is present in the Awesome Bar algorithm. From D.M.O The_Places_frecency_algorithm (more):
For the 10 most recent visits (where 10 is determined by places.frecency.numVisits):
.. Determine percentage bonus for type of visit (ie: the “transition type”):
…. 0 (places.frecency.embedVisitBonus)
…. 120 (places.frecency.linkVisitBonus)
…. 200 (places.frecency.typedVisitBonus)
…. 140 (places.frecency.bookmarkVisitBonus)
…. 0 (places.frecency.downloadVisitBonus)
…. 0 (places.frecency.permRedirectVisitBonus)
…. 0 (places.frecency.tempRedirectVisitBonus)
…. 0 (places.frecency.defaultVisitBonus)
It’s also very cool that you’re rewarded by using your bookmarks with serendipitious improvement of the location bar. In short, there’s a heck of a lot of personal magic in this dataset, as well as potential insights into how humans interact with the web.
So that’s 42% of my page views are revisits — given an increasing usage of FF3 as my primary browser over 3 months.
Now it’s entirely possible the selective pruning for history is making these numbers not quite right. Comments pointing to how the places db is pruned with time appreciated. This analysis should also be informed by a more intelligent filter of visit types. In any event, this is just the tip of the iceberg on both the research and feature fuel that can be garnered from the Places db and the rich API around it. Drop me a line or comment if you’d like my SAGE “workbooks”.
Regarding large scale data, in the vein of ComScore metrics, Mozilla.org is very concerned about the trust user’s have placed in them. Perhaps it’s actually appropriate for the market to determine what works within the open ecosystem of Firefox. In any event, I expect lots of interesting add-on work utilizing places and increasing user research enabled, data driven innovation from Mozilla.
The mozdevguys gave me a shoutout as they planned the tagging system for mozdev projects. We agreed we were using tags to provide a browsing UI, not facilitate search, meaning that no aspect that would never be shared across 2 or more projects was a good candidate for a tag.
I proposed two key semantic guidelines for tags:
Tags should address the application component (bookmarks, attachments, history, etc)
Tags should address the user goal/task (find,share,refind,search,forwarding)
This is pretty much what’s happening, with the addition of software product name identifiers. I’ve used the nifty kwout service to capture a clickable screenshot. Copying partial source doesn’t get the CSS on the tag cloud.
I was also interested in a architectural level of tagging along the lines of a classification of wishlist bugs.
feature addition (features)
feature change (modification)
feature tweak (enhancement)
application (app)
There are lots of interestingenhancements happening over at Mozdev. The tagging system offers a much richer way to browser extensions at mozdev (needs more data tho!), and offers a nice organic alternative to the perhaps overly rigid addons.mozilla.org.
The latest in deep cogitation on tagClouds is from the IA Summit ‘08 (B&A podcasts). While the research found utilization of online video and customization high, only 35% of users reporting clicking on tagclouds. I think the jury’s out on tagclouds as a whole — I value them a lot for providing an aesthetic overview of a topic space.
The version of wordPress powering this site was massively out of date (v2.0) and I paid the price. Two spam infilitrations inserted links into older posts and created a whole directory of spam content.
The upgrade was smooth and mostly painless… as Dougal says “Upgrade or Else“. Indeed, while most of the spam was non-destructive, a handful of older posts seem to have been chopped.
The Web Search Data Mining conference held it’s inaugural meeting recently. This is a spinoff of SIGIR, WWW, and related conferences specifically dedicated to learning from user activity traces & web topology.
While I keep informed and occasionally learn new things from Ted talks, UIE podcasts, etc., that doesn’t compare to the density & sophistication of attending a research conference. In a great sign of the times, many of the WSDM talks are online at videoLectures.net.
As we wrote about in the PIKII paper (full text), the web has evolved to partially support most of the key tenets of traditional hypertext. A notable exception is trails, or maps through information space, that can be shared, followed, and augmented. For a indepth primer on the history here, see the web that wasn’t. Let’s look historically at efforts in this space.
Trailfire, a web 2.0 company, has a Firefox plugin that facilitates authoring and a playback mechanism with the claim of personalized topical recommendations. Trexy was an earlier foray that focuses on supporting search and uses search queries as starting points for trails. This type of parallel, collaborative search of an information space was most recently addressed with Microsoft’s Search Together prototype, scheduled for release soon.
The most interesting recent entry into the trails implementation space is PMOG. This pseudo role-playing game with it’s passive mode of operation and economy around user attention and web browsing may have realized the concept of trails more fully than anyone else at this stage. The notion of “passively multiplayer online games“, and especially pmog.com got some buzz this year at SXSW, fueled by a panel and strong attendance by the gaming community.
While PMOG has a lot going for it beyond just the notion of trails, let’s consider the history of popularity of trail-based systems with a graph from Alexa (yes, dubious source I know). The winner here is Walden Paths, an academic project that produces a useful tool for teaching with websites - while past it’s peek it’s still a defining project.
Alexa shows PMOG rising fast, Trexy with a small hayday years ago, Trailfire waning, and Waldens Paths also waning but with more reach than all the others combined.
Are trails a good idea? With the popularity of social bookmarking sites, my guess is yes. There is another precedent — third voice, which left annotations scattered around the web. The richer and more contiguous information provided by good trails could go beyond the random web graffiti that third voice generated.
Authoring
The annotations offered by the trail services described here are typically attached at the page level, not the page element level as the most robust specification of annotation requires. Even with simple page level annotations, PMOG and Trailfire take very different approaches to authoring.
PMOG uses a 2 step authoring process, with step one being light posts and step two stitching them together into a mission. Trailfire has a more robust interface while browsing allowing annotations & sequencing to happen during the browsing activity.
I like that PMOG allows me to gel trails without a massive interruption in the browsing flow. No metadata is specified when you add the trail. That said, the UI for adding marked urls to specific trails (or missions in PMOG lingo) is clunky. This is partially due to the number of marked urls being tied to the currency of the land, so the expectation is that the set of marked URLs will remain relatively small. I crafted a trail/mission with PMOG as I did the research for this post, the buzz on PMOG, 3/15/07.
Playback Subtleties
The area of playback is also a key experience delta between Trailfire and PMOG. Partially due to the recasting of trails as missions, the contents of the trail are hidden from the user prior to playback in PMOG.
Some placeholder ideas for future playback systems:
Branching: The natural solution for this is a visualization, with additional support for backtracking both from the subtrail and to the subtrail.
Discovery of overlapping trails: either augmentations or unaffiliated. Trailfire does this aggressively, all the time when it’s active based upon current url.
Additionally, if annotations were attached to elements, there’s are some new challenges about helping the user orient to the site before deep diving to content. Due to cross-site security restrictions and the general weakness of frames, playback is likely to always require a browser mod.
Next?
I believe the key missing piece here for mass adoption is the ability to publish trails to a blog. Imagine a microformat for trails, with creation tools and blog posting integration. Ideally, this could support multiple representations (overview map, episodic rendering with visual previews). Existing efforts have added widgets, but the topical focus of trails makes this type of global include detrimental.
This is the blog of Andy Edmonds, HCI professional, web hacker, and cognitive science enthusiast.
Check out my agile software develpment / user experience blog at UXagile.com. The opinions here are expressly mine.
