When #! Goes Wrong

It’s time to have a serious talk about #!.

If you’re a sharp-eyed Web user, it will not have escaped your attention that, for many Web sites (Twitter among them), the characters #! have started to appear in the address bar when visiting certain pages. Try it now. Go to my page on Twitter but check the URL I’m sending you to first – it should be “http://twitter.com/torgo”. Now – when you visit that link, check the address bar at the top of the page. Abracadabra, a mysterious #! (pronounced “hash bang” in geek-parlance, and we are firmly in geek territory here) has interposed itself between the twitter.com and the torgo bits of the URL. The appearance of #! is an artefact of a certain approach to Web application architecture. Many in the Web community have decried this approach (see more detail in Jeni Tennison’s blog entry), but to cut a long story short, the argument against using #! has been painted as largely academic by many Web application developers.

This morning, I woke up and found my (very) local paper, the Archer, had been slipped through my mail slot. Something drew my eye to a box at the bottom of the page. “The Archer is now on twitter,” it pronounces. “Follow us on http://twitter.com/#!/TheArcherN2.”

Ok, #!. Now, it’s personal.

I’m not pointing fingers at the good folks at the Archer, by the way. They just did what Twitter told them to do. In good faith, they copied and pasted the URL that appeared at the top of the page. But surely this result could not be what the people at Twitter intended. So, now everyone who sees this URL and types it in will be forced to type in three more characters than is strictly necessary. And can you imagine someone trying to tell that URL to someone else down a phone line or over the radio? It’s like “H, T, T, P, colon, slash, slash” all over again, except worse! (For one, “exclamation point” has 5 syllables.)

And by the way if the person trying to type in that URL happens to be a Mac user in the UK  then they are going to be doubly confused because there is no # on the Mac UK-English keyboard layout:

Spot the #

From a Web architecture perspective: Jeni’s blog entry goes into great detail on the pros and cons of this approach to keeping application state. In the TAG, Ashok Malhotra is working on a document on Identifying Application State which discusses this issue in detail, discusses and some alternatives to #! and some approaches that Web site developers can employ if they want to use #! and not confuse Web users. If you’re a Web site developer, I urge you to read these.

But thinking of the bigger picture: when you build a Web site or application, you are not building it in a vacuum. Stuff you do, including what appears in the address bar, will have unintended consequences. People doing stuff like passing around URLs out-of-band is part of the Web. So think of the children already and get with the program!

(And to my good friends at the Archer: I love you, and you have done absolutely nothing wrong, but please the next time you print this blurb, print it as  “http://twitter.com/TheArcherN2”. For me. OK? Thanks.)

Example 1: The TechCrunch Europe article on start-up frustrations with BT’s Fibre roll-out. I was trying out an App, Pulse, and ended up reading the original article through this. Pulse is essentially a smart feed reader that downloads and caches articles – one advantage being that you can read them while off-line. (NB: when will someone write an HTML5 “app” that uses local storage to do the same thing?) Anyway, I wanted to share the article using the app’s built in Twitter button. That brought up a pop-up int he app which gave me a link shortened with Pule’s own link shortener (pulsene.ws). I wasn’t too comfortable with this because while I’m happy for the company behind Pulse to know that I’m using their app to send the tweet, I’m not so happy for them to be able to collect information on who’s reading my tweet. Plus I don’t know what that user experience will be when any other random person pulls up that URI. And I have no other option to use a third party shortener.

Example 2: The David Mitchell column in the Observer. This was the most dysfunctional. First of all, I had been reading the article itself in the paper version. I, like many other people in the world, read actual newspapers sometimes printed on real paper. But I find myself increasingly frustrated that there is no easy way to share directly from this medium. In this case, I read the article and I wanted to share it so I got out my phone then went to observer.co.uk (which thankfully does browser sniffing so automatically redirected me to a mobile-friendly home page). Then I had to find the article I had just been reading. I wanted to share right from the article page but no dice. I spent a few minutes looking around for “share me” buttons or menus – nothing (why?). Ok – so I shared directly from the browser menu. That brings up a contextual menu of all the ways you can share something (confusing!) – I chose Tweetdeck. I then had to ask Tweetdeck to shorten the URI (which it did using Bit.ly) so that it would fit (why?). Then I sent out my Tweet, which now had a bit.ly version of a URI to pointing to a mobile-friendly page (which itself does not do browser sniffing). So when another person on a Web browser on a regular PC views that link it shows up as one long column in the middle of their page. Yes – it worked, but totally unsatisfactory. And what about the link to print? Shouldn’t print newspapers start publishing a shortcut, QR code or short link in their print editions to close the gap between print and online? I’m just sayin’.

Example 3: Today, I shared a link from Plancast about the upcoming Federated Social Web Europe event. When I shared from plancast it send me over to Twitter.com which automatically re-shortened the already shortened link with Twitter’s own link shortener. http://planca.st/VRG became http://t.co/RPJrPTN. Well – I didn’t want to use t.co, I wanted to use planca.st. Why? I have this idea that using link shorteners that are controlled by the same organization that produced the original link is slightly less brittle than using third party link shortners. That’s fine – I re-pasted the planca.st link over the t.co link and sent my Tweet. Except, wait! Go to the tweet in question and use the contextual menu (right click) to copy the link and then paste it somewhere else for inspection. Oops! http://planca.st/VRG becomes http://t.co/RPJrPTN again! Thanks, Twitter! Then, to add insult to injury I get the following reply from my friend Ajit (presumably because his Blackberry browser balked at whatever t.co or planca.st or plancast.com decided to sling at him – but who knows at this point?

One more note of frustration: even in putting together this post, I had to manually edit the URIs that Twitter gave me as the canonical pointers to my Twitter posts. Why? Because Twitter insists on using the #! convention and I know that doesn’t work very well in some (especially mobile) browsers whereas if you use the same URI without the #! it works universally. (See Jeni Tennison’s blog for a great post on #! URIs.) Also, I have the “always use secure http” flag set to “on” in my Twitter preferences, meaning that if I copy and paste the URI, that will be the “https” URI – which is also probably not the best one to use in a post.

It’s clear that in this burgeoning era of the Social Web, people are sharing more, and in more ways, than ever before. This is a good thing. But we’ve got to get the user experience right and we’ve got to do it in a way that stands the test of time. The current maze of link shorteners, apps, interfaces, etc… has been great for innovation but it’s not sustainable or scalable. I hope that 2011 brings us some innovation that makes it easier to share (and to understand what’s being shared), in a way that is more in line with how the Web works (open, transparent, scaleable, traceable).

* For those confused by my use of the term URI, this is what Web standards wonks call URLs. Now that I’m a member of the TAG, I am legally obligated to say URI instead of URL. So be it.

photo credit: Charlotte Gilhooly

I’m an early adopter, or possibly a serial alpha tester. I’m always willing to give something new a go, especially when it comes to new ways to get around my city, London. I was first off the block to get an Oyster card – a fantastic innovation that has transformed Tube and Bus travel, in my opinion. I was an early customer of the “OnePulse” combined Oyster-Visa-contactless payment card – less than fantastic, but that’s the subject of another post. So it should come as no surprise that I was one of the first to sign up for the new “Cycle Hire” scheme in London – cheerily called “Boris’s Bikes” by the press. (Us Londoners know they’re really Ken’s bikes but “Ken’s Bikes” suffers from a lack of aliteration so “Boris’s Bikes” it is.)

They probably had enough work to do just launching the service and getting basic e-commerce systems up and running to worry about mobile app development and I’m aso guessing they didn’t have the expertise in house (though that’s just a guess). Many companies and organizations launching new services, particularly in government, might be in similar situations. They could have decided to bag mobile all together, but that would have been shortsighted. Clearly, this is a service that needed a mobile component.  So, as reported in the Guardian, TFL decided not to roll their own mobile app associated with the service but rather opened the field up budding mobile developers. They did so by releasing their data as an API to the developer community and seeing what emerged. And what emerged was a host of mobile applications, some of which have been reported on in the Londonist and CNet UK.

To guide me on my (so-far) three cycle hire journeys, I’ve used the Android Cycle Hire Widget by Little Fluffy Toys. It gives you instant feedback on your home screen on the location, direction and status of the 3 nearest docking stations: invaluable information at the beginning and end of your journey. (I’m also glad to report that we will be featuring a session from Little Fluffy Toys at this year’s Over the Air on how they built that app.)

The main take-away here is that by opening up their data through an API, TFL enabled a market to develop around how to best visualize and package that data for mobile use. And what we’ve seen emerge so far is only the tip of the iceberg. I fully expect to see mashups and other creative uses of that data in the near future.

Photo credit: Rob Pongsajapan.

Last week I had the pleasure of attending the Lift conference and helping to run a workshop on user privacy. This was a workshop with a difference. My colleague Franco Papeschi came up with the idea of a privacy “game” (“Denopticon“) which would help participants explore the issues around privacy, personal information and data sharing. The game started with participants filling out an ID card with personal information about themselves. Participants earned points for finding out and recording personal information from others and additional points for fulfilling various secret missions. It was enormously fun and I hope to help run it again at other events. But besides being fun, it helped the participants, and the moderators, think about the key issues around user privacy.

This was against the backdrop of enormous upheaval in the area of user privacy on the Web. I remember when privacy on the Web used to boil down to “turning off cookies.” Now-a-days if you turn off cookies, you might as well use your computer as a doorstop, and anyway the privacy conversation has so moved on. In a world where more and more of our communication is happening through social networks and socially connected applications, the whole concept of privacy is being turned on its head, to the extent that some (such as Christian Heller) are claiming that we are now living in a “post-privacy” world. And, of course, Google’s Eric Schmidt is on record saying “If you have something that you don’t want anyone to know, maybe you shouldn’t be doing it in the first place,” which (if he truly believes this) I think betrays an almost pathological misapprehension about the human condition.

The truth is, we need privacy, as a society. Anyone who claims we don’t is (forgive me) either terribly naïve, stupid, or a sociopath. Privacy, and a reasonable expectation that some of our actions and communications are and will remain private, is a social lubricant that allows for healthy exercise of denial and other mechanisms that keep us sane. Anti-privacy pundits are quick to reply that “the kids” don’t care about privacy – but this notion doesn’t bear up under the facts. (For evidence, an article from today’s NY Times reporting on the increasing awareness of youth to privacy issues.) In fact, there seems to be a backlash against the “private is public” mentality which has led to over-sharing and social networking fatigue. (Do I really care that you’ve won the medal of the badge of being the mayor of the Duncan Donuts at 33rd and 8th? Maybe that’s something better kept private.)

Over-sharing as promoted by servers like Foursquare may be annoying but it’s basically harmless. You may be opening yourself up to stalking or having your house burgled but that’s a choice you’re willing to make for the benefits that social sharing bring, right? OK, but what happens when you’re not just making that decision for yourself? What happens when your sharing impacts your family, your sexual partner, your children? Emerging usages of  social networks will require more trustable, private environments. With their ability to share structured data, social networks could be a great environment to interact with your stock broker or financial advisor. What about health service communications – such as your blood sugar levels or the results of your AIDS test? What about parent-teacher communications? The list goes on – all of these intrinsically private types of communication could benefit from the rich communication mechanisms that social networks bring to bear. But people would (rightly) be reluctant to use Facebook or other existing social networks in these ways.

Unfortunately, although social platforms like Facebook are adding richer privacy controls, there remain problems both with the implementation of these controls and in making them understandable to regular users. I think Facebook has actually made a lot of progress in making privacy options visible and usabile – at least on their Web site. In fact, my personal trust level of Facebook’s privacy mechanisms has increased enough that I’ve begun sharing family photos and other information with family members on the platform. I’ve been very frustrated by the lack of privacy controls on their mobile clients and mobile web site, but it seems to me they are on the right track. There are challenges on the horizon, though.

One challenge will emerge from the wealth of availability of data that is opening up to Web developers. With a few lines of JavaScript code, a Web application or widget can access your location (via the Geolocation API). Soon, that information will expand to capturing your camera image or digging into your address book. Although browser and web runtime makers are building in privacy controls, are they working and are they the right ones? These are the issues we’ll be exploring at an upcoming W3C workshop I’ll be co-chairing on privacy and device APIs.

Another challenge is going to be implementing trustable privacy in the post-Facebook world. How would my family photos use-case work if my family members were not all on Facebook  but were members of a series of federated social networks? These are some of the problem spaces we’ve been exploring in the W3C Social Web Incubator. The OneSocialWeb project is building an open source platform that uses XMPP to bring some of these ideas to life.

One thing is clear: privacy is becoming a key industry topic and a flashpoint in the intersection between mobile, social and  the Web. The common wisdom is shifting away from the idea that “people don’t care about online privacy” which is good, but it throws a spotlight on the mess that privacy on the Web has become. Cleaning up that mess is going to take some effort.

For a development project I’ve been working on for the past few months (“Agora” – more on that soon), we have largely abandoned this “stand up” style and moved all of this daily status sharing into Google Wave. This was partially necessitated by the fact that the developers and designers we have working on this project are in a few different physical locations, but the results have been surprising.

I was sceptical at first that Wave would be a rich enough environment, but it has really enabled a kind of rich collaboration that.

Developers can post screenshots of UI mockups or code samples or error codes and then generate discussion and get feedback on these. Commenting on the Wave enables people to talk about priorities. And as some of our developers are not native english speakers, using the Wave combined with web translation has enabled them to converse more easily and clearly than via voice or real-time chat.

I approached this with an attitude of “Wave must be good for something” and I can honestly say I am a convert. It has really shown its strength in helping to manage a distributed small-team development project. And while it hasn’t eliminated the need for meetings, it has certainly reduced their frequency, and (I think) boosted productivity.

W3C is organizing an online course to introduce Web developers and designers to W3C’s Mobile Web Best Practices.

In this course you will:

• • * learn about the specific promises and challenges of the mobile platform
  • * learn how to use W3C’s Mobile Web Best Practices to design mobile-friendly Web content and to mobilize existing content
  • * discover the relevant W3C resources for mobile Web design

Participants will have access to lectures and assignments providing hands-on practical experience with using W3C’s mobile Web Best Practices. They will have direct access to W3C experts on this topic who are the instructors for this course. Participants will also be able to discuss and share experiences with their peers who are faced with the challenges of mobile Web design.

More information at

http://www.w3.org/2008/03/MobiWeb101/Overview.html

Register now at
http://www.3gwebtrain.com/moodle/

The week of W3C working group meetings was punctuated by a day-long open workshop which we named W3C Mobile Wednesday. (Yes, my intention is to mobilize every day of the week – already we have had Mobile Monday and Mobile Sunday. Now Wednesday has fallen. Can Thursday be far behind?) Mobile Wednesday was actually a unique opportunity to hear from people working in Korea on the sharp end of the Mobile Web and to do a little bit of a sales job about the work we’ve been doing in the W3C Mobile Web Initiative and why it might be relevant there.

One factor that greatly helped create a feeling of open dialog was the presence of simultaneous translation during the whole event. It’s a luxury I almost never get to experience, but it really can help to facilitate discussion when someone else is worrying about the burden of translation. The translators were a wonder – deftly dealing with sometimes very thick technical discussion, especially during the panel sessions.

Besides Mobile Wednesday, I also had the pleasure of speaking to many Koreans living and breathing the Mobile Web, including representatives of the Mobile Web 2.0 Forum, the Korean W3C office, ETRI, and of the Korean companies involved in W3C activities, such as SK-Telecom and Samsung.

So – what impressions am I left with after this week?

I have more questions than answers, but my overall impressions are that the challenges to the growth of the mobile Web in Korea are similar to the challenges the world over. Perceptions about usage and comparisons to the “real web” are also a problem. I have to respectfully disagree with a statement made by one of the other conference speakers that Korean use of the Mobile Web hasn’t taken off because “Koreans already have very high speed access to the Web at home and at the office.” Yes, Broadband penetration is really high in Korea. However, the use cases for using the Web on the move are different from the use cases for using it in front of a computer. Other speakers at the event highlighted some of these, particular social gaming and one-to-many messaging. Interesting side-note, nobody seemed to know what Twitter was but there are apparently a couple of similar Korean services.

One basic challenge Koreans might have to bringing the Web to the phone is the high use of Flash. Seems that most Korean (PC) Web sites are full of Flash content. Even the photo of me that appeared in the Korean tech news article was embedded in Flash for some reason. The fact that these sites aren’t working on even highly sophisticated mobile browsers is no doubt putting people off the concept of mobilizing the Web.

In any case, it was a very educational an informative week. Special thanks go out to Jonathan Jeon (전종홍) for his role in putting it all together and for posting some great video of the Mobile Wednesday event (see link).

In an hour I’ll be on another flight on my way to Seoul, South Korea. I don’t speak a word of Korean, I have no local currency and I’ve most likely packed the wrong plug adapters. But on Monday morning, I will convene the next face to face meeting of the W3C Mobile Web Best Practices working group. After that, I plan to participate in an event called W3C Mobile Wednesday, a kind of east-meets-west open conference-style event bringing together people working in mobile Web standardization and those working on the sharp end of the mobile Web in Korea: people from manufacturers and operators, yes, but also entrepreneuers, bloggers, developers. It’s all thanks to the Korean Mobile Web 2.0 Forum, ETRI, and the people at the W3C offfice in Korea. I’m very excited about this event and this whole week. Besides making some real progress on the work of the Mobile Web Best Practices group, I hope to get a real flavor for how the mobile Web (and other digital services) are being delivered in Korea, a place that showcases (according to Jim O’Reilly and Tomi Ahonen in their book Digital Korea) the “Convergence of Broadband Internet, 3G Cell Phones, Multiplayer Gaming, Digital TV, Virtual Reality, Electronic Cash, Telematics, Robotics, E-Government and the Intelligent Home”.
That and enjoy some good kimchi.

This week, we have followed up the release of MobileOK with a new document that details exactly that: describe the relationship between Mobile Web Best Practices and Web Content Accessibility Guidelines. Why should you care? If you’re trying to provide a service on the Web, you need to care about both accessibility and mobility. Both of these topics require some investment in skills, tools, and development time, so understanding where the overlaps are should greatly help to reduce development costs and time to market. At the end of the day, it’s also about maximizing the potential audience for your service, regardless of a user’s disability or the device used to access that service.

The speaker line-up was really an all-star cast. I won’t go into too much detail here. I was lucky to be asked to participate on a panel on Mobile-Web convergence (chaired by Ouriel Ohayon) which was notable not only because I was on it, but because it was really well attended. I was a bit worried that the non-sexy name of the panel might put people off but we really packed the room – clearly an indicator that the mainstream “Web” community is eager to hear about mobile. We had a really good discussion on stage as well.

Apart from the panel thee was a lot of mobile interest at the event. French mobile widgets company Goojet won the start-up competition. Echovox’s Zong was out in force as a sponsor and David Marcus, Echovox’s CEO, was a great addition to the mobile panel. Of course, Nokia had a great presence there.

Other highlights?

Martin Varsavsky of Fon – a very entertaining speaker. Much more so than his appearance at Web 2 summit where he was incongruously sandwiched between Google and Verizon.

Jason Calacanis on why blog spam and paid search-engine “optimization” is destroying search in the same way that spam destroyed usenet. I especially enjoyed this presentation because it reinforced my view that trust and identity are going to have to play a role in the future of the Internet if it’s going to remain a useful medium.

Talking to Scott Beaumont and Prashant of Refresh Mobile and hearing about their experiences deploying Mippin. Sounds like things are going really well!

Tom Raftery’s talk on green data centers and power demand management – though it was slightly far a-field for a Web conference. :)

Getting a demo of Mobiluck, a mobile social network. They are doing a great Web-based UI and they “get” what mobile adds to the social networking game. Must do a post on them later.

Meeting Lisa Sounio, CEO of Dopplr. If you’re not on Dopplr, go there and join.

Hanging out with my good friend Rudy De Waele.

Running into Paul Walsh wearing Sam Sethi’s badge.

Unfortunately I missed Philippe Starck’s presentation which I regret because it sounds like it was really good.

Lots more I won’t go into here.

There were a few negatives. Shahram Izadi from Microsoft research presented the Microsoft surface concept but I felt like I didn’t really see anything new here. Janus Friis talked about Joost but I really didn’t feel that he presented it in as good a light as Mike Volpi did at Web 2.0 Summit. Also, he kind of brushed off Paul (BT) Downy’s question about the proprietary nature of Joost (vs. the Web) which I think was a shame. He might have said (as Volpi did) that they are working on a browser-based version.

The Wifi was a problem. Wifi is always a problem at these kinds of events because it’s so easy to underestimate need. At Mobile 2.0 we got Swisscom to come in and build out a custom Wlan which was great and nobody complained. We only had 300 people there though. Le Web 3 (where Swisscom also did the wireless lan) was about 1700 to 1800 people – different kind of scale. And how many of those people were using iPhones as well as laptops? Plenty.

Marc Canter had possibly left his lithium on the plane because he was even more rabid than usual. This had both positive (entertaining!) and negative (shut up and let someone else talk!) aspects.

So. All in all, it was a great way to spend two days. Great speakers. Great networking. Great city. Great work, Loic and team. Looking forward to next year!

Well watch out.

My Yahoo! is back in beta, and it’s about time. What’s new? First of all, there are some outrageous new styles on offer, like this Leopard-print (incongruously named “faux fur” – as a colleague pointed out, why do they have to call it “faux fur” when it’s on a computer?). Unfortunately, many of the styles on offer are not very readable, and readability of text is pretty important in an application like this. It also offers a reader pane built into the application so you can read full blog entries (for those that publish a full feed) right in My Yahoo! This, and the UI tweaks are probably the nicest features. The other bits, such as a page-dominating animated ad (see upper-left-hand-corner) I am finding a little hard to swallow.

Even as a long time, user I was about to give up and move to Google Reader, Netvibes or Pageflakes. The new release is keeping my attention. I’ll definitely give it a spin around the block, but I’m thinking that what they’ve delivered, feature-wise, may be too little too late.

I’m interested to see if I get the same vibe at Future of Mobile tomorrow in London.