May 3, 2012

Need to re-invent the Web (badly)? There’s an App for that!

The Mobile App Trap

The App Trap: Why have just one Web App when you can have hundreds of mobile Apps? A selection of popular Android apps from Google Play, also available for iPad and iPhone from the Apple App Store

I love the convenience of mobile applications but hate the way they re-invent the wheel and are killing the Web. What can be done about it?

I’m in love with the mobile Web

I’ve been smitten with the Web since first venturing out on the information superhighway back in the nineties. This love affair is taken to a new level with the advent of the mobile Web. As an incurable information junkie, having access to news is on the move is great. Using location based services like Google Maps is fantastic, on foot, bike or in the car. I love nerdily scanning barcodes to read Amazon book reviews while browsing the shelves in bookshops, much to Tim Waterstone’s annoyance. And it can be great to have wikipedia in your pocket to settle arguments down the pub.

I hate the mobile Web too

But there’s a big problem with all this appy clappy mobile fun, it’s killing the Web through fragmentation, both for producers and consumers of information. Let me explain.

One of the great things about the Web is that you there is one app to rule them all; a “killer app” called a Web browser. There are several flavours, but they all basically do the same thing using similar technology: they let you surf the Web. One software application (a browser), gives you access to an almost infinite number of Web applications. Wonderfully simple, wonderfully powerful – we’ve got so used to it we sometimes take it for granted.

Now compare this to the mobile Web where each page you visit on a mobile suggests that you download an app to read it. Where there used to be just one application, now there are thousands of glorified “me too” Web browsers apps many of which have re-invented the Web, badly.

Consider the applications in the table below and illustrated on the right. They are all accessible from a Web browser on one of the “four screens ”:  desktop, mobile, tablet and smart-TV:

Native mobile app Purpose Web app
Amazon mobile Online retailer Amazon.com
BBC News mobile News and propaganda news.bbc.co.uk
The Economist mobile More news and propaganda economist.com
eBay mobile online garage sale ebay.com
Flickr mobile photo sharing flickr.com
Guardian mobile Even more news and propaganda guardian.co.uk
Google Reader mobile Feed reader reader.google.com
Google Maps mobile Maps and navigation maps.google.com
MetOffice mobile UK Weather metoffice.gov.uk
PostOffice mobile Postcode / Address finder royalmail.com/postcode-finder
Google Search mobile Search engine google.com
Google Translate mobile Language translator translate.google.com
Twitter mobile Entertaining time-wasting application twitter.com
Wikipedia mobile Encyclopædia en.wikipedia.org/wiki
WordPress mobile Blogging tool wordpress.com
YouTube mobile Videos youtube.com

As you can see, users are encouraged to download, install, understand and maintain sixteen different apps to enjoy this small part of the mobile Web. And this is just the tip of the iceberg, there’s bucket-loads more apps like this in Google Play and the App Store. As a user, you could just use a mobile Web browser on your phone, but you’ll be discouraged from doing so. We’ll return to this later.

Producers and consumers both suffer

As well as being a pain for users who have to manage hundreds of apps on their phones and tablets, the pain is magnified for producers of data too. Instead of designing, building and maintaining one Web application to work across a range of different screens (a challenging but not impossible task), many have chosen to develop lots of different apps. Take twitter for example, in addition to the desktop and Web apps, twitter currently makes no fewer than five different applications just for tablets and phones:

    1. twitter.com/download/ipad (for iPad)
    2. twitter.com/download/blackberry (for Blackberry)
    3. twitter.com/download/wp7 (for Windows phones)
    4. twitter.com/download/android (for Android)
    5. twitter.com/download/iphone (for iPhones)

So a challenging task of delivering content onto a range of different devices has now been transformed into an almost impossible task of building and managing many different apps. It’s not just Twitter, Inc. that chooses to play this game. Potentially any company or organisation putting data on the mobile Web might consider doing this by employing an army of android, blackberry, iPhone and windows developers on top of the existing Web developers already on the payroll. That’s good news for software engineers, but bad news for the organisations that have to pay them. Managing all this complexity isn’t cheap.

Not Appy: How do we get out of this mess?

In the rush to get mobile, many seem to have forgotten why the Web is so successful and turned their back on it. We’ve re-invented the wheel and the Web browser. I’m not the first [1] and certainly not the last [2] to notice this. Jonathan Zittrain even predicted it would happen [3,4] with what he calls “tethered devices”. One solution to this problem, as suggested at last months International World Wide Web conference in Lyon by some bloke called Tim, is to develop mobile Web apps rather than native mobile apps:

There are lots of examples of this. Sites like trains.im provide train times via a simple Web-based interface, no app required. Many Web sites have  two versions, a desktop one and a mobile one. Wikipedia has a mobile site at en.m.wikipedia.org/wiki, Flickr at m.flickr.com, The Economist at m.economist.com, BBC at m.bbc.co.uk/news and so on. But in many cases these sites are poor cousins of the native mobile apps that software developers have focused their efforts on, diluting their work across multiple apps and platforms.

Maybe it’s too late, maybe I’m suffering from the suspicious of change” syndrome described by Douglas Adams like this:

  1. everything that’s already in the world when you’re born is just normal;
  2. anything that gets invented between then and before you turn thirty is incredibly exciting and creative and with any luck you can make a career out of it;
  3. anything that gets invented after you’re thirty is against the natural order of things and the beginning of the end of civilisation as we know it until it’s been around for about ten years when it gradually turns out to be alright really.

The mobile Web makes me suspicous because many apps re-invent the wheel. I’ve argued here that it is against the natural order of the Web, we’ve waved goodbye to the good old Web [5] and its the beginning of the end. I really hope not, it would be a tragedy to carry on killing the Web as it’s given us so much and was designed specifically to solve the problems described above. Let’s hope native mobile apps gradually turn out to be alright really.


  1. Gary Marshall (2011). Could smartphone apps be taking us back to the days of “best viewed with … ”? Net Magazine
  2. Jason Pontin (2012). Why Publishers Don’t Like Apps: The future of media on mobile devices isn’t with Apps but with the Web Technology Review
  3. Jonathan Zittrain (2007). Saving the internet. Harvard Business Review, 85 (6) PMID: 17580647
  4. Jonathan Zittrain (2009). The Future of the Internet: And How to Stop It Penguin, ISBN:014103159X
  5. Hamish MacKenzie (2012) Web 2.0 Is Over, All Hail the Age of Mobile, Pandodaily

September 5, 2007

WWW2007: Workflows on the Web

Don't PanicThe Hitch-hiking novelist Douglas Noel Adams (DNA) once remarked that the World Wide Web (WWW) is the only thing whose shortened form – ‘double-you double-you double-you-dot’ – takes three times longer to say than what it’s “short” for [1]. If he were still with us today, there is plenty of stuff at the 16th International World Wide Web conference (WWW2007), currently underway in Banff, that would interest him. Here are some short, abbreviated notes on a couple of interesting papers at this years conference. They are relevant to bioinformatics and worth reading, whichever type of DNA you’re most interested in.

One full paper [2] by Daniel Goodman describes a scientific workflow language called Martlet. The motivating example is taken from climateprediction.net but I suspect some of the points they make about scientific workflows are relevant to bioinformatics too. Just like the recent post by Boscoh about functional programming, the paper discusses an inspired-by-Haskell functional approach to building and running workflows. Comparisons with other workflow systems like Taverna / SCUFL are drawn. Despite what they say, Taverna already uses a functional model (not an imperative one), it just hasn’t been published yet. The paper also draws comparisons between Martlet and other functional systems, like Google’s Map-Reduce. It concludes that the (allegedly) new Martlet programming model “raises the interesting possibility of a whole set of new algorithms just waiting to be discovered once people start to think about programming in this new way”. Which is an exciting possibility.

Another position paper [3] (warning: position paper = arm waving) by Anupriya Ankolekar et al argues that the Semantic Web and Web-Two-Point-Oh are complementary, rather than competing. Their motivating examples are a bit lame (Blogging a movie? Can’t they think of something more original?) …but they make some interesting (and obvious) points. The authors think that aggregators like Yahoo! Pipes! will play an important role in the emerging Semantic Web. Currently, there don’t seem to be too many bioinformaticians using Yahoo! pipes, perhaps they just don’t share their pipes / workflows yet?

Running in parallel to all of the above is the Health Care and Life Sciences Data Integration for the Semantic Web workshop, where more detailed discussion on the bio semweb is underway. As its a workshop, there are no full or position papers, but take a look at The State of the Nation in Life Science Data integration to get a flavour of what is going on.

Wether functional, semantic, Web-enabled or just buzzword-friendly, there is plenty of action in the scientific workflow field right now. If you’re interested in the webby stuff, next years conference, WWW2008, is in Beijing, China. I wonder if they will mark the 10th anniversary of the publication of that Google paper at WWW7 back in 1998? The deadline for papers at WWW2008 will probably be sometime in November 2007, but around 90% of submitted papers will be rejected if previous years are anything to go by. If you’re thinking of doing a paper, DON’T PANIC about those intimidating statistics, because bioinformatics is bursting full of interesting and hard problems that challenge the state-of-the-art. The kind of stuff that will go down well at Dubya Dubya Dubya.

(Photo credit: Fire Monkey Fish)


  1. Douglas Adams (1999) Beyond the Brochure: Build it and we will come
  2. Daniel Goodman (2007) Introduction and Evaluation of Marlet, a Scientific Workflow Language for Abstracted Parallelisation doi:10.1145/1242572.1242705
  3. Anupriya Ankolekar, Markus Krotzsch, Thanh Tran and Denny Vrandecic (2007) The Two Cultures: Mashing up Web 2.0 and the Semantic Web doi:10.1145/1242572.1242684

Creative Commons License

This work is licensed under a
Creative Commons Attribution-Noncommercial-Share Alike 3.0 License.

Blog at WordPress.com.