This document is available on the Internet at:

The Future of Web Applications

Date:  10th February, 2006


Today I'm trying something new. For the first time ever (but hopefully establishing a precedent we can return to), Urban Mainframe features a guest writer, Noah Slater.

Noah recently attended a conference on "The Future of Web Applications" and, writing exclusively for Urban Mainframe, presents his conference notes here.

I'd like to thank Noah for taking the time to write up his notes and for sharing them here with us today.

Over to you Noah...


My name is Noah Slater, a web applications development leader from the north of England, friend of Jonathan's and attendee at the Carson Workshops Summit on the 8th February - "[a] one-day conference focusing on the development technology you'll be using tomorrow."

It should go without saying that the opinions expressed here are my own and not those of Jonathan, my employer or anyone else for that matter.

The hotel was swanky, the venue was impressive and the atmosphere electric.

I do remember marvelling at the size of the main screen which is quite possibly the biggest projector setup I have seen in my life. The majority of people had laptops, of a decidedly Apple variety. Via the wonders of the event's WiFi people were blogging, sending pictures to Flickr using their mobile phones and even chatting to each other on IRC. I must admit however that all this was a little too geeky for me, preferring instead to actually listen to what the speakers were saying.

The Speakers

Joshua Schachter - Delicious

Joshua was the first up, which was a shame because he gave a good talk and I was still quite hungover from the night before. That, and a certain someone thought it would be funny to have the hotel porter deliver some crutches to my room at 5:30 for no reason, thus waking me up far to early.

The majority of his talk was devoted to the history and technical side of setting up Delicious. I was a little disappointed he didn't speak much about the concepts behind tagging, social taxonomies and implications for the future of meta data. This was a session about the future of web applications after all.

Some soundbites that stood out:

The idiots out there are a lot cleverer than you. They will find ways of breaking your site you never dreamed of.

Build features that people will actually use - not ones they ask for

Though I couldn't really get my head around this one:

Unique IDs are bad for scaling, but you'll figure that out one day right?

Cal Henderson -

Cal gave a nice talk, focusing on the future of [the] web as a whole. Already, however, a few recurring topics or themes began to emerge, something that would prove almost tiresome as the day progressed. The concept of clean URLs made it's first appearance of the day, something that I am (like Tom Coates) far too interested in.

Amusingly he was also the first of a few people to comment on the name AJAX:

Quite possible the worst name for a technology ever.

I for one completely agree, what with it involving absolutely nothing specific to JavaScript or XML. In fact, as a few people pointed out, it should probably be named "A" as this represents the only unique (within the concept of traditional website architecture) thing about it: asynchronism. At least the naming is consistent with other aspects of the technique however, such as the XMLHttpRequest method which is quite possibly the most ill thought out method nomenclature I have ever seen.

Good points on the importance of i18n and L10n but contentious issues raised about tailoring content to mobile devices. Despite the fact that (as he pointed out) this issue comes around every few years (who remembers WAP?) I remain unconvinced. The web should be device independent.

Cal also talked about propriety data lock-in within your system. "If you let people think they can escape with their data they will be more likely to stay with you." I couldn't agree with this more. I think we are going to see a growing importance placed on this issue over the next few years as people start to expect transparent access to their data.

Cal rounded up his slot by mentioning licencing issues, specifically open content with respect to the Creative Commons licencing enabled by default on all Flickr accounts. He finished off nicely by explaining that every image he had used in his slides was openly licenced and available from Flickr.

I think we are going to see a continued increase in an emphasis on open content over the next few years and it was with great interest that I learned of the new Creative Commons enabled searching via Google and Yahoo.

RSS feeds are a read-only API.

Not too sure I agree with that. This "API" term seems to be getting thrown around quite a lot lately...

Tom Coats - Yahoo

Web 2.0 is all about rounded corners and gradient fills Tom states in his humourous start to an entertaining talk. But what else is Web 2.0 he asks:

  • A buzzword
  • A conference
  • A marketing term
  • A new bubble
  • All these and less

I am inclined to agree with all of them. It struck me after listening to Tom's talk that "Web 2.0" sounds like a way of selling the new dot com bubble to the VCs by distracting them from the dot com bust.

"This is such a good idea, it's going to make us all rich!"

"What about the dot com bust?"

"Nah, that was the old web. This is Web 2.0 baby!"

This was the best talk of the day by far. Tom is a seriously amusing guy, very entertaining.

Looking back at my notes, I see my first remark is "Funny guy, jokes about testicles..." I guess you had to be there for that one though.

One of Tom's more interesting ideas, and one which I have been thinking about ever since, is how current data sources are located privately and exposed to the public via websites. He argues that as the web evolves and these websites provide interfaces, the data sources will metaphorically move from behind the interface into the public domain with everyone sharing what is essentially a huge collective pool of data. Or a collection of testicles, depending on how you looked at his diagram.

Tom gave us a preview of Astronewsology, a private project developed with Simon Wilison. So good they can't release it to the public because it would "blow [our] minds". Astronewsology allowed you to browse the news by star sign, comparing what was forecast by the stars with what actually happened. Should make for particularly interesting reading when looking though the obituaries as Tom pointed out.

The point of all this was to demonstrate how easy it is currently to take such disparate data sources — when provided with an interface — and combine them in a useful, often unexpected, manner. The example was funny, yet inspired and lead on to some nicely summarised consequences for the future of web applications or in this case, mashups:

  • Massive creative possibilities
  • Accelerating innovation
  • Increasingly competitive services
  • Increasingly componentised services
  • Increasingly specialise services

How is money to be made from all of this though?

  • Use APIs to drive people to your content
  • Make your services more attractive and useful with less central development
  • Use syndicated content as a platform
  • Turn your API into a paid service

Did I mention he was funny? I would pay to see this guy in stand-up.

The future is the "aggregate web."

David Heinemeier Hansson - Ruby On Rails

No, not Ruby on Rael but a web development framework written in Ruby.

David gave a nice introduction to the framework, but I remain unconvinced. The main idea seems to be that code should be implicit not explicit, something I disagree with strongly. I think that while someone experienced with the code base may find this marginally more useful, all it really leads to is confusion. I don't like the idea of assuming full system knowledge and hiding implementation specifics away.

I turned off when he mentioned that Rails encourages the embedding of application code within the templates. He obviously hasn't worked for a large company with a separate design department.

Most of the slides throughout his talk were filled with Ruby code, a very daring move indeed. You run the strong chance of loosing most of your audience like this, something exemplified by my only notes from his talk being:

You are not unique. You are not a beautiful happy snowflake.

This came from his speech about how 80% of problems are mundane and these are the ones Rails helps you with. The "snowflakes" are the unique problems, the interesting ones. The beautiful ones that make us happy. Something like that anyway.

Shaun Inman - Mint

My apologies to Shaun, but his session was rather weak. It felt more like a student's end of year presentation than a professional talk. I found it hard to keep my attention and it looks like I wasn't the only one.

Shaun mostly talked about his non-free web application Mint, but only had enough content to speak for around 10 minutes. The rest of the session was spent filling in the time answering largely off topic questions.

A little like some of the other talks, the time Shaun shared with us was mostly looking back on how his application was built. While this information might be useful in some respects I felt it was a little out of place at a conference about the future of web applications.

Like I was hoping, someone in the audience mentioned the issues he has been having with his licencing choices. His tone was one of surprise as he explained that people had chosen to redistribute his application's source code which he freely provides when you purchase a licence.

To me this seems absurd. It was going to happen. It's that simple.

Instead of worrying himself about how to prevent redistribution why doesn't he do the good thing for the community and re-release his application as free software. Freedom 0, baby!

Also, interesting to note Ryan Carson's comments as Shaun left the stage:

What Shaun did is impressive because it had an API. That's really the reason we got him to speak today.

Sounds like he was trying to justify something to me...

Marketroid - Adobe

I'm not going to say much about this talk other than the fact it was quite clearly just a marketing opportunity for the closed-standard and non-free Flex product line.

It's a shame this 15 minute advert was included at the summit. A much more interesting, and relevant, replacement would have been a talk about the future of SVG - a very interesting and openly specified media format.

Thinking about it, one of the main gists of the sales pitch seemed to be "write once, run anywhere." Wow, maybe I was wrong after all, this novel concept could really change the web.

Interesting to note that Carson Workshops have removed all reference to this talk from the summary page.

Ryan Carson - DropSend

Ryan gave a brilliant talk about the real money behind building a commercial web application and setting up on your own. I think this was an excellent addition to the day and really drove home some simple hard facts about the challenges you will face. Another good focus was on the techniques used to analyse and exploit your potential revenue streams.

Notes on Ryan's talk (and others) can be found at Simon's site.

UI design is the most important aspect of designing your app, if you don't get that right your screwed.

I disagree, the most important aspect of designing your application is making sure it does exactly what it says on the tin.

Steffen Meschkat -
Google Maps

I'm too sure what to say about Steffen. His talk felt like an over verbose introduction to JavaScript, briefly touching on the interesting yet irrelevant topic of closures, and nothing much else. It reminded me more of a programming 101 lecture than a lively talk about the future of web applications.

Of interest however was that before Google he claims to have built a 1-to-1 virtual reality model of the entire word. I have no reason to doubt him, but it did remind me of an old colleague of mine who once claimed to have written a fully conformant HTTP 1.1 server in 40 lines of Java without using any external libraries.


The conference was the first of it's kind I have ever attended, a great experience and I suspect the first of many. I expect Jonathan to be joining me for the next one.

The conference ended with a Q&A session for the speakers. The last two blew a metaphorical hole in the entire day's events:

  • How do we address the numerous accessibility issues when implementing these development techniques?
  • If you base your business on the aggregate web, what do you do when someone in the chain doesn't maintain the service level expected. Are SLAs really a feasible approach to this problem?

The answers to these questions are left as an exercise for the reader.

On a personal note, the highlight of my two days in London was spending a magical evening before the conference with a long lost love of mine. It reminded me again what the time before new love feels like... when everything is just new. "When a smile or a wave would keep you going all day, and laughter is like emotional crack."

You are a beautiful happy snowflake.

Everything is new again.