Twingly – Social Data with Passion

The second and final day of the conference offered many interesting speeches, both by new startups and the giants. Not only did the big ones speak, by accidentally meeting someone who knew someone we were given 2 minutes to present our thesis, the blogosphere screensaver. The response we got was very rewarding, everyone who has been speaking to us afterwards have been impressed. Over all a very good conference day where we got to speak to a lot of interesting people.

Mark Anders at Adobe
Mark gave a speech about Flex, with can be described as a combination of HTML and Flash. He meant that web browsers in general is not good for cool content, you need Flash for that.

Flex is a developer friendly way to create flash, it’s very visual and rapid when developing. It also gives much better performance than Flash (with AC 2.0 instead of 3.0 as in Flex). We were shown an example where the performance had increased with almost a factor 100.

More info at flex.org and about Apollo at adobe.com/go/Apollo.

Chris Willson at Microsoft

Chris gave us some history about Microsoft. The development of Ajax and about new stuff in IE7. IE7’s aim is to give an amazing user experience, with integrated RSS platform. It focuses on secure and trustworthy browsing. About security he pointed out:

• Protect against webfraud.
• Putting the user in control.
• Advanced malware protection in Vista.

Khoi Vinh – NYTimes

Currently there are no tools for making real time design for specific news stories. Everything is template controlled. When tools are made that can do real time design this kind of publishing will be dramatically more mature.

 We are also moving away from the traditional narrative design which is all about how well you tell a story. We are moving into newscentric interactivity. NT will add more and more tools to help with making this happens.

 The issue of control: the designers have been given more and more control over the content, but the users sometimes want to control the design themselves. NT is losing control over how their data is presented.

 There is nothing like a free feature. All features have a hidden cost in additional code, testing, long-term support and feature noise. NT tries to think of their applications as machines: you wouldn’t add a crank to a machine just to see if it works.

 Finally we got some good advice on how to make design like the NY Times does it. Some core points were:

• Options are obstructions. Things are only in the settings page if they couldn’t be reasonably solved within the interface.
• Keep you navigation within reason. You don’t have to be able to go anywhere from everywhere.
• Test like you mean it. Don’t let your boss do the testing, let your users do.
• Different textual clues to input fields make people write different things.
• Let things be what they are. Let tabs be tabs, button be buttons, links be links etc.
• Design with a maximum of elegance through a minimum of ornaments.

Simon Willson on OpenID
Simon gave a speech about the hazel with different usernames and password at different sites and how OpenID will solve this.

OpenID is decentralized and you pick who should manage your user information. This ensures that no single company will have full control over your identity.

OpenID is convenient to use because your id is a url, which makes it easy to remember and share. It also helps with filling in webforms with information that has been uploaded when you created your OpenID. You can also create personas to avoid having to create accounts everywhere. These personas are linked to your ID.

This system is built on trust between people. When you move in to a new webpage that requires OpenID it will find another webpage where you have already confirmed your identity. That webpage will then tell the other webpage that you are who you say you are, if you let it. It is much like friends introducing each others, but way more technologically advanced.Â

It is important to remember that as just as with other identities on the Internet your identity is not stronger than the way it was create. That is, having an OpenID does not prove that you are who you claim you are in the real world.

 So what sucks?

• Phishing. If you have your identity stolen it’s worse when it’s the key to a number of places.
• What happen if my provider goes down? This should be done by allowing multiple providers.
• Privacy! You might not want to use the same identity at all times. This is solved just as it has always been in e-mail: by having more than one.
• Difficult to explain. It took a little over half an hour to explain to a number of hard core geeks.

Â

 Jonathan Rochelle – Google Docs & Spreadsheets

Jonathan told us about how Google docs and spreadsheet was and is developed. We have taken down his main points for you.

 Why?

• Do yourself content creation is growing.
• Accepted/familiar interface of spreadsheets and documents.
• Accessibility from anywhere (with internet connection).
• Easy-to-use collaboration was something that a lot of people would like.
• Do-it-yourself community creation.

 Key features:

• Collaboration (real time).
• Online storage.
• Publishing (make data available for other people).
• API for development of specific-use apps (makes it adaptable for others specific needs).
• Over all: be Easy To Use.

 General architecture

• Server side calculation engine – lighten the client.
• Ajax.
• Collaboration layer.
• Storage layer.
• Infrastructure & shared services.

 Decisions

• Who is our initial user? People who already know spreadsheet? In this case yes, they were the first target.
• What are the must have version 1 features?
• Collaboration features (rich conflict resolution or simple “trust me” collaboration? Simple trust me was enough since people in general were working with people they know.
• Integration choices (which Google products, when?)
• Technical decisions:
  • Storage services… which fits best?
  • Stateful vs stateless (needed both in the end).
  • “Homing” – how to split the load.

 Lessons learned

• Have user interface help … and get it early.
• Test new ideas early.
• Use test harnesses, automation and benchmarks – early.
• Speed is critical (the speed when using the application).
• The needs of the many outweighs the needs of the few … or the one.
• Sometimes insight does NOT come from the user. The users didn’t know they wanted real time collaboration until the feature was there.
• User data is sacred (don’t ever give the user a reason not to trust your product).
• Provide a very specific value that users can describe. Not just “this is cool” but “I can increase productivity by 80% by using this”.
• Focus on features which align with your product’s value.
• Talk to users.
• The team is the most important thing.

 If you want to read more visit these resources:

Google docs & spreadsheets http://docs.google.com
Google API:s http://code.google.com
Spreadsheets API http://code.google.com/apis/spreadsheets/overview.html

The Virtual Office
This was a talk about experiences of having a company that does not have a physical office. His company has four employees and he gave some information on what tools they are using.

 The benefits of doing like this is a significant saving of cost, time on travel and an increase in flexibility. The disadvantage is that this is not for everyone, it takes discipline working when not having an office to go to.

Â

Bruno Figueiredo – Yetta
Bruno was talking about the importance of making sure people get exactly the information they need – developers should not have to read through a 600 pages manual for a two week development project.

We should ditch extensive documentation and instead start using a methodology, and focus on “the user experience of the documentation”.

To join the discussion on the subject send an e-mail to patterns@ixda.org.

Â
CrowdStorm
What will succeed and what will not in 2007?

It’s hard to say, but here are seven criteria to consider:

1. Don’t roll out another Me 2.0 App
2. Simplicity is key
3. Don’t sell technology – sell user empowerment
4. The Selfish Individual before the Social Good
5. Don’t try a play f the big boys can copy you
6. Entering the mainstream
7. Ability to generate a big loyal following

Â
Daniel Appelquist – Vodafone

First Daniel gives us some idea about the market for mobile web: there are way more mobiles (1200 millions) than pc:s that are webcapable. Next he tells us about the work in the W3 group to make standards and best practice for the mobile web. Two important, but possibly obvious things to consider when developing for mobile, is that you should remember the user is on the move and that the same URI should give a coherent experience over all platforms.

 More reading:

www.w3.org/2004/CDF/ (standardization).
www.vodafonebetavine.net (for individuals and small companies to upload your betas

or download and try others).
www.torgo.com/blog (Dans blog).
www.w3.org/Mobile/Planet (Planet Mobile Web).

Â
Rasmus Lerdorf – Creator of PHP
Rasmus gave a speak on various bits of pieces of things of interest. We list some of his main themes:

 Why do people contribute in these projects?

• Self-interest – I want to use this.
• Self-expression – look, I can do this.
• Hormones – I get a nice fuzzy feeling of belonging by helping out.
• Improve the World – I enable non programmers to take their idea and put it on the web.

 Next something about how to form a good group: “It’s not what people think about you, but rather what they think about themselves”. A lot of people feel like they own php. Important to think about how people feel about themselves when they are in your project.

 Some factors to make a webpage successful:

• Performance is important. If your site can’t handle the load you will fail. Your site has to be fast. To help with this you can use the fairly advanced tool KCachegrind to analyse where your resources are used in your webserver.
• You can never click a link and be sure unless you can understand everything in the url, otherwise you might have code injection. This makes the Internet quite useless, but that’s the way it is. A partial solution is in php 5.5. It is equipped with a filter function to only let exactly the right kind of data through. This kind of strict filtering ensures nothing slips through that is not supported.

 Some final thoughts:

• Avoid participation gimmicks. There should be a real interest of participating, award schemes do not work.
• Get their feeling of being part of the team going.
• Solve one problem.
• Clean and intuitive user interface.
• API:s is the future.
• Make it work.

Â
Tariq Krim – Netvibes

Netvibes is today releasing a universal widget API. If you build a Netvibe widget it will be available on all other widget platforms, by running through a virtual machine, very much like Java. The JavaScript runtime environment will be open source so anyone can make the widgets work on their platform.

A preview will be available next week on their developer network: http://eco.netvibes.com/uwa

Richard Moross & Stefan Magdalinski – MOO

These guys was talking about how you take a product that is very non-web2.0 and turn it into a success. They are selling business cards over the web.

 Their first challenge was to be different enough to be worth talking about and hire the right people. As they say, “if you look in the usual places you get the usual people”. Moo’s business cards has an odd size, can be made with different photos on each and has a great quality. Their employees are all hired through personal connections.

 The right partnership was also important. From the very beginning this company teamed up with flickr, making it very easy to find photos for you business cards.

 Finally 8 points of things they learnt:

1. We run our business on outboard brains. You don’t need to hire for everything.
2. No beta, no gamma, just permanently 1.0. Continuous improvement.
3. Code (not too) recklessly but watch closely and fix fast. If there’s an error every developer gets an e-mail =)
4. APIs are cool, but dangerous. This because you are letting other people have control over you.
5. Don’t internationali{sz}e. From the beginning did not try to make the website in many different languages. Instead used clear explaining pictures and as little and simple English as possible.
6. Do internationalize. Ship to all over the world. Support Unicode properly.
7. Give more by giving less. How much flexibility to give the users? Don’t build the Ajax Photoshop. Don’t give the users chance to create really ugly cards.
8. Our biggest problem: how to sell a 2.01D (not 3D) product online. You can’t properly see and feel the cards.

The solution to technical problems is often human.

Brice Le Blévennec – contactoffice.com

This company has designed a collaboration and messenging web application. Brian says it is built with state of the art tools and is more web 2.0 than similar applications. This mainly because it integrates with a lot of other stuff. A beta can be found at  http://beta.contactoffice.com