How distributed analysis helps building the right context

After many years in working on software projects I realized that the most common reason of misunderstanding the design of an application is the lack of context or to be more precise, the lack of a shared context among the participants to the meetings.

The chance of working with a distributed team according to the CalAgile approach highlighted the need of clarifying this aspect. According to this approach a representative of the software consultancy firm kicks off the project at the customer’s premises by setting up communication channels between the customer’s business-oriented team and the remote software-development team. This kind of organization requires a continuous rethought of the customer’s needs as the representative actively reports to the team the model they should build.

Professional Java Day in Frankfurt

As it was nearly a week since I last attended a conference I couldn’t be happier when on Wednesday my friend Miguel told me about the new chance available just at a few kilometers from Frankfurt. The conference allowed me to get up to date with Sun’s latest plans about the Java ecosystem.

The first presentation was from David Hofert (Sun) and it was titled “Taking Java to the next level”. Among the news I traced down the following:

StackOverflow DevDays in Cambridge

When last June I ran into an article from Joel Spolsky announcing the StackOverflow DevDays in Cambridge I took the chance on the fly and booked my seat. Last Thursday I eventually got to this wonderful city in the UK and after wandering around for a whole day, I attended the conference on Friday.

The first presentation was from Mr Spolsky in person… great way to start. Mr Spolsky discussed about compromises to make between simplicity and power citing general cases first and going in depth into the software business afterwards. It seems that an experiment run by some academic within a groceries store showed that when the number of the different varieties of jam shown on a shelf was low, the number of customers stopping to have a look was much lower but the number of items sold was 10 times higher. This result suggests that the lower the number of choices is, the easier it is for people to make up their minds and take a decision. On the other side, Mr Spolsky has a personal experience suggesting the opposite and based on the numbers coming from the sales of FogBugz, the flag product of his company, FogCreek.

Java, Webcam & Linux

In my opinion, the interaction between real world and computers is one of the most exciting part of software development.

To see the world with an electronic eye and be able to add information to reality creating an augmented reality is really amazing.

One of the fields we are strongly focused at Calameda labs is Augmented Reality.

Scrum Gathering in Munich – day 3

With this post I’ll complete the series about the Scrum Gathering in Munich. I started the third day of the conference by attending the presentation “Growing Self Organizing Teams” held by Harvey Wheaton of Supermassive Games (http://www.supermassivegames.com). Supermassive Games produces titles for the Sony PS3 and it was founded in 2008. After less than one year the company has already got nearly 60 employees.

Scrum Gathering 2009 in Munich – day 2

The first presentation of day 2 was from Patric Palm, one of the founders of Hansoft (www.hasoft.se), the Swedish company behind a product with the same name meant to manage agile and lean projects. The title of the presentation was “Making Agile matter to the bottom line by understanding group dynamics”. An impressing thing is that Patric started Palm when he was slightly older than 20 and I found also interesting his experience in the Swedish military air force.

Second presentation of the day was “Good product owner panel discussion” moderated by Simon Bennet. A very interesting point made during this discussion regarded the prioritization of stories of the same importance belonging to the same Product Backlog but referring to different projects. A simple solution to such an issue is to reverse the question and consider instead what happens if that specific feature is not delivered. Sometimes replying to this question is much easier.

Scrum Gathering 2009 in Munich – day 1

Between Monday 19th and Wednesday 21st October I attended the Scrum Gathering 2009. Since early 2007 I got  interested in Agile methodologies and attending events such this helps me to keep in contact with other professionals working in the same field and discover new approaches and new solutions to issues related to software projects development.

This year’s Gathering was hosted at the Hilton in Munich. I had never been to Munich before and I really liked the city. Thanks to Michael, a friend, I was able to visit the nice town center and understand something of the history and the architecture of the city.

The venue was nice but in some cases the rooms were too small and also, very bad, we didn’t have a wi-fi connection on the first day because the Scrum Alliance didn’t manage to have an agreement with the Hotel. Luckily after a lot of complains about this, some connectivity was made available to us attendees.

ZKoss: ajax without (visible) Javascript

Zkoss id by far one of the most interesting frameworks we have been using this year.

Very simple to use, well documented, it provides a very powerful tools set to build amazing RIA applications in a very short time, almost without having to write a single line of javascript.

Popups, lists, forms and every kind of components are made available through a xml language called “zul”. A aspecial event mechanism make you able to write your web application in a very intuitive way, following the same paradigm of desktop applications. Zkoss works behind the scenes, performing serialization, events dispatching and data binding.

A lot of demo applications made with zkoss are available here. There is also a useful developer’s reference, to help you finding all details about zkoss components, atributes and events.

Ammentos 2.0 beta 1 available for download!

We are proud to present the first Ammentos Persistence Framework release by Calameda. A lot of work has been done behind the scenes in the last months to improve the framework in many sides.

• Performance: This is going to be the faster Ammentos version you’ve ever seen. A fully revisited query engine has been integrated, which speeded up our stress tests execution of about 25%.
• Polymorphism: as requested by many users, Ammentos is now able to recognize the most specific type of loaded object instances. A new set of simple annotation attributes has been added to support this feature through the new types table mechanism.
• Facilities: specific new attributes like syncKeys, typeTableAutoCreate, typeTableAutoInsert have been added to make Ammentos do the dirty job, and let you free to think at your business code.

The updated user guide is already available at www.ammentos.org/userguide_2.0.htm.

Ammentos 2.0.b1 is available for download at sourceforge: http://sourceforge.net/projects/ammentos/files/.

Important: this is a BETA version. The last stable realease is still ammentos 1.3.7. We are making this beta available because we need your feedback, which is always the most important engine for any opensource project. Feel free to write your comments about this release, in response to this post, in our forums or by email.

Software for the human race

Computer science is a strange topic.

A lot of progress has been done since the first home computers appeared, in the 80s. Now computers are 1000 times faster, 1000 times more powerful, 1000 times smarter than those at that time. Our hard disks can store hundreds movies, hundred of thousands of pictures, millions of pictures, billions of documents.

And still… we interact with them in the same old, obsolete, slow way. A keyboard and a mouse!!

This paradigm is so invasive that almost all the software in the world (except for some experimental laboratory) suffers of this big handicap. The web, which presents himself as the new frontier of interaction, is obviously in the same situation. The result is that today’s software is made for some people with one digit in one hand, so not for the human race.

Developers (we are included) should really take in mind that: software is not meant for Lineland inhabitants, is meant for three-dimensional people living in a three-dimensional world, and most of that people have two hands and five digits per hand. Moving a cursor into a flat screen, clicking, dragging, dropping, selectioning and so on are NOT verbs describing normal interaction with any kind of real object.

Yes, I know, it is not so simple to develop this new kind of software. Of course we need a new kind of hardware to do the full job. But there are lots of opportunities already available that can help to move the first steps. Multi-touch screens, for instance.

At Calameda labs we are seriously addressing this point: making software really available for the uman race.