The OOP conference (Object Oriented Programming) was held in Munich, Germany, from 24th to 28th January 2011 with “Business Impact through Mastering Change” as its general motto. Despite of its name, the OOP represents one of the largest and long-lasting events on the general field of software engineering.
This year the conference and its chair, Dr. Frances Paulisch, could celebrate their 20th anniversary, which the organizers took as opportunity to invite many internationally renowned speakers.
While the conference program covered almost all areas of software engineering, the main topics comprised Cloud computing, SOA, Software Architecture, Large Scale System Development, development for Multicore Systems and Mobile Applications. Interestingly, the five most attended talks all covered the area of software architecture addressing topics like complexity, the future of software architecture, frameworks, and design tactics.
As the keynote talks given by widely known speakers such as Tom DeMarco were among the highlights of the OOP2011, judging from the feedback and many tweets, this news article covers some of them in more detail.
Erich Gamma, one of the authors of the Gang-of-Four book, addressed the past, presence and future of software patterns. He emphasized that patterns have become common sense in the meantime. To prove his claim, Erich was referring to the large amount of literature available and all the professional platforms and applications such as Eclipse which leverage patterns. According to Erich, in the future, existing patterns are supposed to change or disappear, and new patterns to arrive. He offered some ideas how a new edition of the seminal design patterns book could look like, if it ever were written.
In his two-part keynote, Martin Fowler addressed Domain-Specific-Languages and Agile Development. As Martin explained, DSLs are very powerful tools which developers often neglect due to their fear of building parsers and lexers. However, internal DSLs can be expressed in the syntax of a given host language, and external DSLs need parsers and lexers that, according to Martin, are much simpler to create than their counterparts for general purpose languages. Martin kept emphasizing, that engineers should come up with a semantic metamodel when defining a DSL. In his second part, Martin introduced the history of agile development and why the Agile Manifesto is often misinterpreted. For instance, values such as “Responding to change over following a plan” do not imply, as he pointed out, the right hand side were something bad.
In his talk “Collaborative Design Imperative”, Tom DeMarco stated that the human brain is only capable of handling a limited capacity of information. In an entertaining estimation, Tom concluded, that the human brain might have only 1 Gigabyte of memory capacity. While software engineers had been able to know almost all aspects of IT in the 1960s, today the knowledge exceeds by far what an individual can handle. Thus, according to Tom,, complex systems development requires multiple people to collaborate. Communication as the most important means of collaboration should constantly adapt to the volatility of design. But communication alone is insufficient. It is important to get and give trust. “Trust is the bandwidth of communication.”
Scott Berkun, author of the best-selling book “The Myths of Innovation” held a keynote entitled such as his book. As Scott explained, most innovations do not happen in a short time frame. According to the “The Myth of Epiphany” an initial idea may come up very soon, but it takes a lot of efforts to create an innovation from the idea. Experimenting is an important vehicle in this context. In “We understand the history of Innovation”, Scott motivated why innovation requires to explore the idea’s conceptual area and also to take risks. Using the third myth “People like it when you innovate” Scott tried to illustrate that innovations also represent dangers for other people which is why sociology and psychology should also be taken into account.
Kevlin Henney chose 17 contributions from the book “97 Things Every Programmer Should Know” to structure his keynote talk. One example he mentioned was “Comment Only What the Code Cannot Say” which addresses that comments often verbosely explain properties of the code that could be easily obtained by reading the code itself. Thus, the guideline suggests, only commenting what cannot be derived from the implementation. Kevlin made the assumption, that good code should contain only a few comments, while a reviewer should always get suspicious if he recognized an overwhelming amount of comments.
According to the organizers from SIGS-DATACOM, 2000 participant have been registered for the event which is a 21% increase compared to the previous year. The next year’s OOP will take place in Munich from 23rd to 27th January 2012.