Stride – A new programming language for beginners

At the Greenfoot headquarters, we – that is: Neil Brown, Amjad Altadmri and myself – have recently worked on creating a new language within the Greenfoot environment: Stride.

The interesting thing about Stride is not so much the language design itself, but its interaction design: editing programs involves different interactions than existing editors.

The goal is that Stride sits halfway between block-based systems (such as Scratch, AppInventor, PencilCode, Alice, etc.) and text-based editors (such as Java or Python), maintaining advantages of both.

I will start a series of posts (text and/or video) here over the next few weeks trying to tell you what Stride is and why you should care. In the meantime, you can get Greenfoot, install it, and have Stride ready to go when the introduction here starts. Stride is built into Greenfoot from version 3 onwards.

Check back soon for the first overview video.

Java 7 for introductory programming — does it make a difference?

The recently released BlueJ 3.0.5 brought full support for Java 7. While it is always good to be up-to-date with the latest versions — what does that actually mean for introductory programming teaching?

Java 7 brought a considerable list of new features to the Java system. Most of them deal with fairly advanced or specialised topics, and will have no influence on introductory teaching with Java. Some, however, will become visible even close to the beginning of programming. Here, I give a short overview of what’s new, and what you — as a teacher of introductory programming — should be aware of.

Java 7 – What’s new

Most of the new features in Java 7 are under the hood or in specialised libraries. These are very unlikely to affect an introductory programming course.

Improvements to the internals of the JDK include changes to the VM to support dynamically typed languages and improvements to the class loader architecture. On the library side, there are new frameworks for concurrency, cryptography and new versions of JDBC and Unicode. All these will not be relevant for most introductory courses.

The two areas that are relevant are “Project Coin”, the code name given to a project defining several small language enhancements, and NIO.2, a new library for I/O. We’ll discuss these in more detail in a moment.

Equally important, what’s not included in this release includes support for closures (known as “Project Lambda”). This would have been the most significant change to Java since the introduction of generics in Java 5 or, arguably, since the original definition of the Java language in 1995. Project Lambda has, however, been deferred to JDK 8 in what’s known as Plan B. JDK 8 is currently scheduled for release in late 2012.

So, which of the new Java 7 features are actually relevant for our intro programming course?

NIO.2

The I/O classes in Java have been improved several times over the years. The release of JDK 1.4 in 2002 brought some significant improvements in the somewhat short-sightedly named NIO library. The naming is rather less than ideal, since it stands for “New I/O”, and the passing of time dictates that everything new will eventually get old.

Which brings us to Java 7 and the New New I/O. So what will this be called now — NNIO? Well, the designers settled on NIO.2.

NIO.2 brings a number of new classes and interface, some with very useful methods. If you are doing any programming that accesses the file system, it’s worth familiarising yourself with these. Particularly interesting are the Files and Paths classes, and the Path interface, which you can find in the java.nio.file package.

All other relevant new features are part of Project Coin. They are diamond notation, strings-in-switch and improved exception handling.

Project Coin

Diamond notation

Generic type details on the right hand side of an assignment can now be inferred. For example, where in Java 6 we had to write

   HashSet<String, Monster> monsters = new HashSet<String, Monster>();

to declare a HashSet variable and initialise it with a fresh HashSet object, in Java 7 we can write

   HashSet<String, Monster> monsters = new HashSet<>();

In other words: We do not have to repeat the generic types of the HashSet on the right hand side, and can instead just write <>. (This syntax is the reason that this construct is known as the “diamond notation”). The effect of this is exactly as the Java 6 version above. It is a purely syntactic shortcut. The compiler will fill in the generic types by copying them from the variable declaration on the left, and all behaves just as before. Of course, the old notation is still allowed.

Strings in switch statements

The variable used in switch statements can now be of type String. For example

   switch (command) {
       case "go":
          goForward();
          break;
       case "help":
          showHelp();
          break;
       default:
          unknownCommand();
          break;
   }

Before Java 7, strings could not be used in switch statements.

Improved exception handling 1: multi-catch

It is now possible to catch multiple exceptions in a single exception handler. For example:

   try {
      file = new File("readme.txt");
      process(file);
   }
   catch (FileNotFoundException | UnsupportedEncodingException ex) {
      ...
   }

As you can see in this example, the catch clause can list multiple types of exception in its header, separated with an OR symbol, to catch any of these types of exception.

Improved exception handling 2: try-with-resources

The second new feature relating to exceptions is called try-with-resources. It solves a hard problem: It was previously surprisingly hard to correctly guarantee that resources (such as files or network connections) were correctly closed in the case of an exception.

The new construct solves this. Look at this example:

   try (FileWriter writer = new FileWriter(filename)) {
      ...
   }
   catch (IOException e) {
      ...
   }

Before Java 7, the standard way to close a resource (a FileWriter in this example) would have been in a finally block, which followed the catch block. In Java 7, the FileWriter is opened in round brackets following the try keyword, marking it as a resource to be auto-closed. Once the try/catch block is completed, the resource will automatically be closed by the Java runtime system.

For classes to be used with the auto-close mechanism, they must implement the new AutoCloseable interface. All the relevant classes in the Java library have been retrofitted to implement this interface.

That already sums up the relevant Java 7 changes. As we can see — not much to worry about here. (This will be different next year, when JDK 8 will bring us closures. I expect we will have to have a lively discussion then about how to treat these in a first-year course. But we can leave that for a while.)

To find out more about the new constructs in Java 7, see this summary. The next version of the Object First book (5th edition, to be released in late October) includes discussion of these new features.


Note: This post was first published in the Blueroom.

Use the Microsoft Kinect with Greenfoot

Those of you who know Greenfoot know that one of its aims is to make programming for beginners exciting and engaging. (Those of my readers who don’t know it should have a look here.)

The most recent addition to Greenfoot is a library that allows programmers to easily use the Microsoft Kinect module with their Greenfoot scenarios. This means that you can now write simple Greenfoot games that are controlled by players body movements.

Probably the easiest way to show what I mean, is to show you what I mean. Here’s a short video:

Programming the Kinect with Greenfoot is probably the easiest way to write programs with the Kinect module. Neil Brown, one of our developers on the Greenfoot team, has adapted open source server software that communicates with the Kinect and designed and implemented a Greenfoot library that makes access surprisingly simple.

If you are interested to try it yourself — here are the detailed instructions. But beware: you might stand in the middle of your room waving your arms around for the next few days! Some people might look at you strangely, but it’s great fun.

Sharing of teaching resources – it’s about people, not about stuff

iconAt the beginning of April this year, we opened a new web site: the Greenroom.

The Greenroom is a web site where teachers who teach with Greenfoot can share resources and have discussions. It was clear for a while that sharing of resources was a powerful thing that was urgently needed for the Greenfoot community. Greenfoot is different from many other environments, teaching with it requires different projects and different ideas, and thus getting started with it, as a new teacher, is challenging. Having a community to talk to, to ask questions, to get ideas, to get tried and tested material, makes a huge difference.

Yet, this is a space where many have failed.

It is often said, with only slight exaggeration, that there are more teaching resource repositories than there are teaching resources. The fact is, countless resource repositories have been created, and most of them have tumbleweed blowing down the main street.

The typical pattern is this: A repository is opened, a flurry of activity follows, resources are submitted (often by the creators and other people personally involved or contacted), and then it dies down. A few months later, little is happening, resources are not maintained, few new resources are added, you can hear the cold wind blowing through empty spaces.

A high profile example is the repository of resources on the ACM SIGCSE website (one of the largest organisations in computer science education). It was – as far as I know – opened in early 2004, and initially attracted a good number of submissions. However, this quickly died down. Looking at the recent submissions, it seems that only four resources were submitted in all of 2008. This went down to three in 2009, with one single submission (so far) in 2010.

Now, I don’t want to pick on that one particular repository specifically. This pattern is not unsusual. I am pointing to it here as a typical example. Making repositories thrive is hard. (We have recently published a paper about this.) So, when we designed the Greenroom, we were very worried about meeting the same fate: spending a whole lot of effort in creating a repository site, only to have it die a slow, quiet death after a few short months.

I am happy to say that we seem to have avoided that fate. The Greenroom is alive and well.

From September 2009 to March 2010, the Greenroom existed as a Google Group. This gave us an excellent discussion forum, and quite poor storage of resources. Over these first six months, about 170 people signed up. Then, at the end of March 2010, we finished our custom implementation of the new Greenroom, and we were amazed at the result: even though only half of the Google Groups subscribers moved over, we surpassed the old subscriber numbers within two weeks, and then they continued to climb. Now, after eight months, we have about 900 people signed up. Lots of resources have been posted, and many interesting discussions are going on.

subscriber graph

Subscriber numbers of the Greenroom; old (red) and new (orange)

So, what makes this site work, when so many others have failed?

I think, in building such a site, you need to address a number of questions:

  • Who is allowed to upload resources? Everyone, or only trusted people?
  • How do you ensure quality? By prior review (work-intensive, hard), or do you let everything in (and risk flooding by poor quality, untried material)?
  • Do you restrict access? (If only teachers can access, you can also post exercises with solutions, exams, etc. But then you need verification.)
  • How do you maintain the resources? Who looks after them?
  • Do you store local copies of resources (which may go out of date) or links to outside resources (which may break)?
  • How do you organise searching and browsing? This requires relating resources to each other – who does that work?

And most importantly:

  • How do you make people come and submit interesting material, and participate in interesting discussions?

I think there are some design choices you can make to address many of these things. As a result, we have built a repository site that is quite different from others. We decided to not use a standard platform (because they only gave us the same old stuff that we didn’t want) and build our own from scratch. The main design choice, which makes most of the difference is this:

The site must be about people, not about stuff.

So we designed a site that is primarily about community, and then about material. There are many design details that embody this idea:

  • The main page of the site shows activities of people, not lists of things.
  • People sign up to the site, and have an ongoing relationship with it. It’s not something where you go once every two years when you redesign your course, it’s a place where interesting things are happening every week or maybe even every day.
  • People sign up with their real names, and are encouraged to post profile pictures. In the Greenroom, you move in a community of peers, not in a big, anonymous black hole.
  • The mental model is that of a staff room in a school, not of a library. It’s about meeting people, not about searching through piles of stuff.
  • Submitting unfinished resources is welcomed. We all know that most things never get finished – submitting something is better than submitting nothing.
  • The resources are owned by the community, not by single people. The edit rights follow a wiki model: everyone can edit everything. (I remember the moment in our design discussions when Neil Brown, one of our team members, suggested this. This changed almost everything. Many of the questions posted above are affected by this. Like many good ideas, it seems obvious in retrospect, but wasn’t at the time.)
  • Small contributions are welcome. Linking some related resources, for example, is a good thing. It’s quick, anyone can do it, and the wiki model allows everyone to do it.

In short: I think people should re-think teaching repository design. It has worked for us. (More details about this are in our paper.) There is daily activity in the Greenroom, many people visit regularly, we have frequent interesting discussion in our forum, and a large number of excellent resources. It has been going much better than we dared to hope.

So, if you think about designing a resource repository, think about the people, not about things. And if you are a teacher interested in Greenfoot, join the Greenroom. There are many friendly people there happy to help.

Comparing Scratch, Alice and Greenfoot

logosAt ITiCSE 2009 and again at SIGCSE 2010, we had panel session: Comparing Alice, Greenfoot and Scratch. The session came about because all three development teams – the Scratch, Alice and Greenfoot teams – were regularly asked one question: What’s the difference?

All three systems aim to let young people learn about programming. Many teachers (as well as parents or kids) have heard of more than one of them, but deciding which one to use can be difficult. While there are clear differences, the time it takes to evaluate all three of them is not trivial.

The panel session turned out to be very popular. The room was packed full, and we got plenty of questions afterwards. So we decided to create a written version of this session. And it’s now available.

We wrote a set of papers, which have now been published in a special issue of the ACM Transactions of Computing Education (TOCE). Each paper was written by a key member of the design team of one of the environments. The papers are more extensive and more in depth than the panel was (we all took the chance to write about various aspects our systems, which we had intended for some time, but never got around to doing), but they also aim to record some of the discussion that we had at the time.

Target age groups for Alice, Scratch and Greenfoot

The TOCE special issue consists of an introduction, three papers (one each) about the three environments, and a discussion section at the end. They are


(NOTE: All papers are available in the ACM digital library. ACM allows authors to also publish the papers on their own web sites. I will link to those freely available copies when they have been made available by the authors. So check back in a little while if you could not get all you were interested in.)

Scratch, Alice, Greenfoot—What’s the difference?

Do you remember the feeling when you were a kid and you had the fantastically rare chance to go into a sweet shop (or, as the Americans among us would probably say: candy store), and you actually had a bit of money, and you could buy something, and there was just so much choice? Wow.

(I still get a similar feeling now – not so much with candy, but with chocolates. Hmmm, chocolate…)

This is what it seems like with educational programming software. Some ten years ago there was not very much around (at least not much that was in widespread use and had good resources), and now there is plenty. For a teacher of introductory programming, it’s a bit like being the kid in the sweet shop: So much here to look at, so much good stuff. But so hard to choose!

If you haven’t used any of the current educational programming environments before, it’s hard to get your head around what’s in them. Just like the well known box-of-chocolates problem. When I talk about Greenfoot, I often get the question “How does it compare to Scratch?” or “How is it different from Alice?”

To help a little with that situation, we organised a panel session at the last SIGCSE conference called Comparing Alice, Greenfoot and Scratch which compared these three environments. “We”, in this context, were Sally Fincher and Ian Utting who organised the session, and Steve Cooper (Alice), John Maloney (Scratch) and myself (Greenfoot) presenting the environments. I really enjoyed the session – it was great to get one of the leading people involved in each of the development teams to present the environments, and it was well received: the room was packed full and the feedback was good.

But panels are transient – no good record is available for those who were not there. So we are now working on a series of articles for a special issue of the Transactions on Computing Education doing the same in writing. They should be published together in a single issue later this year, and include a paper on each of the three environments and a discussion section where we talk about commonalities and differences.

To read it, you have to wait until it comes out. But as a teaser, here is a graphic that we made for the papers, showing the target age groups for each of the three systems.

Target age groups for Alice, Scratch and Greenfoot

Maybe this answers one of your questions already. For everything else, you’ll have to wait a little longer.

Berlin – after one week

I have just reached the end of the first week of teaching here in Berlin, at the “Institut für Informatik“, part of the Freie Universität Berlin.

I am teaching a module called ProInformatik III: Objektorientierte Programmierung – a block course that runs every day with five hours of contact over four weeks.

So, what have I learned so far?

Apart from the fact that translating technical terms into German poses regular challenges (what’s the German word for Debugger?), that apparently my German sounds funny now (well, 18 years out of Germany have left their marks…) and that I can still use a blackboard with chalk (we had a power-out yesterday), one of the most interesting observations is that the students here are much more talkative, more engaged, and ask much more interesting questions than my students at home. Where does that come from?

Well, partly this comparison is unfair. This is a summer course that students can take as a taster before they start to study. If they pass it, and then enrol in the degree programme, they get full credit. This naturally leads to a selective audience. I would guess that it is much more likely that the well-motivated, organised, keen (and possible good) students participate in these voluntary summer courses, so I’m dealing with a subset that is probably easier to teach.

However, I wonder whether there is also another factor at work. In Germany – compared to UK – school starts a year later, and lasts a year longer (although that is just changing now, with most states in Germany changing from 13 years to 12 years of schooling). That means students enter university two years later.

Add to this the fact that Germany has compulsory military service (for males only) we get a situation – in this male dominated discipline – where many of our students are three years older than their UK counterparts.

In other words: Many start here at an age where UK students finish.

The effect is startling. Instead of dealing with kids, I’m dealing with young adults. I wonder whether another year of life before study wouldn’t be a good thing for UK students as well.

Greenfoot book out now in German

Now in German: Greenfoot!

The Greenfoot book, which teaches Java programming and object orientation with the Greenfoot environment, is now available in German (as well as the original English).

That’s great news: Germany is probably the country with most Greenfoot users outside the English speaking world, and having the book in German will make use of the book – be it adoption in the classroom or casual reading by hobbyists – much easier.

I have just received my copy, and it looks great. The publisher has done a great job with full colour printing and good quality paper.

The translation is great, too – it was done by Carsten Schulte from the Freie Universität Berlin. I was really lucky that the publisher found such an excellent translator. It shows that he’s a computer scientist and an educator as well. His deep understanding of the concepts has clearly led to a much better translation than someone without subject knowledge could have produced.

I’m happy to see it out, and to see it end up looking this nice.

Royal Society report on the state of computing in UK schools

A few days ago, the Royal Society has announced a study on the state of computing education in UK schools, its problems and possible solutions. This has been quite widely reported in the press, for example here. The announcement itself makes for interesting reading. Its introduction starts:

Numbers of students studying computing are plummeting across the UK, with a fall of 33% in just three years in ICT GCSE students, a fall of 33% in six years in A level ICT and 57% in eight years in A level Computing students in England and similar declines found elsewhere in the UK.

It contains some good quotes from a range of people, such as this by Matthew Harrison, Director of Education at The Royal Academy of Engineering

“Young people have huge appetites for the computing devices they use outside of school. Yet ICT and Computer Science in school seem to turn these young people off. We need school curricula to engage them better if the next generation are to engineer technology and not just consume it.”

This is the latest development in a growing trend in the UK that recognises the dismal state of computer science education in UK schools, and starts to work on finding solutions. I became involved in this topic a couple of years ago, through the Computing At Schools (CAS) group, a fantastic movement of really motivated and smart individuals initiated by Simon Peyton-Jones. This has grown over the years with initiatives such as developing curriculum for schools, organising computing teacher conferences, and founding a computing teachers association.

While the announcement of the study contains little new to those who already had an interest in the topic, it’s great to see that the Royal Society is getting interested and getting involved. This helps getting more organisations on board with making a change (already apparent from the announcement) and will push the issue higher on the agenda of those who are in the position to make decisions.

There is no doubt at all in my mind that such change is urgently needed and important. Currently, there is a whole generation of kids, some of whom would make great and enthusiastic computer scientists, who never find out about the possibilities and joys of this exciting discipline.

Altogether a great development.