Java 7 is the first major release of Java in half a decade, but even so it's more of a stage-setter than a real blockbuster.
After a decade of pretty regular bi-annual major releases, Java went into a lull from 2006 on. The Java 6 release, known as Mustang, saw a number of incremental releases, but even those slowed this year. The last few releases mostly consisted of the ubiquitous "improved performance and stability" and an update to the Tunisian time zones. And no, I'm not kidding about that! A significant portion of the releases to Java in the past year or so have been time zone updates. The biggest change other than time zones came in the latest update, number 27: Firefox 5 is now a certified browser version.
But things seem to be percolating a bit more nowadays. I think it took a little while for the folks at Oracle to really decide what they were going to do with Java once they purchased Sun Microsystems, but it looks like they've gotten their sea legs and are starting to move the platform forward again, albeit with a somewhat different tack.
Java Version 7, codename Dolphin (seriously?) has quite a few new features, but not a lot of blockbusters. Most of the more significant pieces will be coming in Java 8. Actually, Java 7 was originally slated for sometime in mid-2012, but instead Oracle opted to reduce the scope and get Java 7 out this year, with the remaining features to be included in Java 8 with a target of end of 2012. Practically speaking, that means that some of the larger-ticket items, such as closures, didn't make it to Java 7, which instead focused on a few mid-level additions and some interesting syntactical candy.
One of the biggest additions is support for 2D rendering engines. Called the XRender Pipeline, what this should do is provide a boost to 2D graphics on X11 implementations (particularly Linux). Unfortunately, this does nothing for Android, which doesn't currently support X11 graphics, so it's a bit of a thud for me.
A second major upgrade is the support for JSR 203, also known as NIO.2 or "more new I/O APIs." While the new feature covers a lot of ground, the gist of the enhancement is that NIO.2 will provide a more robust interface, with much more complete support for things like exceptions and metadata, features that were very much missing in earlier versions of Java. A goal here is to make Java more amenable to file-related applications, which could play well in the business application space. Note, though, that when I say files I basically mean stream files as opposed to relational data. Stream files are more and more prevalent, including everything from PDF files to transactions stored as XML to images to video. I think that the ability to properly read and write these files on whatever new devices come along will be an important facet of the next generation of applications.
An almost entirely self-contained part of the release comes from something called Project Coin, which tried to identify minor Java language changes that could provide the most advantage with minimal impact. These are interesting changes in that they require modifications not to the Java Runtime Environment libraries but to the very syntax of the language itself. Syntax changes are often very difficult because they tend to threaten backward compatibility, and changes need to be made to avoid breaking existing code. For example, that's why the new opcodes in RPG have hyphens: no variables could have been defined with a hyphen in the name because it's an invalid name.
Java has billions of lines of code out there, and breaking that code would be a very painful exercise. And yet they've managed to add a few very nice capabilities. One is incredibly simple and yet so needed: the ability to use strings in switch statements. The fact that the switch statement has been forced to use the 40-year-old C syntax is incomprehensible. Kudos for this change.
The most important change from a standpoint of pure tediousness is an enhancement to the diamond operator. If you're not familiar with the diamond, then you haven't done much with generics. In Java, generic programming requires you to define not only a class for a new object but also the type with which that generic class is identified (such as a List of Customers). Down the road that reduces a lot of code, and it enforces type-checking. However, when you define the constructor that initializes that object, you must re-specify the generic parameters, and this can get very cumbersome. A new syntax (an empty diamond) allows you to avoid that redefinition.
Another little gem that might sneak under a typical "Top 10" list is the ability to use underscores in numeric literals. Some of the examples are a little far-fetched; you could, for example, include a Discover credit card as 6011_1234_5678_9012, or define social security number 123_45_6789. But for me, the two situations where it helps are large numeric values, where you can include an underscore rather than a comma (9_999_999_999), or even better, the ability to segregate a long hexadecimal value such as 0x_CAFE_BABE (which actually is a very important number in Java; if you're interested, check here).
So Was Version 7 Worth the Wait?
Beauty is in the eye of the beholder. But unless you are in dire need of a boost in your X11 graphics performance, or you are just plain tired of writing verbose generics code, then I don't know that Java 7 has a lot to recommend it. At the same time, it's unlikely that Java 7 will break anything, so moving might not be a big deal. To me, I think Java 8 is going to be the big deal, and we'll be hearing a lot about that release in the next six to twelve months.
as/400, os/400, iseries, system i, i5/os, ibm i, power systems, 6.1, 7.1, V7, V6R1