Bruce Boyes's Blog

JavaOne Archives

Dare we ask for or (shudder) try to impose code standards on submitted projects?

How many of you have downloaded some code from the web, and been unable to use it because of 1) no meaningful comments in the code and 2) no other documentation such as useful examples? One of my senior project student teams struggled to use a poorly-documented AI engine downloaded from the web and finally gave up and wrote their own. Not exactly the kind of re-use of which Java code should be capable.

What standards? An easy metric is this: if your docs look like the Sun JDK javadocs you are probably OK. In fact Sun wrote some Java coding standards, so let's start there.

All of this was brought to the front of my consciousness by the JavaOne Monday session "Files From Mars" TS-1416 - analyzing Mars Rover data. The presentation focussed on three areas which are perhaps as interesting for what they don't mention as for what they do:

1) simplicity - to handle risk, cost, schedule
2) extensibility - to handle uncertainty
3) reuse - to handle risk, cost, schedule

Then, some corrolaries to the above: a) Know your users: b) one size does not fit all c) Requirements docs are not enough d) User assumptions are often wrong or incomplete

Nothing here about razzle-dazzle, performance, ease of use (per se), graphics capability, etc. But when you think about it, it's hard to argue with the first three topics. Simplicity should help with keeping on schedule, good performance, and ease of use, right? Extensibility means that others will be able to build on top of what has been done thus far, without changing the core code. Plus you too will be able to extend your code, when users demand a feature you didn't originally contemplate. Reuse means that you expect to get a lot of future benefit from present work, by applying (possibly modified) core code in other projects. Sounds like a pretty good mantra for open source projects.

The first enabler for extensibility and reuse is - documentation. In this session's project, fifty percent of the lines of code were comments. That sounds about right to me.

"It is doubtful whether a man ever brings his faculties to bear with their full force on a subject until he writes upon it."

-- Cicero

A long day in SFO - and - Fun with Puppets

We started at 8AM with a breakfast cooked on the spot - thanks to Flip! All day we have been hearing from java.net community leaders and organizers about what works (and what doesn't) in their various communities and projects. I've learned alot but it's like 'sipping from a firehose'.

For example, there are some great Eclipse tutorials in the Java tools incubator. There is a new Java user group mascot who made the journey all the way from Brazil. I'm thinking that Juggy needs a cousin Robby from the Robotics community...

On a more serious note, Boeing recently selected JXTA(TM) Technology to provide the peer-to-peer (P2P) discovery service for the System of Systems Common Operating Environment (SOSCOE) for the U.S. Army's multi-billion dollar Future Combat Systems (FCS) initiative.

Speaking of JXTA, the JXME project seems to be alive and well. This is what we want to see as an optional application layer on top of RF links between mobile robots.

Tomorrow is the second and last day, and it starts at 8AM again. At least there is no shortage of good coffee in San Francisco...

This work is licensed under a Creative Commons License.