Saturday, June 13, 2009

Seeking Input For My Next Opensource Project


I'm seeking your input to help me think about my next opensource project. I have two ideas, either one of which I'd like to do, probably using Java as the primary language, just as a matter of preference. I'd particularly like to know whether either is already being done, so that I don't duplicate work, and whether or not you think it might be something useful in work that you've done.

The first is a SOA Directory Service alternative to UDDI. When I worked to help implement a middleware/SOA framework in the mid-90s, one of the pieces we built was a directory service. While it didn't offer the metadata storage capability that UDDI offers today, it had some advantages over UDDI. It was simple: easy to register services, do lookups, etc. It was very fast. Services registered using a lease mechanism, so you could get a list of matching service instances, knowing that the instances were probably still up. Next, it was replicated. Certainly, many UDDI implementations support replication. What I don't want to do is create another UDDI implementation, but rather to build an alternative Directory Service that is more like what we did in the 90s, consistent with today's framework needs, but more lightweight. To my mind, UDDI is far more heavyweight a solutions than most enterprises need, and a simpler solution might offer some appeal, provided that it integrated well with whatever framework they're already using. That is, that it would be easy to choose it as an alternative to UDDI.

The second possibility is to do an opensource implementation of a data federation system. We built one for a client that was never used, but there were some good ideas in there. I'd like to do it again as an opensource project, because it offers some useful capabilities. It essentially allows users to publish documents to a master node, then replicate documents to regional servers, that is, to push the data close to where it would be used within the organization. For example, if a document were flagged as pertinent to an organization's European region, it would be be pushed to that region's server, and to its backup server in a neighboring region. Users in the region can then make annotations to the documents as needed, and push those back to the original author for consideration. A federation system such as this offers some availability and performance benefits relative to having a monolithic document server. When a user wants a document that not stored in his or her region, the system goes back to the master or another regional server to fetch it. As an added capability, the original system supported plugins that could fetch data from external sources, and that might be useful to include.

Sunday, June 7, 2009

We Don't Have Time To Skip That Step

In the mid-90s I was privileged to be on a team building a service-oriented middleware architecture (Datagate), which I have mentioned before. We used an underlying library that implemented XDR, for which we had no unit tests. Since much of our software was built on this and a couple other core technologies, it was important that they be as solid as possible. Make the foundation solid, and the rest will follow. I decided to build a suite of unit tests against this software, suspecting that there were some bugs in there. Using a coverage tool, I wrote tests that covered the entire XDR library, and we found that the suite would not run on one of our twenty or so supported platforms. We fixed that bug in the library.

After that, we found that one of our nagging bugs went away in our Directory Service. It turns out that the two were related. I took some time to build the test suite, but not that much. It saved us time in supporting the Directory Service. It probably saved application and service developers time, too, but we didn't research that.

After that, when it came to the value of testing, I told everyone who would listen, "We don't have time to skip that step." And it's still true today. Can you afford the extra time it takes to skip testing?

Saturday, June 6, 2009

OT Courtois Section Trip Report - 5/30 and 5/31

I hiked the part of this section from Hazel Creek to Bass' River Resort last weekend, camping after it crosses FR2265 the first time, before the trail cuts west there. What a great weekend to hike - a little on the warm side, but tolerable by watching my pace. Finished up Sunday morning about 11 before the heat really got going, and sat in the shade by the creek, enjoying a couple free beers from some new-found friends and skipping a few stones. Good times.

If you hike this direction, you need to know that when it goes into the field near Harmon Spring, you need to head forward across the field, taking the trail that goes slightly to the right. There's a trail to the left and to the right, and I went left, getting about a half-mile up the road before convincing myself that that wasn't the right way.

Also, if you're going this way, make sure you get plenty of water at the Beecher Artesian Well spring, especially in warm weather. The next decent water isn't until you're almost at Bass', and a long trek on the gravel road section will surely dry you out. There is some water before there, but I decided it wasn't for me, even purified, and I'm not too picky.

South of Highway 8, there are about 4 or 5 deadfalls. North of 8, maybe a couple.

The trail's pretty horsey in sections - plow through, and it gets better.

The ticks were bad, too. Had a couple in new places from this trip. Both standard ticks and seed ticks were on the prowl.

Not a lot of people out - just saw three guys in the Berryman section, but there were at least three people ahead of me, but making better time.

I really like this section. At least on the part I hiked, the hills are pretty docile, and there's some pretty trails through some really impressive pine groves.

Thursday, June 4, 2009

Manage Your Eclipse Install With A Local Git Repository

I had something of an epiphany this morning. If it wasn't the real thing, it sure felt like it. With Eclipse, I'm often trying out new plugins that purport to do one other or another, and there are usually a few that do the same thing, but with different features. So the quandary is, what do I do when I pick one of the several, try it, and don't like it? Perhaps it's hard to use, perhaps my IDE crashes more often than it used to, or perhaps one of the other similar plugins suddenly seems more appealing.

Traditionally, I've solved the problem using rsync. I rsync -a eclipse/ eclipse.beforeWonderfulPlugin, then install the plugin. If the plugin turns out to be a flop, I go back to the old version, which takes a little while. First, delete eclipse, then rsync -a eclipse.beforeWonderfulPlugin/ eclipse. It's slow, but it easier and faster than say, cp or tar. Also, I now have (at least) two entire copies of eclipse laying around.

There are further issues, too, though they're perhaps less important. It's hard to remember what plugins I have installed, for example, above and beyond what's already included.

Lately I've making the move to git to manage source for various projects. Then it hit me - I should make my Eclipse install a git repository. So that's what I did. There's a master branch - that's where the downloads from Eclipse go - ganymede->ganymedeSR1->ganymedeSR2, and so forth. Then, there's the working-branch, which is where I'll normally run from. When I'm trying a new plugin, I'll branch off of working-branch, try the plugin for a while. If it's good, I'll merge it back into working-branch.

Presumably, there will be an Eclipse Ganymede SR3 at some point, and I'll rebase there. When Galileo comes out, I'll put that on master, and probably start a new working-branch.

If I merge a plugin into working-branch, and find out a week later that there's a problem, it's easy to re-branch from a point prior to the merge, and get rid of the plugin. Sometimes it's the case that there's some quirky behavior I didn't notice at first. It's a matter of a few minutes to go back to a previous version, try it out, and see if the quirky behavior was there all along, or started with some new plugin.

Sure, Eclipse lets you uninstall plugins, most of the time, but there are often problems with that.

Could this be done with CVS, Subversion, ClearCase or other revision control systems? Very possibly - but not in practical terms, simply because of the performance issues.

If there are alternative plugins, say, Subclipse and Subversive, and I want to try them both, I can have them both on separate branches off of working-branch, and explore them both for a while before picking one. As an added bonus, gitk lets me see where I've been, and when I started using which plugins.