Yesterday, I sat down with Donald Smith to discuss the Business and Industry Tracks at EclipseCon 2007. Here’s an exerpt from our conversation:

“EclipseCon has a reputation as being a hard-core developer conference, and there’s definitely a lot of developers there. But we have a lot of marketing folks, a lot of business development folks, and a lot of engineering management types that are looking for content that is a little more business oriented.”

I captured the conversation in a podcast which you can access here.

There are currently seven long talks scheduled in the Business and Industry Tracks. These talks are:

• Making the World’s Best and Brightest Developers Happy …
• The Garden & the Market: The Value of Eclipse Open Source
• Working with the IT Press
• Open Source - The New Face of Technology Innovation
• Open Source Business Models: A Wall Street Look at a Wild…
• Building the 100 Bundle Application
• Eclipse and Industry Verticals: The Healthcare Experience

Short talks will be scheduled soon.

The early registration deadline for EclipseCon 2007 is approaching. Register today!

I’ve been looking over the recommended tracks for EclipseCon 2007.

Based on what I know of folks like Chris and Alex, their recommended tracks are more-or-less exactly what I’d expect. There is some overlap between their recommendations and mine, so I’m sure to bump into them. Actually, I’m virtually guaranteed to run into both of them anyway; they are the kings of mingling.

If you can’t get enough of the OSGi, check out An All OSGi Track. There’s OSGi and more in the OSGi/RCP/SOA Track. RichServer and RichClient is a similar track. There more on RCP (and related issues) in Open Source and the use of RCP in Commercial Applications.

An IDE Developer’s Track by Scott Kellicker includes numerous talks on building debuggers, supporting additional languages, language toolkit (LTK) and more; useful stuff if you’re building an IDE based on Eclipse technology (and I know that so very many of you are). Matt Chapman has provided the Aspect-Oriented Programming Track (with an obvious focus). If modeling is your thing, Joana M. F. Trindade has provided a useful track.

Andrei Loskutov, Benjamin Pasero, Benjamin Muskalla, Bjorn, and Gunnar have all provided personal tracks. A quick glance at these tracks reveals a great deal about what interests these folks.

Ken’s Embedded Track caught my attention as my own interest in embedded has increased quite a lot this year. There’s a lot of overlap between Ken’s recommendations and my own, so I’m looking forward to running into Ken in a few of these talks.

Interesting tracks for a database vendor has an obvious focus, but is a little more diverse than the title suggests. I love that database vendors are using Eclipse technology.

I updated my recommended track today; a few talks had changed slots, so I had to drop one, but managed to add a couple of new talks that I couldn’t wedge in before. I also decided to rename the talk. I was in a bit of an odd mood when I gave it the original title and decided to take this opportunity to provide a more useful and descriptive one. With the change, the track has disappeared from the site (it’s still in EclipseZilla), but I have faith that it will reappear at some point soon.

The Eclipse Nebula project has grown a lot in the last few months and it now includes an impressive collection of widgets, including a slick looking date and time picker. For the uninitiated, Nebula is a project where new widgets that complement SWT are created.

Nebula’s CDataTime widget is pretty slick looking, and is pretty intuitive to use.

This is what it looks like when you only care about the date. When you care about the time, it shows an analog clock:

You can choose to show either one or both parts (depending on the information that you need).

The widget is pretty robust, but it’s missing a few features that I’d like to have, so I’m going to try an submit a patch or two. The Mylar folks are also looking at this widget with an interest in integrating it into their stuff.

Happy Robbie Burns Day to all. Gie her a Haggis! and all that…

There are a great many features in Eclipse that I take for granted. One is the handy way that the "Add CVS Repository" wizard handles CVS URLs. The wizard is opened from the "CVS Repositories" view, by selecting menu item "New > Repository location…".

When you first see the wizard, it looks like you have to cut and paste the different bits of your CVS URL into the different fields.
However, if you just paste your URL into the first field, it automagically parses the bits out and populates all the fields appropriately. For example, pasting the CVS URL, ":pserver:anonymous@dev.eclipse.org:/cvsroot/org.eclipse", fills in the fields as follows:

I don’t do this very often, but it’s sure handy when I do…

I think that everybody agrees: choice is good. And with Eclipse comes lots of choice; just look at Eclipse Plug-in Central (EPIC) (which is pretty much the tip of the iceberg when it comes to Eclipse plug-ins). Given that choice is good, it may seem strange that there are so many discussions about features that should be included with a standard Eclipse download (currently the Eclipse SDK). A lot of folks out there really want an Eclipse IDE that comes standard with a GUI builder, XML editor, J2EE support, etc. But if it’s included in some standard download, doesn’t that preclude choice?

We all think that having choices is good, but when it comes down it, having choices means that you have to put in some effort to research the various options and make an educated choice. EPIC makes it easier by providing a rating system, so you can leverage the hard work of others; it’s reasonable to assume that a feature that’s highly regarded by others might be the right choice for you. Still, you have to do some amount of research to find features.

I don’t mean to make it sound like I think folks who don’t want to make their own choices are in any way lazy. We make lots of choices in our day and sometimes, we just want tools that just work. I think that a lot of us would agree: it’d be cool to have an Eclipse distribution that includes all the features we need "out-of-the-box".

Today, we have a few options in the form of Eclipse distributions. If you want a handy all-in-one download packed with the features you need, check these out. If you don’t want to assemble your own IDE, you can get a prebuilt one here. Some of them even come with installers. Think of these in the same way that you think of Linux distributions: very few people use "pure" Linux (I’m not even sure where you find such a thing); most folks use a distro that bundles Linux with some customization, a bunch of handy applications, and neat-o installer. As far as I am aware, the fancy installer isn’t part of the Linux that Mr. Torvalds produces. Hopefully you get the point.

The Eclipse Packaging Project was recently proposed. This project intends to make it easier to build distributions of Eclipse. At this point, the project is just a proposal, but there is a very active community developing around it and I expect that a creation review will occur shortly. The focus of the project is to provide the infrastructure required to create Eclipse products with proper installers and what-not. Along the way, several exemplary products will be created.

Anyway, if you want to get add your input into this project, the eclipse.technology.packaging newsgroup is the place to go.

I’ve been putting a lot of thought into Eclipse Corner lately. We’ve been getting a lot of proposals and submissions lately, and some of them have given me cause to reflect on just what kind of content should be posted on Eclipse Corner. The process today is pretty straight forward:

• The author submits a proposal through Eclipse Bugzilla
• The community comments, based on comments, whether or not the article is worthwhile
• The author submits a draft, the draft is reviewed, and an iterative process of refining the article begins
• At some point the article is peer reviewed, and if deemed acceptable, it is posted
• Kudos and some form of Eclipse SWAG are passed onto the author

Frankly, the process is a little broken. It’s a little too informal for my liking. We try really hard to get articles peer reviewed, ideally by the projects, but a lot of articles make it though merely because the editors like them. At present, there is no release schedule; we just push out articles to the site when they’re deemed ready. Probably the biggest problem is that there is no formal description of what kind of article belongs on Eclipse Corner.

It’s time to provide a little editorial direction. Like everything else at Eclipse, Eclipse Corner is open and transparent. In that spirit, I have some thoughts about what I’d like to see formalized. I invite you to comment.

Here are some thoughts:

Articles should be timeless. Or at least relatively timeless. That is, articles that contain time-sensitive material or are likely to expire in a relatively short period of time do not belong on Eclipse Corner. An article that discusses the ins and outs of some Eclipse API belongs on Eclipse Corner because (apparently) APIs are forever. An article discussing how great EclipseCon 2007 was or even the Europa release probably doesn’t fit.

Perhaps a little on the gray side of things are articles that discuss how to use exemplary products produced by Eclipse products. These aren’t API and so may change and evolve, potentially rendering the content obsolete. My opinion is that despite this danger, this sort of article belongs on Eclipse Corner.

Articles must pertain to Eclipse projects. Ideally, articles are about how to make Eclipse things (like APIs) work. Eclipse Corner is not the place for articles that discuss the specifics of proprietary products or code/products/whatever from other sources. More on the gray side of things would be an article that discusses how Eclipse APIs are leveraged to build a commercial product. IMHO, as long as the article focuses on how to do great things with Eclipse it’s okay; overt advertising of any commercial product doesn’t belong on Eclipse Corner.

Articles should be explicitly associated with an Eclipse Project. I think that this is reasonable and may even be assumed by some. However, there is currently no such restriction. I wonder if it might be too restrictive as there is huge value in articles that span Eclipse projects (using BIRT with RCP, for example).

Perhaps, articles can be explicitly associated with multiple Eclipse projects. Ideally, I’d like to get the PMC or PL from the projects to agree that the article is something valuable that should make it to Eclipse Corner, but I’m concerned about adding one more thing to their list.

Article proposals must be commented on in Bugzilla. Comments indicate interest. If there is no community support behind an article submission, then it might be best to just pass on it. This, of course, raises the question: how do you measure it? My current thoughts are that 3-5 people have to believe that the article is appropriate content for Eclipse Corner. Perhaps some number of committers from the associated projects shoud be included in that number. For example, 4 total with at least one committer from the project.

Push out new articles on a schedule. I’m thinking of lining up the articles on a monthly basis. This would give the editors the ability to properly focus their attention. Also, it would give authors realistic expectation of when they can expect to be published. Given the current volume, one solid article a month is probably reasonable. Any decisions that we make with regard to scheduling would have to change and I imagine that we’d have to accommodate exceptions; there’s no point it artificially holding back an article that’s ready.

Articles must be released under the EPL. This one’s not up for discussion. This is just the way it is. By putting the article on our server, it’s subject to the EPL. The content of the article must be licensable under the EPL. If the article shows code, that code must be compatible with the EPL along with the rest of the content.

---

I’ve been working on some more mundane things like getting the existing articles into a more consistent format and adopting the Phoenix look, but with 84 articles packed with various abuses of HTML, it’s a lot of work. It’s coming, be patient.

I look forward to hearing your thoughts on the topic.

Eclipse Developer’s Journal has published an article from Todd Williams titled “Eclipse: A Solid Desktop, Rich-Client, or Embedded Application Framework“. It provides a good overview of the Eclipse Rich Client Platform (RCP). If you’re thinking about building a rich client application, it’s definitely worth reading.

I’ve updated the link in this post to point to the first page of the article. Oops.

EclipseCon is my favourite show of the year. And that’s not just because I’m an employee of the Eclipse Foundation. Honestly.

I’ve put together a recommended track; I’m not sure if I’ve done it in the right spirit, since I’ve just made a listing of the talks that I’m planning to attend. It would be more correct to call it a combination of three of what I would call tracks.

The first "track" is concerned with Mylar. Unfortunately, there’s only one talk on Mylar, but hopefully we’ll also be able to schedule a BoF on the topic (there does seem to be a lot of interest, so this should probably be easy to do).

The second "track" focuses on eRCP. I’m facinated with eRCP and the whole embedded space, but haven’t had the time to properly focus any attention on it. That’s changing this year, and I’m starting by attending as many eRCP/embedded talks as I can. Leading up to EclipseCon, I’ve started tinkering with MTJ and eRCP/eSWT so I’ll be able to come to the conference armed with some good questions. At least I hope that they’ll be good.

Ajax in Eclipse is the third "track". There’s a lot of very cool stuff going on in the RAP and ATF projects that I want to learn more about.

After that, there are a bunch of other topics that I find interesting that I’ve just lumped together.

Anyway, the track is here if you’re interested. I look forward to seeing you at EclipseCon. Go forth and register. Do it now.

I’ve been playing on and off with the Graphical Modeling Framework (GMF) for a few months. I am by no means an expert, but I’m coming along. Over the past week or so, I’ve been playing with the the 2.0M3 milestone build (I see that 2.0M4 is out today). Setting up your environment to work with the milestone build is a little mindnumbing, as it prereqs several other projects’ work that must be manually assembled, but that little problem will disappear when the Europa update site is up and running (besides, the prereqs are pretty clearly laid out and it’s not all that hard to bring it all together).

I’ve cobbled together a little demonstration of GMF in action. I’ve built a very simple graphical editor that lets you assemble an orgchart. The implementation is very simple (complex things are hard to demo) and so only gives a small taste of what you can build if you have about four minutes on your hands. It actually took me a little less than four minutes to assemble the first pass of the application; I then go back and tweak it a bit which brings the length of the demonstration to about 7½ minutes.

The demonstration uses the GMF Dashboard which is pretty cool. To be honest, I’m not sure that I’d use it much after the first couple of tries, but it does reduce the total number of mouse clicks which is nice. It should be pretty clear as you watch the demonstration that GMF is designed for far more sophisticated use. Where possible, I’m doing the simplest possible thing and accepting defaults. You can do a lot more with it, but that’s something that I’ll save for a later time (besides, I’m only a beginner with GMF myself).

I can’t say that using GMF is dirt simple, but I don’t believe that it’s intended to be. However, after only a short amount of browsing, you can start building some pretty amazing things.

If you missed the link above, you can access the demonstration by clicking here.

In November, I wrote about some experiments I did with Eclipse running on Apache Harmony. At the time, I was concerned exclusively with how much of the Eclipse SDK I could get running on the Harmony JRE. Today it occurred to me to try some of my Eclipse RCP applications running on it. Without exception, every application I tried worked fine with Eclipse RCP 3.2.1.

I’d post some pictures for Gunnar, but there’s really not much to show. The applications look and act exactly as they did before. Nothing particularly facinating from the user’s point-of-view.

Combine this with Chris’ fancy new PDE feature in Eclipse 3.3, and the packaging story for your RCP application gets a lot rosier.