A Look Back At 2009 For Eclipse

as we reach the end of 2009, it's the perfect time to take a look back at another successful year for the eclipse community. as well as giving my own thoughts on the past year, i got some feedback from mik kersten (mylyn & tasktop), ian skerrett (eclipse foundation), scott lewis (ecf project) and oisin hurley (2010 eclipsecon program committee chair).

some of my own highlights:

• another successful release: galileo once again proved that the eclipse release train concept is one of the best ways to gather community projects to release together. for this year's release we did a series of podcasts at eclipsezone. just in case you missed them, we covered pde osgi tooling improvements , the cdo project, ecf and distributed osgi , the rich ajax platform project and the pulsar initiative. the final interview in the series covered provisioning with p2 with jeff mcaffer.  and the work on the next release train continues, with 3.6m4 becoming available recently.
• the rise of e4: the e4 project continues to look promising - a new ui model defined in emf and the ability to use css to define and decorate your swt widgets. of course there's lots more to it, that you can read in the e4 whitepaper . one of the coolest things i've seen recently was the concept of using e4 as an open social gadgets container . in 2010, e4 will be the project to watch, both in the eclipse community, and for java ui developers everywhere.
• my own contributions: i've been lucky enough to have some time to contribute code to the ecf project this year with tweethub . although it's far from complete, it's been great to finally get into an open source project and see what the whole process is like.
  i also got the chance to follow up on the essential emf refcard that i worked on last year with ed merks with two more refcardz - getting started with eclipse rcp and eclipse plug-in development .

mik kersten, ceo tasktop & project lead mylyn

looking back, the eclipse ecosystem sailed steadily through a downturn that had a broad effect on our industry. one of the biggest software events of the year was vmware's acquisition of springsource for $420m. this has
already had a positive effect on the eclipse-based tools for the spring and grails frameworks, and springsource could make an even bigger impact on eclipse in 2009.

on the mylyn end, the most notable thing that happened in 2009 is the rapid growth of our open source and commercial alm integrations. we're nearing four dozen, and have seen particular interest from the agile
space. to wrap up the year, just this week we announced versionone integration plans, released support for thoughtworks studios mingle last week, and have more coming in the first quarter of 2010. an additional
impact this has had on eclipse is that we're in the process of moving and generalizing the mylyn connector discovery technology into p2. this will enable a much easier install story for key 2010 initiatives such as the
eclipse marketplace.

beyond that, keep an eye out on the runtime project at eclipse, since the world is slowly but steadily discovering the benefits of osgi runtimes on the server. also expect some major activity from the mylyn project in the first quarter, as we are in the process of a planning a restructuring in order to handle the growth of our ecosystem, and expect to make more key players on the alm side of eclipse a key part of that effort.

ian skerrett, eclipse foundation

the eclipse modeling community really shone in 2009. everyone seems to be using emf and are exploring other projects like cdo, xtext, gmf, etc.

2009 was the year the e4 team starting producing code. they did their first 0.9 release and are making nice progress for a 1.0 in 2010.

from an industry perspective, cloud computing and mobile development were the two hot trends. from an eclipse perspective, eclipse is pretty much the de-facto ide of choice for both trends.

in 2010

• e4 will be the talk of the community. i think features like support for opensocial gadgets, bundles written in javascript and the use of emf to model the ui will have a huge impact on the community.
• javascript in general is going to be important to eclipse and the entire industry. i see it in the mobile space, e4 and even in the osgi world.
• i think we will see even more interest in eclipsert and osgi. eclipse gemini will add the features to equinox that are important for use of osgi in large enterprises. the large application server vendors are already talking more about osgi to their large customers.

scott lewis, ecf project lead

i think the big community news is the 'tragedy of the commons' issue, and the unsustainability of the current eclipse foundation model...i.e. as highlighted by bjorn freeman-benson's set of posts over the middle part of the year. i think it's becoming obvious to everyone that the eclipse foundation kowtow's to the corporate membership...and with all of the participation/contributions of the corporate membership going down, and ibm's continuing to go down rapidly, we are seeing eclipse stagnate. it's no longer considered particularly innovative, and in fact, that's because there's not much innovation happening...except in a few projects (i would venture to say that ecf is one of those innovative projects...but even that's not without a struggle). certainly, i think that most would agree that not much innovation is happening at the core of eclipse (eclipse platform, ui, equinox, etc).

the tragedy of the commons makes clear that the ef's model for ongoing participation (that the bulk of the actual code development is done by corporate *dues-paying* members) is not sustainable. and i don't happen to think that a recovery of the crappy economy will 'fix' this problem. consequently, *all* of the community will suffer terribly if eclipse (the commons) becomes irrelevant/is no longer innovative...and that suggests...no requires...that something should be done at the level of the system/foundation. it's pretty clear from mike's unbelievably unprofessional (imho) blog posting (i.e. calling bjorn a 'jerk') that the current foundation leadership is going to do nothing significant about this problem...but rather try to stonewall and continue on their present course. this is not good for the community, imho.

wrt the 'good side'...i think the thing to watch out for for the future is osgi/equinox on the (web) server. this is a real area for growth for eclipse projects imho and includes several of the eclipse rt technologies...i.e. equinox, ecf being paramount among them .

osgi on the server is taking off, i believe...particularly for the 'cloud' (another rapidly-getting-a-lot-of-hype thing for 2010). this is where i intend to concentrate my own business efforts with my own venture.

and naturally i think ecf will have a lot of important stuff to say...e.g. remote services (what was distributed osgi), as well as support for google wave. that's an exciting trend/technology as well, i think (google wave).

oisin hurley, chairman of eclipsecon 2010 program commitee

it's been a busy year at eclipse. the top five things for me, and this is a personal view, are, in no particular order:

1. tuning of the release train - this year the release train is humming along on much more solid tracks than it has hadin previous years. what's more, the requirements of the big train build has actively driven some of the technology in the community, resulting in the buckminster aggregator for constructing p2 repositories from multiple sources and lots of bug fixes.
2. this is related to (1) - there's been an increased focus on how to build eclipse plug-ins and projects, and lots of activity thrown at getting the issues out on the table and simplifying for developers. with efforts like athena, buckminster and b3 within the community, and efforts like tycho happening outside the community, it looks like there are very strong commitments to solving developer build issues.
3. the relentless rise of modeling, specifically the big push by the xtext guys. i'm a developer, and i hate mousing around the screen drawing pictures. i have ten digits, most of which take part in efficiently banging out keyboard switches and chords. this is why i think a text-based approach for modeling, in effect creating dsls to solve particular issues, is cool. couple this with enough smarts to automatically create a text editor with command completion and it is full of win.
4. when you develop for eclipse, you make lots of plug-ins. these are awkward things, since there are times when the subtleties of osgi can deceive the normal human brain. there are other occasions when it's too easy for a developer to break api by accident. what saves us all in this case? the pde tools. these are really getting good at this stage and are the only way to go for manifest-first creation of osgi bundles, imho.
5. finally, and this is more of a personal thing for me, the soa tools project has received an injection of new life with the creation of the soa initiative and proposals of the mangrove soa model and ebam. it looks like there will be more coming in there soon and, of course, there is the soa epp package!

thanks to everyone for taking the time to provide their thoughts on 2009.

don't forget that nominations for the eclipse community awards is now open. if you feel that an application, project or individual has made a signifigant contribution to the eclipse community, now is you chance to make your voice heard.