The First 100 Days

The term of the new class of committer representatives officially started on April 1. To their credit, the reps had already established several channels of communication before the term even started.

To get the party started, I posted a few questions on the committer reps newsgroup today asking about their action plan for the near future.

Of course, for them to have an agenda that represents the committers' needs, we all have to chime in to provide input and feedback.

In an effort to initiate a dialogue, I asked the following questions:

1. What is the single most important issue that you would like to push for and bring up for debate in the first 100 days of your term (ie. in the first face-to-face board meeting)? Why do you feel it's so important?
2. If issue stated in question #1 was still unresolved at the end of the year, would you consider that to be a failed term? More generally, how do we (and you) quantitate success of a term?
3. Now that you have had some orientation by the Eclipse Foundation as to what role you are able to play, are there items that you listed on your nomination vision statement that you think is out of the scope of your responsibilities? Another way to ask the question would be: is the position so far what you thought it was going to be or were you in for a surprise?
4. Being a committer rep on the Board of Directors is a coveted position. You get to add 'Director of Eclipse Foundation' to your LinkedIn profile, you get to add it to your EclipseCon biography and you get a chance to start every conversation by saying "Now that I am on the board of directors..." and annoy the heck out of your co-workers. So in short, it must be a lot of fun :).Yet for some reason, it's clear that not too many committers care about that. The vote turnout is low and the nomination field was not very crowded. Why do you think that is and do you plan to do anything about it?

This should make for an interesting thread, but let's keep it in the newsgroup.

Pound For Pound

In a recent interview, in reference to the Eclipse Platform team, Mike Milinkovich was quoted as saying: "...pound for pound, I think it's one of the best software engineering teams on the planet."

While I certainly agree with Mike that we, the Platform team, have collectively gained some weight over the past few years (as it was evident at EclipseCon2007 last week), I still found it inappropriate for him to announce that fact in a press release.

One could argue that the Eclipse Foundation is partly responsible for that. At EclipseCon, they threw an awesome party for the committers and the community where almost $3,000 worth of carbs was consumed past 10 pm. According to the genius Dr Phil, you should not eat past 7 pm.

Also, have you heard of those things called 'cookies'? They are delicious!

I, for one, am really trying. I work out every day (as Wayne would attest), have gone completely off dairy this week and plan to eat a plum for breakfast every morning starting next Monday.

One would hope that the Eclipse Foundation would stand by us in thinness and in fat.

EclipseCon 2007 Post-Mortem

The Best Conference Ever
I have been to the Santa Clara Covention Center for non-EclipseCon and more expensive events (e.g., SDWest) and let me tell you, it's not the same. During EclipseCon week, everything is better and everything runs without a glitch. It is a smooth operation. A big thank you to the conference organizers for an amazing job.

OSGi For Dummies
It was a great idea to merge the OSGi developer conference with EclipseCon and I hope this trend continues. Generally though, I felt that a significant number of the conference attendees did not know how the two technologies related to each other.

It may be a good idea to have a plenary talk next year to educate the masses about OSGi and Eclipse. My choice to present such a plenary would be BJ Hargrave. He articulates those OSGi concepts perfectly and his voice projects very well. He also ought to write an "OSGi for Dummies' book.

A Panel without a Panel, a Point without a Counterpoint
One of the sessions I was most curious about was the What sucks about Eclipse panel. As much as I like healthy debates, I was afraid it would quickly turn into a complaint session about insignificant problems, rather than tackle big issues.

In fairness, the moderator was good, but I think the session was destined to fail before it even started. According to the abstract of this session, "the format of this panel is a little non-traditional. The audience are the panelists". I suspect it was difficult to assemble committers or project leads for the panel if the premise was to be ambushed by the community :)

So the audience brought up some good points, but there was no one in attendance that was knowledgeable enough to respond. So it left us all empty, and after a while the whole thing just felt awkward.

I blame Bjorn and Richard for voting for and accepting that submission. (I am kidding, please don't take the bait ;)

Web Track != WTP; Test and Performance Track != TPTP
The web development track was very heavy on the WTP. Similarly, the Test and Performance track seemed heavy on the TPTP content. I would have liked to see more variety in those two tracks as I am sure there is a lot of Web and test/performance-related work going on in the Eclipse space that is not directly coming out of these two projects.

By comparison, other tracks such as the C/C++, Java, and RCP tracks, had a very well-balanced program.

Tutorials - A Popularity Contest
I suggested to the program committee that, when it comes to tutorials, they should accept submissions based on what sells, not evenly distribute across tracks. A good, but certainly not the only, indicator would be to look at the big hits/misses of EclipseCon2006. Unfortunately, either no one listened, or I spoke up too late.

So we ended up with some tutorials about obscure topics that did not do well this year. In fact, some tutorials that did not do well in 2006 came back to the 2007 schedule with little or no change. There was no miraculous jump in their attendance count this year either.

It was a bit sad to walk outside a huge room and see less than 10 people in attendance. Of course, there was no such problem in the modeling, RCP and embedded tracks. They were packed.

I think some people on the program committee ought to do some soul searching (perhaps travel to Tibet, if it helps) and come up with a plan to not let this problem happen again.

Onward and Upward
EclipseCon2007 was a great event. Looking forward to next year.

Eclipse on Mac - Part Deux

The most popular post on my blog is by far the recent Eclipse on Mac discussion.

Yesterday, as a follow-up, Ken Gilmer posted an itemized list of things that bother him about running Eclipse on Mac, complete with screenshots.

Personally, I tend to stay away from any operating system that does not come out of the state of Washington. But for all you Mac users out there, take a look at Ken's post and let's have another interesting thread going :)

The Greatest Schism of All

I have recently voiced my dismay about EclipseZone running Netbeans ads.

While the criticism is, of course, justified, I may have been a little too hard on our friends over at EclipseZone.

After all, how would I, the community and the Eclipse Foundation react if Sun sponsored EclipseCon?

I am nothing if not reflective, introspective and contemplative; so here is how I imagine such a scenario would play out:

"The Eclipse Foundation director would announce the decision of Sun to sponsor EclipseCon on his blog. He would welcome Sun with open arms and reaffirm the respect we have for Netbeans and how competition has been healthy and good for both camps.

A Sun evangelist, most likely someone from Great Britain, would then post a comment on the director's blog thanking him for his kind words.

The story then gets picked up by ZDNet, where Drama lives. One can expect an over-the-top headline like 'Sun has made a major step toward healing the schism in the Java community, which is the biggest schism that humanity has seen since the schism of 1054."

It will be one big love fest.

Fast forward a week or two...

We wake up one day and we read some earth-shattering news that some random guy from Sioux City, Iowa has just switched back to Eclipse after trying the 3.3M5a milestone, or some single mother from Grand Rapids, Michigan just tried Matisse and liked what she saw.

At which point, bloggers on Planet Eclipse and Planet Netbeans pick up the groundbreaking stories and start interpreting those "honest evaluations". Enter EclipseZone with one of their inflammatory, yet well-intentioned, posts and we are back to square one.

The love fest is over."

Sounds about right? :)

Panel Best Practices

In the second installment of my "I am so filled with the joy of EclipseCon that I cannot contain it" series, I would like to discuss how much I am looking forward to attend this year's line-up of panels.

I am a big fan of the panel concept, but up until now I have never attended a panel that lived up to my expectations at any conference (EclipseCon or otherwise). I am hoping this year would be the exception.

So I compiled a list of factors that have contributed to my negative experience with panels and turned it into a list of Panel Best Practices. After all, if life gives you gator, make gatorade.

It is clearly not intended to be a Letterman-esque top ten list, particularly since it's not funny. I just happened to have 10 thoughts on the topic:

1. Limit the number of panelists to a maximum of 5. That should be sufficient to start and sustain a 45-minute discussion where everybody gets decent air time. No need to assemble a soccer team on stage.
2. When assembling a panel, make sure to invite people who are qualified to talk about the subject matter. At times last year, I kept thinking to myself: "Who are these people on stage and what makes them so qualified to talk about this topic?"
3. Panelists should keep their introductions brief. Stating the name and affiliation is sufficient. Panelist, it would be best if you saved interesting facts about yourself, such as your aversion to foods that start with the letter Q, for the tell-all autobiography.
4. Panelists, who are competitors in real life, tend to take jabs at each other, then pretend it was joke and say 'Oh, I'm just kidding...'. Don't do that. It's not funny.
5. Panelists, who are friends in real life, occasionally (and awkwardly) start picking on one of their own in the name of comedy. Don't do that. It's not funny.
6. It is certifiably lame to still be making Brangelina and/or Netbeans girls jokes and references.
7. The moderator is the only guy on stage who is allowed to be funny -- but not too funny. People are not there to see a stand-up act.
8. The audience Q & A portion should be controlled tightly by the moderator. It only takes one persistent audience member to ask an off-topic question and before you know it, an RCP Best Practices panel turns into a 30-minute discussion of the SWT_AWT bridge on Mac.
9. The audience should lower their expectations. A 45-minute discussion will not result in closure on the open issues being discussed.
10. A 45-minute discussion may not result in closure, but it would be nice if it resulted in an action plan or a concensus on a direction going forward. Otherwise, we would be having the same conversation next year.

When Jupiter Aligns with Mars

For the past few days, I have been so filled with the joy of EclipseCon that I cannot contain it.

However, March seems like an inconvenient time to have the Eclipse event of the year --particularly for the Eclipse developers:

1. EclipseCon preparation time takes away from development time when we could be fixing bugs and adding new features. Add onto that the week of EclipseCon itself where development practically comes to a halt.
2. Teams try to cram as much new functionality as possible into M5 to showcase it at the conference. Quality of some of these features may be somewhat questionable, given that there was not enough time to test.

Wouldn't July, for example, be a better time to hold the conference? It would be right after the release of the Eclipse train; so it is a time when the developers are unwinding, and the quality/volume of new features is at its peak.

As for why EclipseCon is currently held in March every year, here is my conjecture:
"Given our apparent (and inexplicable ;-) fascination with Jupiter, March is probably the month when Jupiter aligns with Mars or something. So we must hold EclipseCon at that time. We must!"

OSGi - the miniseries

Over the past few days, Neil Bartlett has published three installments in his OSGi tutorial series:

1. Getting started with OSGi: Your first bundle
2. Getting started with OSGi: Interacting with the Framework
3. A Comparison of Eclipse extensions and OSGi services

With M5 behind us, I finally had time to read the miniseries. Here is my review:
"Bartlett's OSGi miniseries is the best miniseries since Alex Haley's Roots!"

I encourage every developer to make some time to read these articles. They are a great read and you will learn a lot. The third one in the series is my favourite so far.

By comparison, if you had spent three hours watching the Grammys on Sunday night, the only life lesson you would have learned is that "Hips don't lie" :)

Unfortunately, Neil's excellent series is hosted on EclipseZone, so you will most likely be greeted with a Netbeans ad at the top of the page when you go read it.

When will this madness stop? Is it hurting Eclipse? No. But it is certainly not helping either.

Eclipse on Mac

Today I read a blog post claiming that Eclipse has little, if any, following among Mac users.

That claim took me by surprise. So as a litmus test, I did a bugzilla query to see how many bugs were reported on Mac and I got 3550 bug reports. Not too bad.

Where is the data that backs up or refutes this claim?

It would be interesting if the webmaster could provide Mac download statistics and/or if the marketing wing of the Eclipse Foundation could provide one of them pie charts showing Eclipse's mind share on Mac.

No Thanks

One of the nice things about working in open source is that you interact directly with your customers (ie. the community) in an open forum via bug reports. When we, the committers, resolve a bug as fixed, it gives us a sense of joy that we made some user out there happy.

While implementing a highly-anticipated feature is a reward in itself, it would be nice to hear a thank-you every once in a while. Unfortunately, very rarely do we hear that thank you from users.

72 people have voted for a Text Drag and Drop feature. When it was finally implemented, not one thank you.

Oh well, such is life...

Thank you, Dani!

Foul!

Am I the only one who finds it wildly inappropriate for EclipseZone to be running Netbeans 5.5 ads?

New Year, New Forms

Je love Forms, je can't live without Forms, and we are starting off the new year with the long-anticipated Forms facelift for the Europa release.

Visual designer extraordinaire Kim Peter has just posted new screenshots showing the new vision. New enhancements and features include: color refinements, header/section title updates, one and two-row header variations, drag-and-drop to and from the header, error validations and much much more.

We are now very busy trying to implement that vision and get it ready for the masses by the end of the 3.3 M5 milestone.

Here is a taste of the new Forms on Vista:


I am certainly going to nominate Kim for Top Contributor for being a champion of cool and esthetically pleasing UIs for years.

Incidentally, if you want to learn more, there is an excellent Forms tutorial at EclipseCon 2007, where all aspects of Forms, old and new, will be covered.

Goodbye Janek, Goodbye Roach

After 20 months of amazing productivity and a long list of noteworthy achievements, today was sadly Janek Lasocki-Biczysko's last day on the PDE team.

Instead of joining the work force, Janek decided to take time off after graduation to travel and then go back to school for more knowledge.

In his fourth year at the University of Toronto, Janek has been working on a project called 'Interfacing with video games through a third party'. Janek chose the Madagascar hissing cockroach as the player in his experiment. The Madagacar cockroach is a bonafide movie star, having appeared in the movie Bug and all. So naturally, one would expect the roach to be difficult to work with. It is like trying to direct a demanding diva like Jennifer Lopez.

But Janek stuck by his star. They had their ups and downs over the past few months and the project finally came to an end.

Enjoy the world premiere of Janek's roach video.

All the best, Janek.

The Mother of All Panels

Ed Burnette seems to have suspended his obsession with Google and GWT, for now, and has become active once again in the Eclipse world. Welcome back, Ed :)

The latest cause Ed has been championing is the official expansion of the PDE mandate. Of course, he is not the first one to do so, but he is certainly the most skilled at mobilizing the base. Observe how within a couple of hours, Dave Orme was already talking escalating the matter to the Board of Directors. Even Karl Rove or James Carville are not as good as Ed.

Of course, not everybody is as much of a PDE enthusiast as Ed or Dave. Old-school OSGi developers are resisting to accept the wonderful world of PDE and are sticking to their old tools (ie. sticks and stones). Some Maven enthusiasts think that PDE places too many restrictions on their worfkflow.

Are we right or are they wrong? How can we come together to make Eclipse the best development environment for all plug-in/bundle/module/component programmers?

There is no better forum to address all these timely topics than an OSGi panel at EclipseCon 2007. So I submitted a proposal.

Our competition in the OSGi track is a panel about the "Future of OSGi". Give me a break. Sounds pretty boring. It will probably be a distinguished panel of OSGi-philes patting each other on the back and telling you that OSGi is "the best framework in the whole wide world, blah blah blah..." Who wants to see that? :)

Before we start talking about the "future of OSGi", let's talk about the present.

So if you like drama, if you like heated debates, vote for the fun submission. You will get your money's worth and more. People paid to see a fight on that stage, and I say let's give the people what they paid to see.

I am not going to reveal my most effective moves, but let's just say I used to be a huge 'Saved By the Bell' fan as a young teen.

A Chicken in Every Pot

Last week, Mike Milinkovich sent an email to the eclipse.org committers mailing list announcing important dates for the upcoming 2007 elections of the Eclipse Foundation Board of Directors.

There are four coveted seats on the board of directors reserved for committer representatives to be elected by their peers for a one-year term starting April 1, 2007. This year, the nominations open Dec 17 and close Jan 13. Voting takes place from Feb 7 through Feb 21, and the winners are announced on Feb 28.

So I thought this may be a good time to take a look back at the accomplishments of the current elected representatives over the past year. Wait a minute. We can't, or at least I can't.

With all due respect, I, for one, have no idea what they did on behalf of committers. Not to say that they did not do anything, it is just that the communication was pretty much non-existent. As a result, I am not very well-informed.

There are so many ways and channels already available to the elected representatives (and us all) to communicate effectively with the rest of the committers. Unfortunately, these channels were not utilized well, if at all.

It would have been useful, for example, if the representatives published an interim report after each board meeting to the committers mailing list outlining issues discussed and conclusions reached. Blogging is another complementary and useful outlet to keep their constituents informed.

I also do not recall the representatives soliciting for input from the committers in a public forum on what issues are most important to them that they can bring it up to the board of directors. How do they know they are really representing our interests?

I hope the next class of freshmen and/or incumbent committer representatives take the time to communicate effectively with the rest of the committers and community.

Actually, what I am asking for is nothing new or groundbreaking. Bjorn used to do that when he was an elected member a few years ago, and the webmaster keeps us very well-informed on the issues that affect us the most.

We all appreciate that, and let's have more of it.

Incidentally, the title of this post is, of course, a reference to the famous 'A Chicken in Every Pot and a Car in Every Garage' slogan of Herbert Hoover's 1928 presidential campaign during the depression.

Trivia: The Simpsons had an episode about an election with a title that is very similar to that slogan. what was the title?

The Elephant in the Room

In the last installment of my EclipseCon2007 series, I would like to discuss the elephant in the room.

Let me start by quoting Jeff McAffer, a member of the EclipseCon2007 program committee:
"The job of the PC is to put together a program that attracts and informs the attendees. Lots of submissions in an area indicates a lot of interest in presenting on a particular topic, not necessarily in hearing about a particular topic.... Bu seriously, it is about the attendees, not the presenters." Very well articulated comment.

Now let's check the facts on the ground (i.e. the allocation matrix) to see how we are doing on the "let's give the people what they want" front.

The matrix seems to be very fairly distributed across tracks, suggesting that the program committee believes that the attendees are going to be equally interested in all Eclipse projects. Do the organizers really think that all tracks have the same appeal or are they trying to be politically correct about the allocations? Perhaps the answer is in point #1 in Bjorn's response, i.e. egos and politics play a part. The fact that the track chairs are project PMC's probably does not help when it comes to objectivity on how popular a track really is.

An objective decider, who has the task of coming up with an allocation matrix, should take a look at the EclipseCon2006 numbers to get a realistic and unbiased perspective on what sells and what does not, and use that information as a guide for the 2007 allocations.

In 2006, RCP/Platform sold well. WTP's numbers were very low. Modeling and GEF did well. CDT and DTP tutorials had a relatively low turnout, etc. etc.

Given last year's numbers, it is therefore incomprehensible to me why, for example, WTP and RCP should have the same allocations this year, or why the number of DTP tutorial slots was tripled.

Whichever subject matter did not sell well last year should have its tutorial slots decreased. If something sold well, let's have more of it. To the victor go the spoils.

I know those Santa Clara ballrooms fit up to 113 people, but let's not push it to the limit. It is better to have two tutorials of reasonable size on a popular subject, instead of trying to squeeze as many as we can into one.

If the RCP, Fundamentals or modeling tracks end up having many more tutorials than another technical track with this approach, such is life (as Bjorn would say). It's all about filling those tutorial seats and giving the attendees what they want. It's not about a politically correct "allocation matrix of fairness."

Where Are They Now?

I am looking forward to the Toronto edition of the Eclipse 5th birthday party. In fact, I am planning my whole week around it. I got a haircut. I am buying new clothes this weekend and I am getting my car washed on Tuesday.

As for party entertainment, I originally was hoping that Arthur would hire a ventriloquist. But after I saw the list of people who will be in attendance, I think we will have enough entertainment without the need for talking wooden puppets.

It seems that all past and present Install/Update committers will be in attendance. It will be fun to sit around to see what they are up to now and reminisce about the evolution of that component through the ages. Too bad Chris won't be there to ambush them :)

True dat, peeps!

Bringing Sexy Back

In an effort to help Justin Timberlake bring sexy back, we are planning to give Forms a facelift in the 3.3 cycle and showcase the new look in the PDE editors.

Personally, I love Forms more than life itself. If you love Forms as much as I do, cc yourself on bug 152912 as the new prototypes will be posted there in the next little while.

A question remains though: Have Forms ever loved me back?

Go ahead, be gone with it!

Breaking News/Gossip

Labour Day marks the unofficial start of the new blogging season. Not to be outdone by Katie Couric, I decided to have a new format for my blog, whereby I focus on revealing breaking news in Eclipse 3.3 development.

My definition of 'breaking news' is that of CNN, so you may want to lower your expectations a bit :). Here is the first installment:

Work is under way in the Toronto offices of the Eclipse team on a new killer feature. When completed, this "feature" will help slash the Eclipse SDK download size by at least 30 MB. I do not want to give any more details at this time, as I do not want to jinx it; but if someone guesses correctly on where the 30MB can be trimmed, I will confirm the guess.

Another cool feature that we are working on is codenamed 'Command Conqueror'. It combines two of our favourite things: cool UI and the org.eclipse.ui.commands extension point.

So the 3.3 cycle is shaping up to be a very exciting one. Make sure to come to this blog to get the latest developments as they happen :)

A Talk From The Trenches

Eclipse usability is, and has always been, a very hot topic.

A working group has recently been formed to focus on usability across eclipse.org projects. There also is a very interesting thread on the eclipse.foundation newsgroup regarding the issue.

Yesterday, Julian Jones posted a historical review of usability in Eclipse containing a link to his EclipseCon 2005 presentation on Eclipse UI Best Practices. It's an informative read. Check it out.