FOSDEM 2008 (recap)

I don’t know why I like FOSDEM that much. Maybe this year it was because of the nice weather, maybe it was because a number of IRC friends and colleagues came from other countries, maybe it was because this year I was able to do a lightning talk on Kettle. (**)

Some of the rooms I wanted to go to on Sunday were completely crammed with people and because of the heat radiating from some of them, I was not totally satisfied. Fortunately, an excellent presentation from fellow Belgian Tom Bayens saved the day. jBPM was always something I wanted to play with and gain more knowledge about and this certainly did the trick from a high level.

I did get the feeling that although Kettle is used on most continents of the planet, there are still certain categories of people that we fail to address.

When I sent in my proposal for the lightning talk, the organizers had not heard about Business Intelligence before, let alone specialized stuff like ETL. That is hardly their fault, don’t get me wrong. What is surprising is that tools like Kettle are not penetrating the walls of the academic world (at least not over here in Belgium). Is the gap between the academic world and business really as big as the one between ICT and business?

More important, what do we (OSBI) need to do to get a developer room next time? We had speakers and developers committed and ready to come to FOSDEM 2008 from virtually all Open Source BI players out there. It would have been a historic event, an unprecedented coming together or competitors, a total community of hundreds of developers and potentially hundreds of thousands of users united in one place. Bummer! Next year I’m filing another request. Maybe in a few years and with some luck, we can bridge a few gaps.

For the professors, assistants and students out there: think of all the fun thesis you can make on just the parallel threading and MPP engines in Kettle. What’s that you say? Oh well, at least it’s been a lot of fun to me!  If it’s too easy, pick harder topics, like finding ways to parallelize data mining algorithms.

Until next time,

Matt

(**) The Grand Market and Belgian beer probably has something to do with it as well.

FOSDEM 2008

Dear Kettle fans,

In a few weeks (23-24th) it’s FOSDEM time again. Not only will there be a number of Pentaho and Kettle fans attending, I will also be giving a lightning talk on Saturday on the topic of Kettle. I saw a few of these last year and they are quite brutal: you get 15 minutes to explain as much as you can on your subject and then an alarm clock goes off stopping you dead in your track. As such, some timing & testing will be required for this presentation later this week.

I intend to give a short history of Kettle, the current status, what we’re working on and a few minutes of demo.
Since this is a “Free Open Source Developer Meeting”, I’m going to focus technically, although it should be interesting for all open source data integration fans.

If you want to meet the gang, have a chat or a beer, feel free to contact me and we’ll get together.

Until next time,

Matt

Help tips

Dear Kettle fans,

We have to keep thinking about ways to improve usability of our software.  Over the last year we’ve been knee-deep into hard-core API design work of Kettle and now it’s time to put usability back at the top of the agenda.

One of the first things that hits you as a new user is that often you don’t know how to create a new hop between 2 steps.  Well, why don’t we tell these new users how to do it?  Well we do, however, the obnoxious tips dialogs is one of the first things that people click away apparently.

OK, so we tried, now let’s try harder.  Here’s what I had in mind: a user drops a second step onto the canvas and voila, a tool tip appears…

The advantage of using these tool tips is that the second your mouse leaves the tool tip area, or if you click anything or key in anything, the help tip is gone.  As such, it doesn’t slow you down at all.

What other ideas do you have?  Let us know and we’ll hack it in!

Until next time,

Matt