Read my book

I wrote books about Webpack and React. Check them out!

Monday, June 25, 2012

Thoughts on Devaamo Summit 2012

I visited this year's edition of Devaamo Summit 2012 around a week ago. Devaamo Summit is an event focusing on various open themes held at Tampere (Manchester of Finland). There were two tracks this year: open web & data and open mobile & hardware.

This was the third time it was organized. As far as I understood it was previously focused on MeeGo but as you might know, things didn't go so well with that project. Even this time there were some issues (Nokia just laid of 10k people, 3.7k of which in Finland, a big deal locally).

The event lasted for two days. The first day, really short one, had some lightning talks. The second day contained the actual beef. The topics of the presentations were well chosen and there was probably something interesting available for everyone. In addition there were a few booths around and a hacking challenge even.

Somewhere near Tampella factories

Thoughts on Lightning Talks

I didn't take for any specific notes for the lightning talks so going by memory here. The format was standard 5 min (3 mins for presentation, 2 mins for questions). Sadly this was not enforced strongly enough so some talks kept on meandering around. Hopefully they can deal with this better in the next year's event.

Topicwise it was a mixed bag. There were some business related talks and more technical ones. I'm going to elaborate a bit on some of the technical ones since I found those most interesting:
  • Lauri Paimen demonstrated his implementation of QML in the web browser. QML is a declarative language familiar from the Qt framework. It looks a bit like JSON but allows you to describe some basic logic. Note that Paimen's QMLWEB is still in quite an early phase (proof of a concept). You might want to investigate the presentation slides (implemented using QML of course) to get into it better.
  • Otto Kekäläinen spoke about VALO-CD. Essentially it's a CD (do'h) containing a collection of open source software. It has been meant to function as a gateway for people stuck on Windows. Get them use the apps first and hook them on Linux later. :)
  • Tero Piirainen introduced ITS Factory and Tampere's plans of opening some traffic data. I really hope this catches on in Finland.
  • Antti-Jussi Kovalainen discussed his Beatstream project. Essentially that is a service that makes it possible for you to stream audio from your home computer to anywhere using a web browser. The source is available at GitHub of course.

Thoughts on Presentations

Given there were two tracks I had to make some tough choices about which presentations to see. I focused mainly on the open web & data track, though I saw a few on the other track as well. Overall the quality of the presentations was good.

I have to admit some felt a bit underwhelming given they were clearly aimed for people not so aware of open solutions. I also felt that sometimes the thirty minute slot given for each talk was not enough for some. It is simply not possible to delve very deep in such a short frame of time.

Acceptance Test Driven Development for Web Apps

Eemeli Kantola of Futurice spoke about Acceptance Test Driven Development for Web Apps. The presentation remained on quite a high level. It seems to me ATDD might fit well in an environment where you have a good understanding of various requirements. Mapping these to tests using ATDD feels like a natural approach. That way you get something less vulnerable to regressions and will be able to maintain a sustainable pace.

This was one of those talks that were hindered by the fixed 30 min slot. It would have been really interesting to see how it works out in practice. Fortunately Eemeli has prepared a demo project that ought to give some idea.

Best Practices for HTML5 Servers

Otto Kekäläinen of Seravo went through some Best Practices for HTML5 Servers. Overall the talk was somewhat oriented towards beginners. If you've ever dealt with servers, you probably know this stuff already. It would have been somewhat interesting to have some kind of comparison of old skool servers (self-hosted) and cloud based solutions (Heroku, OpenShift) in the presentation.

Decoupling Content Management with CreateJS

Henri Bergius of Nemein discussed about Decoupling Content Management with Create.JS. He aims to change the way we think about CMS. Instead of treating it as a monolithic entity, he proposed a way to decouple it. Essentially he splits the whole in three parts: web editing tool (content manipulation, publishing), web framework (business logic, serving the content), content repository (database and related ops). The main advantage of this is that you can use the CMS you want with the UI you want.

I believe this sort of approach might have some kind of a future ahead of it. It definitely makes sense architecture-wise. Whether it works out in practice and gains some popularity remains to be seen.

This is another presentation that would have benefitted from a longer slot (60 mins even). The slides I linked to seem to be from a longer talk.

Trends and Visions in 3D Printing Community - Survey Results

Jarkko Moilanen's "Trends and Visions in 3D Printing Community - Survey Results" felt like a blast from the academic past, and to be honest that's what it was. I guess the main thing I got out of this was that an average person dabbling with 3D printing is a European 35 old male with university background. Oh, and the community is somewhat fragmented. The people dealing with 3D printing identify with the maker movement.

Based on the results it seems 3D printing is on its way to further commercial usage. Currently the main problems on hobbyist side seem to be accuracy, speed and material limitations (pretty much just plastic for now). I think B9Creator might solve the first two. That makes the current consumer level printers look like ancient technology (consider the leap between NES and SNES or PS and PS2).

Intro to 3D Parametric Design

Randall Arnold of Tribal Method gave an Intro to 3D Parametric Design. Even though the presentation was interesting it felt a bit underwhelming to me. Sure, it gave a nice idea of the paradigm. Still, it would have been nice to get a bit more into the detail.

Parametric modeling is just one of the basic ways you can use to produce 3D models. It is especially suitable for producing objects to print given it's easy to end up with something that's manifold (ie. the object doesn't have "holes"). Boolean and formative (mesh) modeling are commonly used as well. Each approach has its sides.

In parametric modeling you describe your object as a set of operations. In essence you are constructing a graph of operations. In addition you will define certain kind of constants and parameters. Suppose you are modeling a hex bolt this way. You might want to treat the dimensions of the upper part (the one with hexagon) as constants while the height of the screwy part could vary. This allows you to produce a bolt with given height quite easily.

I prefer mesh modeling myself. Of course it isn't mathematically as pure as boolean or parametric approaches. You can go from boolean or parametric model to mesh easily but not the other way. There are some hacks (subdivision surfaces, modifiers etc.) that make the situation fuzzier, though.

This is the kind of stuff I would've like to see in the presentation. As it was I really didn't learn a lot from it. Again, 30 minutes was way too short a slot to go through comprehensively a topic such as this.

Going into Cloud - New Conceptual Model for Personal Information

Jussi Alanen's presentation underlined the importance of having a proper access to personal information scattered around the cloud. According to him (and Gartner) "personal cloud" is going to replace PC by 2014. These guys are taking the prediction seriously and plan to do something about it.

Currently the data has been scattered all around various services (Facebook, Twitter, Google, Flickr, you name it). Just having access to the data is not enough. A sensible way to organize it is required.

If I understood correctly they aim to combine data from multiple sources using metadata. This way you could construct a certain kind of view that would give you access to data related to some specific pivot. Consider that you participate in some event (like Devaamo Summit). You might want to have access just to photos and contacts related to it.

This kind of scheme might make it possible to implement new kind of services. There are plenty of mashups available already. This approach would probably take it to the next level.

It will be interesting to see what they will be able to achieve based on this premise. If you are interested in this sort of stuff, you might want to check out their site named "Bring my information home!".

Thoughts on Other Presentations

I caught a glance of Pekka Sarkola's "OpenStreetMap and Free Open Geodata Are Great Opportunities". Something interesting happened a while ago at May. Most of Finnish geodata available was published with a permissive license. This is a huge deal. It wouldn't surprise me if we saw various enterprises tapping into the data.

Unfortunately "Hack Avoin Ministeriö" by Aleksi Rossi was cancelled at the last moment (the speaker didn't show up). Avoin ministeriö (open ministry) aims to apparently make it easier for the citizens to take part in the legislation process. It seems like a step to the right direction.


Overall Devaamo Summit was a nice little event. If it is organized next year and I don't have any other appointments I'll likely visit there again. Thanks to the organizers and speakers. Special mention goes to Futurice and their stand for the swag. :)