Gáspár Nagy on software

coach, trainer and bdd addict, creator of SpecFlow

Craft of Budapest (conference summary)

by Gáspár on April 30, 2014

Last week I attended Craft 2014 conference in Budapest. Here is my short summary about it. This was the first time the Craft conference was organized, and I think it was a success. Being able to bring so many good speakers and 900 attendees (380 of them from outside of Hungary) is really a good result, so I hope it will continue this way also in the next years. Budapest was a white spot on the map of international developer conferences, but with Craft we made a huge leap to change this. There were some smaller organizational issues like not enough time between the sessions or problems with the internet connection, but these were rather small inconveniences that did not change my overall good impression. The Q&A part of the sessions was done through sli.do. To be honest, I like the personal aspect of asking questions directly by the audience and not via such a technical channel, but I have to admit that the standard way also has some flaws (eg. the “please repeat the question” problem). Collecting the questions through the internet puts an extra load on the WiFi, and especially in the main room, the network became pretty unreliable, so I even gave up trying to ask or vote for questions. So maybe that’s the future of conference Q&A, but I have to get used to it. imageThe venue is the new gem in Budapest’s architectural diversity – a nice combination of the classic factory-style with modernity. Maybe it’s not ideal for such big conferences (it does not have enough proper session rooms), but was a good choice. Regarding the sessions, I got enough inspiring input and participated in interesting  discussions in these two days, however, of course there were a few sessions I did not like so much.  You can find my comments below. The sessions were recorded and are available for watching online, so you can also check them.

Bodil Stokke – Programming, Only Better

imageBodil’s keynote was funny. There is a trend to say that the problems of software development are rooted in OOP /shared state and functional programming (FP) is praised as a rescuer. This talk followed this trend, but went even further by saying that logic programming languages are even better than FP, because they do not only eliminate shared state but the complexity of the program flow as well.  I think she is right in this conclusion. I also loved Prolog, when I played with it at the university. But it’s like 42… we  know the answer only. The problems people have with FP (let’s not even consider more esoteric languages, like the logic programming languages) is that they lack those important patterns and practices that would make it usable in the daily business, One interesting thing in the session was that Bodil used a coding platform embedded into the presentation. This was really cool. Must check.

Jonas Bonér – Going Reactive: Event-Driven, Scalable, Resilient & Responsive Systems

I think the session was a good summary of available concurrent programming principles and models. I’m currently reading the “Seven Concurrency Models in Seven Weeks” book by Paul Butcher that is pretty much about the same topic.  Both in the book and in the session I am missing the guidelines when to use what.

Dan North – Jackstones: The Journey to Mastery

imageDan North’s session was one of my highlights on the conference. I think this talk was one of his best and it was very well in line with the context of this conference. This was not a talk that would change your work from the next day on, but the kind that comes back to your mind even a few days later and may kick off interesting discussions with your colleagues. It did not try to give answers, but raised questions to think about. I tried to summarize the session to my wife (mastery is different for a piano player, an ice hockey player and a soldier – and we are all like a prepared paper for folding origami Jackstones), but I have not manage to seize the essence yet. Take a calm evening and watch it (with a beer maybe)! (At the time of writing, this talk is one of those few that is not online yet, but here is a recording from YOW!.)

Amber Case – Getting Things Done at Scale

imageThis session was the last on the first day, I was already tired, so I went in without any preparation or expectation. After a few minutes I realized that she was talking about the experience she had when her startup had been bought by esri. I got concerned, because I don’t have a startup to sell (well, no one at all), and I’m not too interested in hearing gossips and networking stories (at least not here). But then I was absolutely positively surprised, Amber was telling her own story, but the outcome is important to anyone: it is possible to make changes even in larger organizations. You have to have clear goals, prepare well for the transition, decide on where it is possible to make compromises and actively deal with people and cultural issues. And of course you need to do this step by step – if you can change 20% in a year, that is already a big achievement, she said.

Chad Fowler – McDonalds, Six Sigma, and Offshore Outsourcing: Unexpected Sources of Insight

imageI liked this one too. It was funny, provocative, but still gave enough for thinking. In some sense it was completely the opposite of Dan North’s talk. The first part when Chad talked about his background and the Six Sigma was a little bit boring, but the second part when he talked about the analogy of McDonalds and the franchise business in software development was interesting. Also when he applied it to outsourcing, he had some very interesting arguments.


Markus Völter – The art of building tools – A language engineering perspective

This was a very good pragmatic technical talk., The topic was language workbenches, particularly JetBrains MPS, which Markus is clearly a master of. The common problem with language workbenches or language oriented programming is that it is hard to get an overview on the usage for usual enterprise applications. This talk raised many real life examples, where a custom DSL could serve well. My favourite he showed was  how to use DSLs for defining insurances. Very inspiring.

Gerard Meszaros – Find the Right Abstraction Level for Your Tests

imageMeszaros is the grand old man of testing. His book (xUnit Test Patterns) is referred to in many testing-related books or articles. I haven’t seen any of his sessions yet, so I was quite excited. It was surprising… positively. First of all, Gerard is not old at all (or at least much younger than I expected) and he is not the guy who once wrote a book and wants to live from that in the rest of his life, He said that even though testing is an important part of his professional life, he  also deals with many other things, too. The first part of the session was a categorization of different kind of tests we might need, The testing pyramid was only a starting point, he went more deeply into the problem of the granularity of the different tests. This is an interesting topic indeed, but I have to watch it once more, because at that Friday I could not absorb it fully. The second half was about to show these in practice (on slides): he started from a detailed and verbose unit tests and ended up in something very similar that we use in BDD scenarios: clear separation of functional and technical aspects; hiding non-relevant information; using the Given/When/Then terms. Was interesting!

Comments are closed.