6 Replies Latest reply on Dec 7, 2012 5:11 PM by spilon

    vCampus Live! 2012 programming hackathon -- recap & evaluation

    MvanderVeeken

      We had a great experience during vCampus Live! last week, and according to some of the feedback we got... that was your experience as well! It was great to meet a lot of you, and thank you for coming to our event in such numbers.

       

      This year, we featured 2 hackathon tracks on Day 0 (Nov. 28), a security track and a programming track. Both were very well received. In this post I would like to recap on the programming hackathon, and evaluate together with you.

       

      Recap

       

      The idea behind the hackathon was very simple: we supplied car usage data (historic, and a live car driving laps around the city), and the job of the participants was to create a valuable application that utilizes that data.

       

      A lot of people showed up for the kick-off. We explained the concept, and people started forming teams. All in all, 8 teams of 5 people were formed. Some people seemed to have dropped off before the creation of teams -- from what I got that was due to different expectations. Some people may have thought that they would be forced to stay in that room for ~20 hours and program all night . Although a lot of people did work until ~3AM, this was not at all mandatory. 

       

      In total, 5 teams made it through. 3 teams dropped off, mainly it seemed because lack of focus or progress. In some situations, some key members had to drop of due to other engagements, this quickly led to the disintegration of the team. While this is certainly possible, we would like to hear from you. Is there a particular reason why people dropped off? Is there something we could have done to set expectations better? (apart from the infrastructure, to which I'm getting at ).

       

      A total of 5 teams finished, and were able to present their applications to the judging panel. After careful deliberation, 1st to 3rd prize were chosen from the finished applications. The judging panel was very impressed with all the idea's!

       

      On Day 2, we had an ad-hoc presentation session where the teams presented their ideas. I was surprised by how many people showed up, and had interest in looking at the awesome creations and idea's of the teams. We also used this session to award the prizes.

       

      The winning teams and prizes were:

      1. Team '01' -- Microsoft Surface Tablets
      2. Driving Miss Daisy -- Microsoft Kinect for Windows sensors
      3. Xceleration Cars -- Digital Picture Frames

      We haven't received a summary of the projects from all teams. We would be very grateful if you can provide us with a short summary of what you did in the Hackathon VCL12 forum. We have recorded the presentations, and we will be posting them soon -- to give you a good idea of what the teams created!

       

       

       

      Survey

       

      After the hackathon finished, and the final products were reviewed and judged, we asked people to fill out this paper evaluation form. I wanted to share the results, and share my conclusions with everyone.

       

      7484.overall_5F00_experience.png

       

      The overall experience was rated Good to Very Good. It seems that people really found the event valuable, and also enjoyed themselves. For me this was one of the most important ratings. You can have technical incidents, but the overall experience should be good!

       

      8156.theme.png

       

      The theme, which was car usage data, was rated as Very Good to Good. While it is not a traditional use of the PI System, it is a very exciting and emerging one. The fact that we had a car driving laps around the city that gathered Live data really added to the experience. The fact that people got to use GPS data was generally thought of as exciting.

       

      0825.infrastructure.png

       

      This is the first rating that shows a very mixed message. It shows a spread between Slightly Negative/Neutral to Good/Very Good. I think this also depends on what people used the infrastructure for. Some people needed more flexibility for their projects than others. Generally speaking, I think the infrastructure is the big item we need to improve on to make the experience much better. We had a single PI System (running on Windows Azure) that was collecting the live car data, and had historic data. We also provided workstations/test servers with PI and AF installed. The issue that most people experienced was constrained security permissions. While we wanted to make sure that teams could not influence other teams data/structure, this lead to annoying situations. Also, the performance and connectivity seemed intermittent. We changed some security settings ad-hoc to accommodate people, but this takes very valuable time from the participants.

       

      We definitely need to improve on the infrastructure for a next time. While the current setup was certainly workable, it was not very pleasant. Currently I'm thinking of having a central PI server collecting data, with personalized PI Systems as slaves with a PI-to-PI connection. This makes sure that teams have total freedom over their infrastructure. Any comments and additional feedback on this is very welcome!

       

      2275.mentoring.png

       

      We had different shifts of OSIsoft people available for questions and coaching. These people were from different parts from OSIsoft (COE, Engineering, TS, etc.). People really seemed to like the level of coaching. I'm personally very happy with this result. For a next time, I'm thinking on maybe embedding an OSIsoft developer to each team, to intensify coaching.

       

      7875.time.png

       

      How people perceived the available time also shows a mixed message. I think this again depends on the projects people engaged in. Some might have been better suited for this time frame than others. Nonetheless, I think it could be a nice approach to start either early in the morning, or have the kick-off meeting the evening before, so people can get up early (or work late). This time the arrangements with the hotel made it difficult for us to start early in the morning. For a next time, this might be a better idea.

       

      3487.competition.png

       

      People generally found the competition element and attached prizes as Important. We had some cool prizes for the winners, and people seemed to care about that. While there wasn't a strong competition spirit, it does add to the spirit of creating something valuable. We had some comments that suggested having a 'sharing' session mid-hackathon, where people would share their idea's on what they were building, as to remove the 'secrecy' of what they were working on. Overall, I think the prizes added something positive.

       

      5344.venue_5F00_food_5F00_drinks.png

       

      The Venue, as well as the provided food and drinks were perceived as OK to Very Good. We had a nice room up on the 36th floor, with an inspiring view. Some suggested having bigger tables, so teams could more easily collaborate. A very good suggestion was to have whiteboards available to each team, to brainstorm their idea's and draw architecture. The Food and Drinks seemed satisfying. We had a constant stream of refreshments available outside the room during the day, the refreshments at the welcome reception during the evening, and chips and soda's during the night. Stuart Collins created waffles (with his trusty waffle iron) at 2AM, which was very cool. We had one small incident where due to mis-communication, people were not aware that the refreshments were not available during a 1,5h time period. Overall, I think we were taken good care off!

       

      6087.attend_5F00_again.png

       

      For me this was the most important statistic. A majority of the participants would do this again without question. Some people are unsure. I think the latter group can be convinced to join again by providing an improved infrastructure, and setting expectations better. I'm very glad people really liked it, and would join us again for a vCampus hackathon!

       

      We also got some additional (free form) feedback. Most were suggestions on how to improve the infrastructure, which we will definitely be taking into account. Bigger tables and whiteboards were a very good suggestion. One attendee suggested having the bar open later , which is of course a very valuable observation.

       

      Overall, I think it was a very good experience. The action items we are taking with us for a next version would be:

      • Improved infrastructure
      • Bigger tables and whiteboards
      Again, I would like to invite everyone to provide feedback on how they experienced the event. I'm also very curious to know of the reasons people had to drop off early. How can we set expectations better?
      A big thanks to everyone who attended!

       

       

       

        • Re: vCampus Live! 2012 programming hackathon -- recap & evaluation
          Bannikov

          Michael @ OSIsoft

          Is there a particular reason why people dropped off? Is there something we could have done to set expectations better?

           

          I can share my own opinion about hackathon as is. I suggest that it should be more personal contest, not team contest. Many people doesn't know each other, and the few who does (more active vCampus member, of course) receive some advantage. So next time I prefer that it would be singe person contest. Otherwise team building is critical for success of hakathon. For example, participants can fill some survey which should contain their programming experience (.NET, PHP, ASP, web services, MS SQL, MySQL...), and team can be formed based on this information - team should contain people that can use the same stuff to create application. It's not useful when one team member prefer PHP, the other - C#, and third can only code Visual Basic version 6. Teams can be formed preliminary -- in this case it would be more efficiently.

           

           

            • Re: vCampus Live! 2012 programming hackathon -- recap & evaluation
              Gael

              A community is made to share expertise and experience. To have a single person contest kill the concept of community (this is just my point of view). People who have no programming skills provide a lot to a team regarding features, architecture, application needs and many more. Everybody has his place in the hackathon: project manager, developers, technicians, end users...

               

              A team is really a team when all members of the team can share each other. Sometimes you agree with the decisions made by the team and sometimes not: this is normal life.

               

              It is not because everybody is using the same progamming technology that the team will work fine. There are always compromises to do a specific task.

               

              Furthermore, we must not focus on programming technology and languages because some applications developed during the hackaton were not necessary using programming but there were using the power of actual PI tools (PI-PB, Coresight...). These applications were amazing.

               

              vCampus is open to everybody. Making selection regarding the skills of people is definitely a bad idea.

               

              In my team, we didn't know each other and we made a good job, taking the best of individual skills and join them. This alchemy works sometimes and sometimes not...

               

              This is just a game.

               

              Gael

                • Re: vCampus Live! 2012 programming hackathon -- recap & evaluation

                  I tend to agree with Gaël on this one... One way we could try and alleviate that in the coming years, would be to try and stimulate the creation of teams ahead of time, here on the vCampus forums. We could announce some high-level details about the hackathon (just enough so that people don't start working on it in advance...) and then discussion would begin, people would share their skills and expertise, and form teams that come to the event ready to rock. Just my 2 cents.

                    • Re: vCampus Live! 2012 programming hackathon -- recap & evaluation
                      Ahmad Fattahi

                      Great discussion. IN real projects things are done in teams; so "being able to work in a team environment" is a critical factor anyway. Using the social features of vCampus can help forming teams in advance - great suggestion by Steve.

                      • Re: vCampus Live! 2012 programming hackathon -- recap & evaluation

                        My executive summary of the hackathon would have to be: Awesome.  

                         

                        I agree with Gael above, not only because he was on our team (Team '01')  but because the team aspect was what worked wonderfully for us.  As an individual I would have struggled given the time constraint to build what we did eventually build; in parallel we had team members analysing Event Frame creation, others writing map control code, and others propping up the bar.   I think we had a Welshman, an Englishman, a Danish guy, an American and a French guy (but don't mistake his accent for Canadian, sorry Gael ).

                         

                        Being 6'4 and a messy worker I did find the tables a little small, but that wasn't an important issue.  I really did want to work my Kinect magic during the Hackathon but just ran out of time...next time...

                         

                        All in all great job OSIsoft.  Looking forward to the next one, going to bring my sleeping bag and a onsie!