A daily blog by Steve Vale on the progress of the sprint - 20 February to 2 March 2012
Sunday 19 February
Arrived in Ljubljana. Air France tried to send my luggage to N'djamena, but fortunately I noticed in time, and do have clean clothes to wear tomorrow. Pre-sprint briefing soon with the Australians and the facilitator, Fiona.
Monday 20 February
Day 1 of the sprint. The cold and snowy weather provided a contrast to the warm welcome from our hosts, the Statistical Office of the Republic of Slovenia. Irena Križman, the Director General, joined us for the opening session and helped to stimulate our first discussions. The participants are starting to share their ideas and to debate some of the fundamental issues. I think it is fair to say that whilst we all agree on the need and the drivers for the GSIM, we have different ideas about what the GSIM should be. Spending the evening together trying Slovenian food, beer and wine has helped to further develop the team spirit, so hopefully we can start to reach some sort of consensus tomorrow.
Tuesday 21 February
Lunch time: We are starting to get to the real issues of the scope and purpose of GSIM. Regarding scope, it was easier to identify what is definitely out (technology, methodology etc.). Harmonising information objects, facilitating communication and enabling easier re-use of information are currently seen as the main purposes, but that could change during the afternoon as we look more at user needs.
Exercise on the purpose of the GSIM (click to enlarge)
18.30: Finished the formal work for the day, though no doubt the discussions will carry on over dinner this evening. We ended the afternoon with an exercise in small groups to come up with a rough draft GSIM targeted at a specific user group. This helped to focus minds on a number of issues that will definitely need more discussion (e.g. is "data" part of the GSIM? Is GSIM about the "blood" or the "arteries" of statistical production?). All participants agreed we have made good progress today.
Wednesday 22 February
A busy and fruitful day today. We spent the morning reviewing related initiatives by the Statistical Network, the CORE project and others, identifying useful inputs and lessons learned. Thanks to Adam Brown (New Zealand) and Carlo Vaccari (Italy) who joined us by computer link to share their experiences. This afternoon we defined success criteria for the GSIM: It should be communicable, stable, applicable, complete, inspirational, fit for purpose and sustainable. After this we defined the work programme for the rest of the sprint. Starting tomorrow, two groups will start to model the GSIM, one taking a top down approach, the other starting from bottom up identifying the objects to be included in GSIM. By the end of the day there was a strong feeling that we had made progress, and an impatience to get started on the modelling work.
Scenes from the Sprint Room (click to enlarge):
Thursday 23 February
Lunch time: The groups have been busy. The bottom-up group are reviewing the relevant literature from previous projects to identify and eventually classify the necessary objects for the GSIM. Their table is rapidly disappearing under post-it notes. The top down group split into two, working on a one sentence definition of the GSIM, and a high-level diagram respectively. They have just presented their first ideas to the whole Sprint team, and have gone back to their room to incorporate feedback and develop their outputs further.
Out of chaos comes order: 
Developing the model: 
"Standing on the Shoulders of Giants": 
Evening: The groups have presented their initial conclusions, and it seems that it will be possible for their views to converge. That is the challenge for tomorrow. We had a useful conference call with experts outside the sprint group this evening, where the feedback was that we are going in the right direction, but we must ensure that the scope of our work remains realistic.
Discussions continued over dinner, and quite late into the evening. There is an emerging issue that there may be something missing in the overall schema. We have the diagram from the HLG-BAS vision, with the 4 corners, GSBPM, GSIM, Methods and Technology, but perhaps we need something above this, to bring together these four components to actually produce statistics. Perhaps some sort of generic management of flows and processes, to ensure they work in the most efficient way. We will reflect more on this in the coming days.
Friday 24 February
We have reached the end of the first week, and the mid-point of the sprint. A lot has been achieved this week, and many of the team were looking rather tired this afternoon. However, by the end of the day we reached a convergence of views on the high-level information objects, and the programme of work for the remainder of the sprint. Whilst we have agreed to focus on the GSIM itself, there is a growing recognition that more work will be needed to specify a model tha links GSIM with the GSBPM, methods and technology. This is a topic to be raised with the High-Level Group.
Saturday 25 February / Sunday 26 February
A couple of days off, so some of the group decided to explore Slovenia. We enjoyed visits to Bohinj, Bled, Piran and the Postojna caves. However, even there we could not entirely stop thinking about GSIM. We are refreshed and revitalised though, ready for the challenges of the week ahead.
Top down meets bottom up, though I think we may be a little quicker: 
Piran: 
Bohinj: 
Monday 27 February
A long day today. The main tasks were ensuring convergence between the top down and bottom up approaches to defining information objects. This resulted in agreement on the top two layers, though the extent to which "external" objects at either end of the statistical production process (such as information requests and statistical products) should be included remains an open question. Work also started on the business cases for the GSIM and the "grand unification" approach, as well as testing the emerging GSIM against real life examples of statistical production.
This evening we had a conference call with the "practitioner group", Adam (New Zealand), Tim, Sheri and Alice (Canada), Dan (US) and Klas (Sweden), to get feedback on the work so far. This was very helpful, particularly in determining which diagrams worked best in terms of communicating our results. The subsequent discussions continued over dinner and late into the evening, resulting in more diagrams and ideas, all of which seem to be taking us closer to our goal.
A top layer view? 
Tuesday 28 February
5pm and we are going strong for a few more hours at least. The highlights so far have been a general agreement on the top level objects (blue circles in the diagram above), and those at the next layer down, as well as some provisional definitions of the terms used. In parallel, the process of testing the model has continued, with a second scenario (a register-based census). The queries from the tests are currently being resolved. At the same time, other members of the sprint team are working on communication aspects, including diagrams and a future view of the environment in which the GSIM will operate.
We have agreed to continue this evening until we are satisfied that our GSIM draft is good enough (or until the security people throw us out of the office).
8pm - arrived back at the hotel. The queries from the test scenarios have largely been resolved, though exactly where to put "quality" within the model is still open for discussion. The model is therefore considered more or less complete for the purposes of Sprint 1, and the remaining time will be used for documentation, refining the business case, and developing communication tools.
Wednesday 29 February
Slightly earlier finish this evening, as we had a group dinner. The sprint team are starting to look a little more relaxed, but also rather tired. I think we also had cause to celebrate though, as we had made good progress on several fronts, particularly in the first part of the day. The documentation is coming together well, and a third test scenario has helped to remove any lingering doubts about the applicability of our version of GSIM. We have been refining our communication material, including a presentation, documentation for the HLG-BAS and proposals for future work. We have a conference call with the HLG-BAS Secretariat Team and our practitioner group tomorrow, which will be the real test of whether the outside world can understand and appreciate what we have done.
Thursday 1 March
Busy morning preparing our presentation and text for the final report, and various other pieces of documentation. This afternoon we had the opportunity to share our ideas with the HLG-BAS secretariat members and the practitioner group. As expected there were some questions and comments on the detail, but we were very pleased (and relieved) to receive a lot of positive feedback about the overall direction of the work on the GSIM, and our ideas on "Grand Unification" (see blog for Thursday 23 February).
Friday 2 March
The sprint is over, and the participants have started heading for home. We spent today tidying up the last loose ends, and preparing and delivering a presentation for senior staff from the Slovenian statistical office. Their feedback was positive, and it was a useful practice for those participants who will need to deliver similar presentations on their return to their own organisations. We said our goodbyes and reflected on what we considered to be a very hectic, sometimes frustrating, always stimulating, and ultimately very rewarding two weeks.
