Last week (July 1st) JTeam hosted its first Search Symposium at our office at Frederiksplein 1, Amsterdam. The idea of the symposium was to get people together with different perspectives on search and discuss their experiences. The meeting was hosted by Adriaan Bloem (Analyst for CMS Watch) and Edwin Adriaansen (CTO from iLocal). They both presented their take on search, where Adriaan focused more on the search market, different vendors and the open source alternatives and Edwin provided insight into the rationale and findings of doing a large-scale search implementation.
As said, Adriaan focused more on a high level overview of what components make up a full search solution and provided some background on the current market. He spent a fair amount of time on open source alternatives, basically Apache Lucene and all the projects surrounding it (e.g. Apache Solr and Apache Nutch). What I found most interesting about his presentation was that he highlighted the importance of organizations doing a search implementation to actually think about the requirements upfront and choose the right solution to meet those requirements. In his opinion, still too many people pick a solution before even knowing exactly what they want to get out of their search solution.
Edwin’s presentation was the business part of a case study JTeam and iLocal did together at the last ApacheCon conference in Amsterdam. It introduced the implementation project they did at iLocal (JTeam helped iLocal by empowering the in-house development organization) using Lucene and Solr. It showed the rationale for the iLocal management trading in their existing commercial search solution and replacing it with an in-house developed solution (on top of Solr). Their main reason was that this provided them with the flexibility and time to market that a company like iLocal needs.
The findings included the fact that the project was very successful, a huge increase in performance they gained and an overall vendor independence. One interesting thing that came up was that Adriaan at some point commented that iLocal is a good example of what he highlighted in his presentation: that you need to know your requirements in detail. At iLocal they knew exactly what they needed and therefore was able to make the right decision. But iLocal is a project where the problems with search implementations are sort of put under a magnifying glass. All the challenges that you are likely to encounter in a search project became very obvious in the case of iLocal. In most search projects they are not that obvious and therefore get easily overlooked until they become too big and are harder to fix. So the need for knowing your challenges and needs upfront was illustrated once more.
After a short break, the meeting continued with an open discussion between all attendees. With some guidance in the form of statements regarding business cases, open source vs commercial, and etcetera, the discussion went on for about one and a half hour. It was a good and vivid discussion, which was also very broad, a little bit too broad in my opinion, from the value of open source in general to large-scale government projects. However, a lot of food for thought 😉
The one thing that I got from the whole symposium was that there are many different perspectives on search and most of them come with their own set of challenges. So again, think well about what you want to achieve with your search solution upfront, try to match the best solution to that and then make it happen!
Due to the success and the enthusiasm of the participants, we will definitely follow up with more meetings related to search. We got some valuable feedback from the evaluations we sent out, which we will incorporate into the next one to make the next one even better. Looking forward to seeing you all there!