Search is a difficult problem. Especially when it comes to enterprise and organization. According to researches, more than 2/3 of people searches are not returning satisfactory results. Enterprise is messy, complicated and contains lot of unstructured data these days. CAD and other 3D files are part of this messiness. For many years, we generally thoughts of web and enterprise search as a place where text begot text – you input some text, press search, and get a bunch of relevant results, also in text form. However, in many places, and 3D and CAD is one of them, search by keywords is not very efficient.
Company have been trying to innovate in 3D or Shape search for the last decade or so. Read my old blog – 3D Shape Search in CAD and PLM. Despite many available solutions in 3D search, I can see a low adoption rate. In connection to that, I saw the opportunity to rethink 3D search.
The following article caught my attention earlier this week – In Challenge To Google, Pinterest Launches Guided Search. Pinterest is a visual discovery tool that people use to collect ideas for their different projects and interests. People create and share collections (called “boards”) of visual bookmarks (called “Pins”) that they use to do things like plan trips and projects, organize events or save articles and recipes. According to the article
At an event at Pinterest headquarters Thursday evening, CEO Ben Silbermann announced Guided Search, a new visual way to explore Pinterest’s more than 30 billion pins—links or images chosen by users and assigned by them to topical collections. Unlike most search engines, where you must choose a precisely constructed string of keywords for what you want to find, Guided Search offers suggestions as you go, based on the associations Pinterest has learned to make between all the objects in its database.
You can learn more about how guided search works here and on the following video:
The idea of "discovery" is very compelling in the engineering discipline. In many situations it is very hard to formulate a specific keyword based query to find a result. Therefore the ability to classify, categorize, slice and dice data can be very powerful to search and navigate 3D data.
What is my conclusion? Guided search is not a completely new idea. You can see that in some old enterprise search systems. However, to combine together with visual data corpus provides some fresh experience in results. The combination of technology and experience is important. Web is slowly becoming a future platform for engineers. We are spending more time online and getting additional web experience. Some web tools are providing ideas, technologies and solutions that can be re-used in engineering and product development. I think Pinterest guided search is one of examples. Just my thoughts…