Adam’s March 2018 Meeting Summary

Peter started with an update about Windows Spring Creators Update, specifically its new “Timeline” feature, which seems to be a task-view-with-built-in-file-history GUI element.

Pete then gave a “Basics of JSON” talk, detailing {“How”: “JSON can hold”, “Data”: [“in”, “many”, “self-describing”, “forms]} I think most group members are extremely familiar with JSON as a universal string-based mechanism for sharing data, and with many different systems for serializing and deserializing data into / out from JSON. Pete pointed members to the C# JSON library: NewtonSoft.JSON

David Capps brought in Google Protocol Buffers, which perform a similar function to JSON, but encrypts and compresses the data. These seem like a really interesting faster alternative to JSON.

After lunch Dave spoke about drawing 3D images using GLScene and Firemonkey A detailed, practical case-study about converting 3D microscopy images into 3D images on screen. The mind-boggling complexity of working first in 3D, but then also having to add light-sources and cameras made for a really interesting talk, with lots of technical complexity. Both GLScene (open-source library created by Mike Lishkey of VirtualTreeView) and the Firemonkey components seemed capable components, solving many of the same problems. Firemonkey seemed more capable of managing fonts/text.

Then it was back to Pete for a session on ElasticSearch which is an open-source solution for storing large quantities of JSON data, and then querying / searching this data EXTREMELY fast. ElasticSearch stores data, and makes it accessible via HTTP requests, using a relatively simple indexing syntax: //<my-data>/<index>/<type>/<document>coupled with a much more complex JSON-based querying syntax.

Pete uses PostMan as an IDE / environment for querying ElasticSearch. PostMan is a tool for developing APIs that send and receive HTTP data, including JSON. The main takeaways were that ElasticSearch is very fast, has a somewhat verbose JSON-based query language, and is a very useful way to build fast mechanisms for structured queries of diverse data.