Our experiences on materials started by watching, smelling, touching, tasting and hearing materials by using our own sensors, which have been developed into a set of advanced characterization tools such as OM, EM, Pyrometer, SQUID, AFM, TES and consequently integrated as IoT/TSU. It corresponds to the first phase of sciences observation science and experimental science, which has produced tons of data requiring feature extraction, generalization, grouping, abstraction, modeling and digitization of observation and experiment.  Due to not only technical and/or scientific limits of observation and experiment but also economic feasibility, we have not been successful in digitizing captured data perfectly except for some special ideal cases of enough physics behind.  Therefore, as practical solutions for dealing with complex engineering materials data of diversities and uncertainties we have developed exchange formats, standards, metadata and ontologies, expecting hand-curated, very high quality reference data are prepared by someone and/or collective knowledge like LOD/RDA.

During compilation works on materials data with perseverance-pleasant but sometimes very boring data curations, useful digital management tools have been developed to digitize, visualize, share, filter, analyze, assimilate, mine, cluster and categorize.  Thanks to ICT drastic revolutions in this century, an ecology of materials data infrastructures is expected to emerge for “Cambrian Explosion” of materials.  So as to make it happen we need to investigate and develop key tools for digital communications among materials(nature), human experts on materials, measuring instruments and materials users/machines by sharing common media.  

Theoretical science and computational science may play important roles to form the common media.  Strategic integrations of first principle calculations and machine learning/AI methods have become very hot topics in this decade, but it is just in the beginning of population genetics.  There are many issues if we trace an analogy between bio-genetics and material-genetics. The central dogma from gene to health in bioscience looks like a fiction as was the case of classic multiscale modeling in nuclear materials simulations since 1950s.  Too many options in the linkage, too much path dependencies and too many missing links still exist before us, which require multi-principles strategic approaches from geometry to reality-existentia based on intelligence as common sense.  It requires a collection of materials landscapes drawn by taking advantage of above-mentioned preparations. 

Data science and design science are expected to deal with connotation and denotation of materials, which implies the essence of engineering.  It is the final goal of MGE, namely, chaoscapes of many additional lines, surfaces and voxels on the original landscapes of materials.  As for such additional information, first principle calculations may play a role of producing lines and surfaces from a constellation of high quality data points. So as get surfaces and voxels, we need thermodynamic frameworks endorsed by a collection of high quality single phase data.  There we need to overcome boring repetitive and routine works to deal with data properly, so that we need to apply intelligent and useful methods as roughly called AI. Bridging needs and seeds by modeling, optimizing, deduction, induction and abduction requires our collective creativity as humans.  Here we need a new thought to go beyond the current data infrastructure supported mainly by original scientific papers.