Operationalizing machine learning models is an increasingly important challenge. How do we run these models on heterogeneous hardware while maintaining the ability to debug these models when they fail. @bsletten at #OReillySACon
Dimensions of service granularity: functionality, db transactions, data dependencies (schema ownership), workflow/choreography (inter-service communication) - except functionality, all dimensions prefer service consolidation.@markrichardssa at #OReillySACon
shared lib, shared service, code replication, service consolidation - 4 techniques for dealing with shared code in microservices @markrichardssa at #OReillySACon
@seldo@briankardell@Amorelandra I so agree with this: The representation of your object in memory depends what you intend to do with it, and context-sensitive representation is not a feature of OO design.
When doing #DDD, always invest up front in establishing a clear ubiquitous language. Not doing this will result in confusion and wasted effort down the line.
@vladikk#OReillySACon
@vladikk@Heimeshoff Do you give the UL a physical form (in documentation) or is the refinement just making sure you have "good quality" talks with domain experts?
@SzymonPobiega@stilkov Haha, thanks, good point! I wasn't entirely sure, but I thought I guess sometimes you might want to advise in favour of a particular vendor.
1+ million lines of code in node_modules for a hello world app with eslint, tap and create react app. No worries! NPM has solved the software reuse problem. @matteocollina#OReillySACon
By performing stream processing in @apachekafka you can do projections/aggregations/joins that you would otherwise do when querying the database. In this way different consumers can store data exactly as they need it.
@benstopford#OReillySACon