Plugging Things Together With Reactive Programming

An airhacks.fm conversation with Gordon Hutchison (@hutchig) about:

Playing chess with zx81, huge computer scene in Glasgow, BBC micro then saving for Acron Electron -- the cheaper BBC Micro, programming text adventure games, Forth on RML 380 Z, Sun's OpenBoot was written in Forth, Dragon 32, controlling the computer world with 13, programming colourful fractals, "do whatever you have permission to", then accessing the printer queue, transactions research and Java, IBM develops Java Transaction Service (JTS), travelling to Javasoft in Silicon Valley to transfer the JTS knowledge, moving from JTS to JVM implementation group at JDK 1.2 timeframe, having fun with IBM Java classloader, heap corruption, "lighter" experience with Eclipse RCP, Java Transaction API, Java Transaction Service and CORBA's Object Transaction Service, tranactions are a gift, just learn databases, "we don't need your transactions" in 2006, reused blog post from 15 years ago will be a big hit, IT became fashion -- everything is just reframed, implementing RAID algorithms, enjoying Java EE experience with OpenLiberty, deploying 50 times a conference session with wad.sh, having more coffee with classic WebSphere, OpenLiberty loose applications, OpenLiberty guide to loose applications, starting TX at facade level, JPA and transactions, getting two copies of the same object in the same request, every request is a transaction, loosing up the thread context, project Loom, transactions are making the developer's live simple, the pre-prepare phase, errors on CICS vs. MTS, solving the transaction diamond problem, reactive programming and backpressure, application servers and backpressure, you are not Google, reactive platform at Uber, too much sophistication, too complex to debug, and the human problem, functional reactive programming, plugging things together in reactive programming is appealing, the simple interface between publisher and subscriber, reactive programming as integration hub, learn Java streams first and reactive concepts will come easily, HTTP request / response model does not fit well with reactive programming, backpressure and kafka, Kafka's configuration, reactive streams operators as enabling layer, microprofile reactive messaging is similar to Message Driven Beans, Event Sourcing with debezium.io and Apache Kafka, event sourcing with GRPC, Apache Pulsar the "Kafka.next", SmallRye, CloudEvents and MicroProfile, SOAP envelope

Gordon Hutchison on twitter: @gordhut, on GitHub: https://github.com/hutchig