2021-06-23 | Meta | June 2021 Status
My demos from April and May were not satisfying for me. I don't think they effectively demonstrated the main points. The feedback I got was meh. I've also refactored my direction a bit.
Philosophically, I have changed my understanding of what my efforts mean. This change is motivated by Laura Jackson's book The Telling. I am sure I have corrupted her book in understanding. I am a newbie and she is a monster in her area. I am re-reading her book. I did like her approach of inclusion. The idea is that her "telling" included everybody. She focused on her common humanity, and on ideas and needs that ran deeper than, say, optimizing links in the supply chain. This seems similar to Kesey's rambling caution in his speech at UC Berkeley that information does not flow from the mic out like an inverse pyramid vs. at the base. We share much as humans, we share more than we have differences. Jackson presented her "telling" in a way that was inclusive, and challenged the reader that if they disagreed to come up with their own telling that was inclusive.
Jackson deserves recognition of another level down in analysis. As an example, the experience of holding new life, a beginning, being pregnant, and understanding your own mortality at the same time, is a profound experience that males do not experience. There is a truth, verity, to this. It is quite a challenge to understand this, and at the same time draw an inclusive circle in a telling.
Finally, Jackson was obsessed with the meaning of words. At the time she wrote, there was no idea of formal ontologies or meaning from an AI/ML perspective, at least, not that she was aware of, I imagine. This matches my ideas as well, in that we can get agreed on meaning collaboratively and inclusively. I'm a bit out of my depth on this, but I even suspect that some of her concerns of inclusion could be addressed by an open world assumption coupled with formal ontologies.
Most current human work is adapting to highly decomposed work streams to match the cadence and needs of the global supply chain. We are bound.
At the same time, though, there is much within humans that is outside of that track of 6,000 years. It isn't like we can decouple in any realistic way; however, being aware that we are human and have relations and needs outside of this, is key to our wellbeing. We need to remember this as we are shuffled into virtual meeting cubes. We have needs outside of the web that delivers an electronic device in a cardboard box in 24 hours or sustains 8 billion people on the planet. We need real human touch to thrive. I need to be aware of this as I develop and share my efforts. While I have been aware of this catch-22 for awhile, that what I am resisting, the insanely complex global supply chain, is required, I have not focused on the inclusive telling from an audience and scope of story perspective.
I was talking to a friend of mine a few days ago face-to-face. She advised that I focus on the fact that my ideas are useful as a form of disaster recovery. Stuff breaks. Stuff needs to be refactored. She said that people don't really know what the word resilience means, and that I should probably avoid using it.
At a high level, this is what I need to show:
I need to add operational streams (log data and related, from ELK stack to Syslog to Splunk to cloud streams) and open the demo straight off with this view. This is what most people in my audience see daily.
After I show how the operational streams can be captured and automatically visualized, I can move into the data flow models. At a most basic level, I can take an alarm for a component from an operational stream and highlight it on the data flow. I did this in the thirty minute demo, but the demo suffered from the introduction which was primarily around capturing knowledge in the form of data flow.
As far as technical progress towards these two goals, I have collectd feeding event streams via AMQP 1.0, and am populating physical domain graphs that will be linked on processes in my data flow models.
After I invert my main presentation and start with operational intelligence vs. ending with it, I have another stretch goal. I realized during the last few months that my models have all of the aspects of CI/CD. My "code" though, is pushed down to data. Data is more important than code in many ways, yet most of our CI/CD pipelines focus on the code. This is the nature of N-Triples in that the knowledge can be streamed sequentially and merged. This also gets at my inclusive stance and ownership in general. Knowledge of a system can be captured, collaboratively improved, visualized, and tested at a data perspective independent of platform. This is the power of a graph approach, an ontological, open-world assumption approach. We don't generally call the creation of this kind of knowledge CI/CD, but it has the features. The models can be continuously integrated and even deployed. This is why I focused on the front-end for the DFD capture first.
Doing all of this while keeping Jackson in mind is kind of weird and challenging, but I think it will make my work more useful.