The timeline

Using Python I was able to construct a simple Monte Carlos analysis tool for use in Project Management scheduling. It takes in the PERT outputs (a weighted average estimate and standard deviation) and runs simulations across a list of tasks, providing outcomes in percentage format (i.e. the percentage of times each estimate occurs relative to the number of iterations). As it just now exists, it will be an important task to assess the accuracy of the output over time. However, it was a useful exercise and cost-saving for the company if it turns out to be an accurate tool. Project Management software capable of running this type of analysis tend to be expensive, which is fair, but with my newly minted Python skills it was a chance to combine career and new skill to generate business value with a much lower total cost. Again, another item I intend to revisit on the blog as results come in, but also an example of how I sometimes stop sniffing the literal flowers in my planters and tr...

Tech aside, I had my first experience of pests the other day. The tomato plants were looking a little yellow on the lower leaves. Flipping the leaves over resulted in an army of little white aphids being exposed. A quick spray bottle of dish soap (washing up liquid, folks in the UK) and a rinse a short time later seems to have slowed the invasion. Another dose in a day or so will hopefully keep the blighters at bay. Positive news, bean pods are appearing, so at least one of my crops will produce a yield. Carrots continue to grow, but they're a little rangy, probably from the limited sun that my functional-but-ugly setup can receive in its out of the way spot. Not to say there isn't direct sunlight, but between the beans and the air conditioner condenser, there's a windows of about 1.5 to 2 hours of direct sunlight on the carrots. Still, for an experiment I'm relatively happy so far. Nothing died, but the time that the auto-watering has been able to run has been li...

Nothing is as easy as it seems. A new attempt at making the watering system operate via a web interface hit a snag. I purchased, and soldered, a new NodeMCU unit. This unit has only 16 pins, as well as the expected ESP8266 chip, but unlike my other two units it also has a 5v pin. I was excited, hoping that, like an Arduino, the 5v would be available through the signal pins. No such luck. The 5v pin is the only pin with 5v logic; all other pins operate at 3.3v, which is insufficient for the N-channel MOSFET used to control the water pump. Anyway, a new attempt: using the new NodeMCU unit with a switch relay. The other NodeMCU units couldn't do this as they were entirely 3.3v; the relay requiring 5v to flip the switch, but only needing 3.3v for the signal telling it to do so. Whether my relay works with 12v logic is another matter - I bought it for mains electric control, so it's only ever worked with AC. Wish me luck.