85th Monthly Technical Session
Every month, HENNGE hosts a session for sharing technical knowledge among its members known as Monthly Technical Session (MTS for short). The 85th MTS was held on August 27th. Continuing the trend of staying safe and working from home as much as possible, this month’s MTS was held online on Zoom.
Erasing Data
Fukutomi gave the first talk at 85th MTS about erasing data. Specifically, once a contract for using our product is concluded, we need to make sure that the customer’s data is erased completely. When we delete a file from our system, normally that operation will only overwrite the index of that file. To make sure the data is destroyed completely, we need to overwrite the entire data completely with new random data. He continued to explain that in the past with low-density HDD, three times is the minimum amount of overwriting necessary to ensure the data cannot be recovered. However with recent high-density HDD, once is enough to cover the previous data, thus the standard is updated accordingly.
NFT in 10 Minutes
With NFT gaining popularity in recent times, Akiko decided to give a talk explaining what NFT is all about. Even though crypto art’s content is something that we deem common, they are sold for an exorbitantly high price. This raises the question of why, with digital content being easily duplicated, would someone spend that amount of money on it? The answer, it turns out, stems from the ideology that “Everybody can see it, but only one owns it”. Ownership is an emotion, and NFT ensures that the ownership can be transferred or sold in the future if necessary. Akiko continued the talk by explaining how it works internally.
Data Fetching is Awesome
As a frontend engineer, Ray shared his experience solving a challenge he found in a project he’s recently working on. Similar to how CSS is “awesome”, he found that data fetching is just as “awesome” as CSS, meaning that there are a lot of problems that might come up when working on it. Among them, there are waterfall problems, loading indicator management, code organization with data dependencies, early data access, etc. Ray continued by explaining the three approaches that we can use, which are: fetch on render, fetch then render, and render as you fetch. He explained how each one works, how the new approach render as you fetch allows us to have the best of both worlds by utilizing a try-catch-like code to throw promises and error objects to be caught by the ancestor components.
Running Parallel Programs using TSUBAME3.0 Supercomputer
Jasmine started her talk by explaining about a supercomputer, which is an extremely powerful computer used for simulations, big data analysis, or machine learning projects. Using FLOPS (Floating Point Operations per Second) as a measure of computing power, the world’s fastest supercomputer is able to achieve more than two million times the computing power of a Macbook Pro that we usually use for work. As a student at her university, she is allowed to borrow a set amount of resources for her projects. She then demonstrated to us how she utilized the TSUBAME3.0 supercomputer to calculate pi with parallelization techniques.
Gavin gave a talk about no-code application builders, which is a software development platform that allows people to build apps without the need to write a single line of code. This is most commonly done through natural language input, so no technical background is required. There are many such builders in the wild, and Gavin introduced one platform in particular which is OpenAI. The way it works is quite similar to how we can train computers to translate human languages for us, in this case, the platform would translate our human language to a programming language. He then gave a demonstration on how to build a simple game on OpenAI, and produce Javascript code as a result.
We ended MTS with a beer bash, also held online on Zoom 🥳
