FOSSPicks
LabPlot 2.8
LabPlot is an interactive data plotter, visualizer, and data analysis tool that's designed to keep you from ever opening a spreadsheet again. It lets you import your own data from any source, such as measurements from an Arduino sensor or statistics from a website, and plot that data into any number of charts, matrices, spreadsheets, and tables. It's also another application that has seen a lot of development since we last looked at it over two years ago, and the 2.8 release alone involved a year of work and a month in beta. But there's one feature in particular that makes this update worthwhile: It solves the biggest problem for those wanting to explore data analysis, which is finding meaningful data to analyze.
This release makes it much easier to explore one or more online datasets by integrating them directly into the import dialog rather than having to source them and download them yourself first. Just select Import from Dataset Collection as a source, and you're presented with the collection view. This has three main panels to hold a category, the dataset names, and an explanation of what the dataset archive includes. Categories include chemistry, medicine, nature, physics, and statistics, and it feels like a superpower being able to access this kind of data from your desktop. To help you explore these new horizons, there are two new worksheets. The first lets you highlight specific lines with a plot, while the second lets you import external images into a worksheet to help make the data more presentable. The spreadsheet view also adds more descriptive statistics, such as quartiles for statistical mode and more normalization methods. The amazing data analysis view can also calculate rough values to smooth a histogram, which makes it even easier to see patterns in your data. It's deep and fascinating and only an install away.
Project Website
https://labplot.kde.org/2020/09/16/labplot-2-8-released/
SSH tar pit
Endlessh
The perennial secure shell, SSH, is cryptographically secure, as attested to by the millions of servers, devices, and humble Raspberry Pis that rely on it for remote access. Even when a device is visible from the Internet and you don't enable a passwordless login, it's still secure as long as your password is complex enough. But this doesn't stop people, or bots, from endlessly trying to guess them. If you've ever exposed an SSH server to the Internet, you'll know that your authentication logs are soon swamped with external IP addresses trying to guess username and password combinations on your SSH port. Disabling password logins helps, as does changing the default port from 22, but we'd also recommend installing either DenyHosts or Fail2ban. These are brilliant Python scripts that watch for login attempts and automatically block attempts from specific IP addresses when they fail to log in, either permanently or for a set period of time.
But Endlessh is another option that can also help mitigate logs full of failed SSH connections. It belongs to the "security through obscurity" category of security utilities, which means it doesn't really make anything more secure, only less attractive in the time wasted/CPU/bandwidth equation likely used by botnets. But it may also make you smile. That's because its main function is to send the SSH announcement you see before you login at a very … very … very … slow … speed. Any ne'er-do-well trying to access your machines needs to wait patiently for this to load before they get the login prompt, and you can configure that to wait for days. It can even trap multiple clients at the same time, making it use very few system resources. But the really clever thing is that it isn't using SSH at all. Even after the wait, there will be no login prompt. Instead, Endlessh is a simple, secure decoy while you secretly run your SSH server on a different port completely, hopefully now free of endless login requests.
Project Website
https://github.com/skeeto/endlessh
Buy this article as PDF
(incl. VAT)
Buy Linux Magazine
US / Canada
UK / Australia