Language Selection

English French German Italian Portuguese Spanish

Programming Leftovers

Filed under
Development

  • A Major Contribution to Learning R

    Prominent statistician Frank Harrell has come out with a radically new R tutorial, rflow. The name is short for “R workflow,” but I call it “R in a box” –everything one needs for beginning serious usage of R, starting from little or no background.

    By serious usage I mean real applications in which the user has a substantial computational need. This could be a grad student researcher, a person who needs to write data reports for her job, or simply a person who is doing personal analysis such as stock picking.

    Like other tutorials/books, rflow covers data manipulation, generation of tables and graphics, etc. But UNLIKE many others, rflow empowers the user to handle general issues as they inevitably pop up, as opposed to just teaching a few basic, largely ungeneralizable operations. I’ve criticized the tidyverse in particular for that latter problem, but really no tutorial, including my own, has this key “R in a box” quality.

  • The deep learning obesity crisis - Vincent Lequertier's blog

    Deep learning have made dramatic improvements over the last decades. Part of this is attributed to improved methods that allowed training wider and deeper neural networks. This can also be attributed to better hardware, as well as the development of techniques to use this hardware efficiently. All of this leads to neural networks that grow exponentially in size. But is continuing down this path the best avenue for success?

    Deep learning models have gotten bigger and bigger. The figure below shows the accuracy of convolutional neural networks (left) and the size and number of parameters used for the Imagenet competition (right). While the accuracy is increasing and reaching impressive levels, the models get both bigger and use more and more resources. In Schwartz et al., 2020, as a result of rewarding more accuracy than efficiency, it is stated that the amount of compute have increased 300k-fold in 6 years which implies environmental costs as well as increasing the barrier to entry in the field.

  • Mint: A New Programming Language for Building Single Page Apps (SPAs)

    Mint is a refreshing programming language for the front-end web development. It is developed and maintained by a large community of experienced developers.

  • Understanding Have I Been Pwned's Use of SHA-1 and k-Anonymity

    Four and a half years ago now, I rolled out version 2 of HIBP's Pwned Passwords that implemented a really cool k-anonymity model courtesy of the brains at Cloudflare. Later in 2018, I did the same thing with the email address search feature used by Mozilla, 1Password and a handful of other paying subscribers. It works beautifully; it's ridiculously fast, efficient and above all, anonymous.

More in Tux Machines

Security Leftovers

How to Apply Accent Colour in Ubuntu Desktop

A step-by-step tutorial on how to apply accent colour in Ubuntu desktop (GNOME) with tips for Kubuntu and others. Read more

The Wine development release 7.15 is now available.

The Wine development release 7.15 is now available.

What's new in this release:
  - Command lists in Direct2D.
  - RSA encryption.
  - Initial Wow64 thunking in WIN32U.
  - Optional support for colors in test output.
  - Various bug fixes.

The source is available at:

  https://dl.winehq.org/wine/source/7.x/wine-7.15.tar.xz

Binary packages for various distributions will be available from:

  https://www.winehq.org/download

You will find documentation on https://www.winehq.org/documentation

You can also get the current source directly from the git
repository. Check https://www.winehq.org/git for details.

Wine is available thanks to the work of many people. See the file
AUTHORS in the distribution for the complete list.
Read on

What Is Virtual Memory on Linux? How to Manage It

Virtual memory is a way of representing your memory that's abstracted from the physical memory on your machine. It makes use of both your RAM and your storage space, whether that's on a traditional hard drive or an SSD. In Linux, this is done at the kernel and hardware levels. The CPU has a piece of hardware called a Memory Management Unit (MMU) that translates physical memory addresses into virtual ones. These addresses are independent of where they physically reside on the machine. These address spaces are known as "pages" and they could be in RAM or on your hard drive or SSD. The OS sees these addresses as one big pool of memory, known as an "address space." Virtual memory takes advantage of the fact that not all of the memory that's being used in theory is being used all of the time. Programs in memory are broken down into pages and the parts that the kernel deems as unnecessary are "swapped out," or moved to the hard drive. When they're needed, they can be "swapped in," or brought back into RAM. The space used for virtual memory on a drive is known as "backing store," or "swap space." In the Windows world, it's usually implemented as a file, known as a "swap file." It's also possible to do this in Linux, but it's much more common to use a dedicated disk partition. Read on Also: nbdkit for macOS | Richard WM Jones