Posts

Posted on

Running WSJT-X on a Raspberry Pi using an SDRPlay RSP2

Using the preconfigured/installed Raspberry Pi .img from SDRPlay (here), it’s pretty easy to get WSJT-X up and running once you work out some stuttering audio issues with GQRX or CubicSDR (both installed on the image). I’m using a Raspberry Pi 3 for this experiment, maybe a new RPi 4 would have better results.

First download and install the .deb for the Raspberry Pi from the WSJT-X page here.

Start up either GQRX or CubicSDR and tuning to the standard FT8 frequency for the band you’re listening to, eg. 14.074MHz for 20m.

Here’s GQRX listening on 14.074MHz:

it was good to see that there must already be preconfigured virtual source/sink configured for alsa/pulse audio, so just selecting the available pulse audio as the input in WSJT starts to get you signals on your waterfall. I was fully expecting to have to configure my own source/sink virtual audio cables to pipe the audio from one app into the other, but this seems to be ready to go straight out of the box with the supplied .img:

At this point though I was stuck for a while trying to work out why I was not getting successful decodes, notice all the signals in the waterfall above, but the decodes window on the left is empty. I checked the audio sample rate in GQRX was 48khz, and I checked ntpd was running so my clock time was in sync.

I went round in circles trying to debug this for a while until I listened to the audio being received and realized the audio from GQRX was stuttering, it was only out putting about 1/2 sec audio in every sec. Decreasing the SDRPlay RSP2 sample rate in CubicSDR all the way down to the minimum 250kHz solved the stuttering (slightly higher sample rates worked too, but I didn’t need more than this for listening on a single frequency, and this same approach worked in GQRX too), and now with constant audio with no dropouts, I was getting good decodes in WSJT-X:

Logging my spots out to pskreporter, 40m at 11pm this evening was not great, but had a few spots on the West coast and also out to Cuba:

This is great that you can get an SDRPlay RSP2 running as a receiver on a Raspberry Pi and also run WSJT at the same time. I have a few ideas for how I can use this as a low power setup running 24×7… more details to come as my project takes shape… 🙂

Posted on

Is the software development industry trapped in a cycle of being ‘perpetually immature’?

A while back an xkcd comic on Voting Software attracted a lot of attention on Twitter for it’s alarmist comparison of the attitudes of the software development industry to other engineering disciplines and in particular, this statement:

“our entire field is bad at what we do, and if you rely on us, everyone will die”

From: https://www.xkcd.com/2030/
Voting Software
From: https://www.xkcd.com/2030/

This was widely shared and quoted on Twitter, and even shared by industry leaders such as Grady Booch:

https://twitter.com/Grady_Booch/status/1027291830101016576

There was one comment in reply to Booch’s tweet that particularly resonated with me:

https://twitter.com/Joerg1968/status/1027417862422573057

“I fear that perpetual adolscence is our curse and our gift – (we’re) always chasing the next shiny thing”

Joerg Muller-Kindt

This thought that as an industry we’re stuck in “perpetual adolescence” caught my attention. As an industry we struggle to advance. What’s unusual when you think back over the past 20 years and even further back looking at the history of our industry, there are developments and ‘movements’ where groups of individuals come together to champion an idea that at first seems to make a positive improvement to our industry as a whole, and then over a few more years the ideas fade away and seem to be forgotten. Some are even occasionally rediscovered again by another group of like-minded individuals years later.

The ‘perpetual adolescence’ of our industry as I see it is not that we’re stuck at a particular point in our industry’s growth and unable to progress, we do progress and advance. This growth however, appears to be in short cycles where we move forward, but then we go back to what we were doing before, and we’re stuck in a cycle of discovery and rediscovery. It’s not as bad as one step forward, two steps back, but it does appear to be something like one step forward, one step back. We’re never continually advancing, occasionally we move forward but we end up going backwards again. The progress of our industry never seems to be continually moving forward. Maybe this is typical of any industry in its infancy.

Examples:

  • The ‘software crisis‘ of the 1970s
  • Structured Analysis and Structured Design
  • Formal methods
  • 4GL development tools focused on developing specific styles of application (e.g. PowerBuilder, and client/server type systems)
  • The hype of Object Oriented languages
  • Fred Brooks seminal article “No Silver Bullet“
  • CASE (Computer Aided Software Engineering) tools and code generation – mid 1990s
  • UML, Rational Unified Process (RUP)
  • The Agile Manifesto
  • The cycle of centralized vs decentralized (distributed) systems
  • The hype of Functional languages
  • Strongly typed vs dynamic languages
  • Platform independence vs platform dependence (of languages and technologies)
  • Software development as an engineering discipline or an art, and everything in between
  • Serverless, as the solution to all problems

There’s plenty of other ideas and concepts that could be part of this list, and I’m not highlighting anything in my list to say these ideas good or bad, but when you look back it’s interesting how ideas come and go over time. People come and go too, some noticeably bringing new ideas and passion with them, sometimes rediscovering ideas again from the past. Sometimes as champions leave the industry their ideas no longer have a voice and their ideas disappear with them too.

If our industry is still in its infancy, is this cyclical, repetitive behavior just the industry trying out new ideas to see what sticks? Or is it that the people actively involved in championing ideas come and go too, bringing ideas in and sometimes taking ideas with them when they move on? Does this mean that as an industry as experienced developers leave the industry and new generations of developers come in, are we doomed to continually repeat our mistakes of the past?

Posted on

Why code is hard to test

Junior Developer: I can’t write a test for this code, it’s too hard to test

Senior Developer: Who wrote the code?

JD: I did

SD: You’re saying the code you wrote is hard to test. If you wrote the code, who made the code hard to test?

JD: I guess I did

SD: Right, so can you restructure the code to make it easier to test?

JD: Yes, looking at it again, I could break it down into smaller methods that would be easier to test individually

SD: Great! In future, how can we write the code so make it easier to test?

JD: Well if I think about how I’m going to test the code as I write it, then I can definitely write it in a way that would make it easier to test

SD: Great! And if you write your tests before you write your code, that would help even further with that approach, right?

JD: Yes, I guess it would.

TDD is not rocket science.

Posted on

Installing Solaris 2.6 under QEMU

I’ve been looking at picking up a used Sun Sparcstation from eBay. It occurred to me that I’ve never installed an early version of Solaris before, so wondered if I could give it a try under QEMU since it’s emulates different hardware, including Sparc.

There’s an awesome step by step guide on Adafruit that takes you precisely each step to get Solairs installed un QEMU. You can follow the steps in their article here, so I won’t repeat all the steps here.

The key steps before you get to the install are creating a disk image:

qemu-img create -f qcow2 sparc.qcow2 9663676416

and then booting with the Solaris iso image as the cdrom and the disk image attached:

qemu-system-sparc -M SS-5 -m 128 -drive file=sparc.qcow2,bus=0,unit=0,media=disk -drive file=solaris_2.6_598_sparc.iso,bus=0,unit=2,media=cdrom,readonly=on

After this point it’s following through the steps in the install.

Here’s qemu booting up for the first time:

Here’s the Solaris installer starting up:

After the install had completed, here’s the rather impressive for it’s time CDE desktop:

Proudly powered by WordPress