Living in the Shell -------------------- I am now using the Devterm R-01 for, what, 11 months or so as my more or less primary computer, which has lead to the "discovery" of some problems: - The screen real estate with its widescreen format and small size isn't ideal for a traditional desktop - It is somewhat slow - The trackball is still a crime against humanity What i also noticed: - The screen real estate is absolutely OK for shell based work - The keyboard, while relatively small. is still rather pleasant to type on - The battery run time is really good, normally i get a full workday with note taking, mails and so on out of it... even the extended nethack round after work doesn't drain it completely. - In the console, running in framebuffer mode, the display is rather pleasant on the eyes So, some time ago i stopped using X and settled for a entirely console based workflow (as any sane person would do in 2024... right?). Little improvements As the R01 is a somewhat experimental system, using a RISC-V chip and was put to market rather hastily by Clockwork Pi it has some... pecularities in its supplied OS image. Thankfully, there is also a quiet active community around it so there are tips on improvements to be found on the forum. In the forum i discovered the work auf the user katmai, who has created a nice toolbox of improvements that makes the life with the R01 way better, he even made his collectian available on github [1]. With this toolset the R01 boots now up in seconds, has a nice console font, checkable batery status from the console and many more niceties. Software that i settled on Email - mutt File manager - nnn Newsreader - slrn RSS - newsboat Calender - calcurse PDF reader - green Image viewer - fim Browser - elinks Media - mpv (works reasonable with low res videos) Gopher - gopher And then there is also the plethora of various UNIX tools like troff, awk etc. that i use for some tasks, some installed games and such things like dictd that are just nice little helpers. Using only the erminal has taught me one thing: To really appreciate the modularity end empowerment of the user a Unix (-ish) OS gives you, you HAVE to dive into the terminal. Having (normally quiet good) manuals for every command right there and the ability to glue programs together to form your own solutions is a really fascinating experience. While viewing images or videos on the Devterm was never a big problem (besides the lack of computational oomph), viewing PDF's required a bit of searching and experimentation. While there are some console based PDF viewers out there, the only one i managed to successfully compile and get to run was the since about 14 years unmaintained green[2] pdf viewer. After getting it to run i faced the little problem that the framebuffer of the R01 seemingly rotates everything 90 degrees anticlockwise. With mpv or fim this is no problem (you simply rotate the image back), but green doesn't have this buildin function. So i put together a little script utilizing pdftk to rotate the pdf before handing it to green: %<------------------------------ #!/bin/sh if [ -e $1 ]; then pdftk $1 cat 1-endeast output /tmp/$1.pdf sudo SDL_NOMOUSE=1 green /tmp/$1.pdf rm /tmp/$1.pdf else echo "$1 existiert nicht" fi %<------------------------------- I often need to scan documents that i receive in paper form to pdf, so, after wrangling around with sane and img2pdf i decided to hack together a little script that streamlines the process to something as comfortable as using a GUI document scanner: %<------------------------------- #!/bin/sh if [ -n $1 ]; then SCANFILE=$1 else SCANFILE=/home/ralf/scan/scan_`date '+%s'`.pdf fi if [ ! -d "/tmp/scan" ]; then mkdir /tmp/scan fi while true; do echo "scanne..." scanimage --format=jpeg --output-file /tmp/scan/`date '+%s'`.jpg --progress echo "Eine weitere Seite scannen (j/n)?" read answer if [ "$answer" = "n" ]; then break fi done img2pdf /tmp/scan/*.jpg -o $SCANFILE.pdf && cowsay "Erledigt! Die PDF Datei liegt unter $SCANFILE" rm /tmp/scan/*.jpg %<------------------------------ For writing documents, reports and so on i settled on groff. Yeah, it has a bit of a leaning curve to it but all in all it statisfies my needs for creating documents. Luckily, today most of the documents i exhange with others are in PDF format which can easily created with the groff toolchain. The only downside i get with this setup is, that the modern web is nearly impossible to use, at least the JS heavy sites are oout of the question. But... is this really a downside? I could now dive into a rant about the modern web, tracking and so on, but i think we all know the drill. To conclude this text let me say that i don't advocate to replicate this setup. It works for me with my particular messy brain and counters very well my tendency to procrastinate away on time sucking websites (now i am procratinating by studying manpages and tinkering with shell scripts...). All in all it is a nice experiment that i am now using for a while and perhaps will use at least for a while in the future. Perhaps in a month, two months or a year i will be back on a more normal system, who knows?