i realized why i don't like the "objects with methods" approach of oop or anything where objects "do" something: because i have a hardware view of the computer i am more comfortable with "the cpu is doing something" (following machine instructions). objects are just memory regions and don't do anything. rather, the cpu (and dma, gpu etc) manipulates the memory. i like the functions and structs approach because it doesn't obscure the machine.
it's me, i'm PCBs image
state of the mnt desktop reform prototype. it's a box with some electronics integrated into the side panels that turn the laptop motherboard into a standalone computer with upgradeable cpu/ram modules. need to order all the components to have blinkenlights and button etc now, and decide on final material options.
shipping of MNT Pocket Reforms to [@crowdsupply]( ) / Mouser continues! (photos by @npub12gl2...frt9)
blender 4.5 (git) somewhat working on panvk (mesa git) on mnt pocket reform with rk3588 with some light patching
new custom trackpad based on iqs9150 works in the MNT Reform Next! trackpad electronics by @npub1829n...s2fp, and i wrote the driver for it today (which runs in the keyboard). the tracking is very precise. gestures&unintentional multi finger rejection need some tuning but we have many options to work with.
ok so @npub1829n...s2fp visited and soldered his custom multitouch trackpad design for the MNT Reform Next. and it works! i just had to bring my windows 10 VM up to date to run azoteq's config/debug GUI which you can see here, and figure out the correct rx/tx col/row mappings. next step will be integrating the new chip in our open source firmware.
i'm a bit confused now image
is the DARE (eu processor) project also somewhere on fediverse? ironically the page lists only X (really?), bluesky etc
i wrote a longer post in MNT Community about the roadmap for integrating the GNOME desktop again for the Pocket/Reform/Next devices in a way that's gonna be a pleasant experience