byte's silly little site

running 9front on arm mac in qemu

byte — Fri, 21 Aug 2026 00:00:00 +0000

getting started</h2>
this guide assumes the reader is familiar with what plan9's</a> deal is, basic features of qemu</a>, and has unhealthy attraction to obscure software like the author does. also some arm mac hardware.</p>

x64</h3>
if the latter isn't the case but everything else is still true then good news! 9front can be installed just by following the guide, x86_64 is quite simple and the semi-auto installer takes care of everything properly. i went through it many times both on qemu and real hardware like thickpad T60. here's a video that can guide</a> you through it.</p>

arm64</h3>
now for the rest of the audience:</p>
first go to this peculiar corner of the internet, grab the latest stuff:</p>

https://9front.org/releases/2026/08/02/0/</a>, specifically the arm qemu image</a></li>
and a pdf manual</a> for it</li> </ul>
after it’s all downloaded i highly recommend skimming through the whole manual, it’s quite entertaining and will be enough for a quick basic introduction.</p>
qemu setup</h2>
get brew (if needed), follow prompt to add it into zshprofile and the current console session</p>
`/bin/bash</span> -c "</span>$(</span>curl</span> -fsSL https://raw.githubusercontent.com/Homebrew/install/HEAD/install.sh</span>)</span>"</span></span></code></pre> then qemu:</p>`
`brew</span> install qemu</span></span> qemu-system-aarch64</span> --version</span></span></code></pre> prepare the disk:</p>`
IMG</span>=</span>9front.arm64.qcow2</span></span> gunzip</span> -c 9front-11952.arm64.qcow2.gz</span> ></span> "</span>$IMG</span>"</span></span></code></pre> the manual says 9front on arm needs a u-boot as a bootloader, and that it needs to be built from sources. on arm mac this would require a troublesome setup with gnu tools for cross-compilation, some kind of build container and an emulation of x64 through rosetta. that’s hella annoying.</p> instead, let’s just use the binary provided by the lesbian linux:</p>
mkdir</span> -p .tmp/</span> &&</span> cd</span> .tmp/</span></span> # fetch it</span></span> UBOOT_DEB</span>=</span>u-boot-qemu_2025.01-3+deb13u1_all.deb</span></span> UBOOT_DEB_URL</span>=</span>https://deb.debian.org/debian/pool/main/u/u-boot/</span>$UBOOT_DEB</span></span> curl</span> -fsSLO "</span>$UBOOT_DEB_URL</span>"</span></span> ar</span> x "</span>$UBOOT_DEB</span>"</span></span> tar</span> xf data.tar.xz</span></span> # get the binary and cleanup</span></span> mv</span> ./usr/lib/u-boot/qemu_arm64/u-boot.bin ../</span></span> cd</span> ../</span></span> rm</span> -rf .tmp</span></span></code></pre> now we can try the initial boot with the parameters more or less close to the ones recommended by the manual:</p>
`qemu-system-aarch64</span> \</span></span> -M "virt-11.0,gic-version=3,highmem-ecam=off,accel=hvf"</span> \</span></span> -cpu host -m 4G -smp</span> 4</span> \</span></span> -bios u-boot.bin</span> \</span></span> -drive "file=</span>$IMG</span>,if=none,id=installer"</span> \</span></span> -device virtio-blk-pci-non-transitional,drive=installer</span></span> -display</span> none -serial stdio</span></span></code></pre> expected output is:</p>`
## Transferring control to Plan 9 (at address 40100000) ...</span></span> 127 holes free</span></span> 0x40523000 0x7ffff000 1068351488</span></span> 1068351488 bytes free</span></span> </span> Plan 9</span></span> timer frequency 24000000 Hz</span></span> cpu0: 1000MHz QEMU</span></span> 4095M memory: 1023M kernel data, 3072M user, 15344M swap</span></span> cpu1: 1000MHz QEMU</span></span> cpu2: 1000MHz QEMU</span></span> cpu3: 1000MHz QEMU</span></span> qemu.bootfs: Thu Jul 30 07:19:08 GMT 2026</span></span> fingerprint: e6b8428c2f58f159a7ddfe8c9cae52a29ca841a9</span></span> nusb/usbd: no hubs</span></span> </span> /dev/sdF0:</span></span> /dev/sdF0/data </span></span> /dev/sdF0/dos dos</span></span> /dev/sdF0/fs hjfs</span></span> /dev/sdF0/nvram</span></span> /dev/sdF0/plan9</span></span> bootargs is (tcp, tls, il, local!device)[local!/dev/sdF0/fs]</span></span></code></pre> here we can just proceed with the guide, press enter a couple of times with the default prompts, and get to the shell:</p>
bootargs is (tcp, tls, il, local!device)[local!/dev/sdF0/fs]</span></span> user[glenda]:</span></span> hjfs: fs is /dev/sdF0/fs</span></span> </span> init: starting /bin/rc</span></span> ...</span></span> % ls</span></span> bin</span></span> lib</span></span> tmp</span></span></code></pre> but it’s just the "livecd", not the real install. let’s fix that by killing the VM and restarting with a user disk image. in the guest shell:</p> % fshalt</span></span> </span> halting.../srv/hjfs.cmd...hjfs: ending</span></span> </span> done halting</span></span> aux/acpi: unknown device in # filename</span></span> It's now safe to turn off your computer</span></span> ^C qemu-system-aarch64: terminating on signal 2 from pid 658 (<unknown process>)</span></span></code></pre> now in the host shell:</p> DISK</span>=</span>9front.qcow2</span></span> qemu-img</span> create -f qcow2 "</span>$DISK</span>" 64G </span></span> # and restart the VM</span></span> qemu-system-aarch64</span> \</span></span> -M "virt-11.0,gic-version=3,highmem-ecam=off,accel=hvf"</span> \</span></span> -cpu host -m 4G -smp</span> 4</span> \</span></span> -bios u-boot.bin</span> \</span></span> -drive "file=</span>$IMG</span>,if=none,id=installer"</span> \</span></span> -device virtio-blk-pci-non-transitional,drive=installer</span></span> -display</span> none -serial stdio</span> \</span></span> -drive "file=</span>$DISK</span>,if=none,id=disk"</span> \</span></span> -device virtio-blk-pci-non-transitional,drive=disk</span></span></code></pre>actual installation</h2> now we can just follow «FQA 4.3 − Performing a simple install» section. mostly. by default cwfs64x setup skipped disk partition for me for some reason, so mountfs fails.</p> % inst/start</span></span> Thu Aug 20 22:08:00 CET 2026 Installation process started</span></span> </span> Preparing menu...</span></span> The following unfinished tasks are ready to be done:</span></span> configfs - choose the type of file system to install</span></span> confignet - configure the network</span></span> tzsetup - choose time zone</span></span> </span> Task to do [configfs]:</span></span> </span> --------------------------------------</span></span> You can install the following types of file systems:</span></span> </span> cwfs64x cached-worm file server for large disks</span></span> hjfs file server for small disks</span></span> gefs good enough file system (experimental!)</span></span> </span> File system (cwfs64x, hjfs, gefs)[cwfs64x]:</span></span> </span> --------------------------------------</span></span> </span> ...</span></span> The following unfinished tasks are ready to be done:</span></span> mountfs - choose and mount file system partition</span></span> confignet - configure the network</span></span> tzsetup - choose time zone</span></span> </span> Task to do [mountfs]:</span></span> </span> --------------------------------------</span></span> </span> Please choose your cwfs64x partitions</span></span> </span> Cwfs cache partition [no default]:</span></span></code></pre> so instead of just falling through let’s choose the skipped steps and approve the defaults.</p> it looks scary but in reality it’s just typing out a couple of words and pressing enter.</p> Preparing menu...</span></span> The following tasks are done:</span></span> configfs - choose the type of file system to install</span></span> partdisk - edit partition tables (e.g., to create a plan 9 partition)</span></span> prepdisk - subdivide plan 9 disk partition</span></span> </span> The following unfinished tasks are ready to be done:</span></span> mountfs - choose and mount file system partition</span></span> confignet - configure the network</span></span> tzsetup - choose time zone</span></span> </span> Task to do [mountfs]: partdisk</span></span> </span> --------------------------------------</span></span> The following disk devices were found.</span></span> </span> sdF0 -</span></span> empty 0 1 (1 cylinders, 3.93 MB)</span></span> * p1 1 25 (24 cylinders, 94.50 MB) FAT32</span></span> p2 25 957 (932 cylinders, 3.58 GB) PLAN9</span></span> </span> sdG0 -</span></span> </span> Disk to partition (sdF0, sdG0)[no default]: sdG0</span></span> The disk you selected HAS NO master boot record on its first sector.</span></span> (Perhaps it is a completely blank disk.)</span></span> Shall we create a blank EFI partition table (GPT)</span></span> or install traditional DOS partition table (MBR)?</span></span> </span> Install mbr or gpt (mbr, gpt)[no default]: mbr</span></span> </span> This is disk/fdisk; use it to create a Plan 9 partition.</span></span> If there is enough room, a Plan 9 partition will be</span></span> suggested; you can probably just type 'w' and then 'q'.</span></span> </span> cylinder = 8225280 bytes</span></span> '* p1 0 8354 (8354 cylinders, 63.99 GB) PLAN9</span></span> >>> w</span></span> >>> q</span></span> </span> --------------------------------------</span></span> </span> Preparing menu...</span></span> The following tasks are done:</span></span> configfs - choose the type of file system to install</span></span> partdisk - edit partition tables (e.g., to create a plan 9 partition)</span></span> prepdisk - subdivide plan 9 disk partition</span></span> </span> The following unfinished tasks are ready to be done:</span></span> mountfs - choose and mount file system partition</span></span> confignet - configure the network</span></span> tzsetup - choose time zone</span></span> </span> Task to do [mountfs]: prepdisk</span></span> </span> --------------------------------------</span></span> The following Plan 9 disk partitions were found.</span></span> </span> /dev/sdF0/plan9</span></span> empty 0 2 (2 sectors, 1.00 KB)</span></span> nvram 2 3 (1 sectors, 512 B )</span></span> fs 3 7515647 (7515644 sectors, 3.58 GB)</span></span> empty 7515647 7515648 (1 sectors, 512 B )</span></span> </span> /dev/sdG0/plan9</span></span> empty 0 134206947 (134206947 sectors, 63.99 GB)</span></span> </span> Plan 9 partition to subdivide (/dev/sdF0/plan9, /dev/sdG0/plan9)[no default]: /dev/sdG0/plan9</span></span> This is disk/prep; use it to subdivide the Plan 9 partition.</span></span> If it is not yet subdivided, a sensible layout will be suggested;</span></span> you can probably just type 'w' and then 'q'.</span></span> </span> no plan9 partition table found</span></span> 9fat 204800</span></span> nvram 1</span></span> other 19143163</span></span> fscache 19143163</span></span> fsworm 95715820</span></span> ' 9fat 0 204800 (204800 sectors, 100.00 MB)</span></span> ' nvram 204800 204801 (1 sectors, 512 B )</span></span> ' other 204801 19347964 (19143163 sectors, 9.12 GB)</span></span> ' fscache 19347964 38491127 (19143163 sectors, 9.12 GB)</span></span> ' fsworm 38491127 134206947 (95715820 sectors, 45.64 GB)</span></span> >>> w</span></span> >>> q</span></span> </span> --------------------------------------</span></span> </span> Preparing menu...</span></span> The following tasks are done:</span></span> configfs - choose the type of file system to install</span></span> partdisk - edit partition tables (e.g., to create a plan 9 partition)</span></span> prepdisk - subdivide plan 9 disk partition</span></span> </span> The following unfinished tasks are ready to be done:</span></span> mountfs - choose and mount file system partition</span></span> confignet - configure the network</span></span> tzsetup - choose time zone</span></span> </span> Task to do [mountfs]:</span></span> </span> --------------------------------------</span></span> </span> Please choose your cwfs64x partitions</span></span> </span> --rw-r----- S 0 glenda glenda 9801299456 Jul 30 09:19 /dev/sdG0/fscache</span></span> </span> Cwfs cache partition (/dev/sdG0/fscache)[/dev/sdG0/fscache]:</span></span> --rw-r----- S 0 glenda glenda 49006499840 Jul 30 09:19 /dev/sdG0/fsworm</span></span> </span> Cwfs worm partition (/dev/sdG0/fsworm)[/dev/sdG0/fsworm]:</span></span> --rw-r----- S 0 glenda glenda 9801299456 Jul 30 09:19 /dev/sdG0/other</span></span> </span> Cwfs other partition (/dev/sdG0/other)[/dev/sdG0/other]:</span></span> Ream the filesystem? (yes, no)[yes]:</span></span> ...</span></span> % mount -c /srv/cwfs.newfs /n/other other</span></span></code></pre> the next step is configuring the network which we don’t actually need since we didn’t start the VM with network device and we have all the dist files in the installation image:</p> ...</span></span> The following unfinished tasks are ready to be done:</span></span> confignet - configure the network</span></span> mountdist - locate and mount the distribution</span></span> tzsetup - choose time zone</span></span> </span> Task to do [confignet]:</span></span> </span> --------------------------------------</span></span> </span> We will configure the ethernet.</span></span> </span> </span> Please choose a method for configuring your ethernet connection.</span></span> </span> manual - specify IP address, network mask, gateway IP address</span></span> automatic - use DHCP and SLAAC to automatically configure</span></span> </span> Configuration method (manual, automatic)[automatic]:</span></span> starting ethernet automatic config</span></span></code></pre> now choose the distribution mount (/) and the OS files will be copied to the disk:</p> …</span></span> Task to do [mountdist]:</span></span> </span> --------------------------------------</span></span> Please wait... Scanning storage devices...</span></span> /dev/sdF0/data</span></span> /dev/sdF0/dos</span></span> /dev/sdF0/fs</span></span> /dev/sdG0/9fat</span></span> /dev/sdG0/data</span></span> /dev/sdG0/fscache</span></span> /dev/sdG0/fsworm</span></span> /dev/sdG0/other</span></span> </span> The following storage media were detected.</span></span> Choose the one containing the distribution.</span></span> </span> /dev/sdF0/dos (microsoft fat)</span></span> </span> Distribution disk (/dev/sdF0/dos, /dev/sdG0/fscache, /)[no default]: /</span></span> % mount /srv/boot /n/distmedia</span></span> </span> Which directory contains the distribution?</span></span> </span> Location of archives [/]:  … The following unfinished tasks are ready to be done:</span></span> copydist - copy the distribution into the file system</span></span> tzsetup - choose time zone</span></span> </span> Task to do [copydist]:</span></span> </span> --------------------------------------</span></span> processing /sys/lib/sysconfig/proto/allproto</span></span> file system made</span></span> </span> --------------------------------------</span></span></code></pre> next steps are trivial, just pressing enter some more:</p> …</span></span> Task to do [ndbsetup]:</span></span> </span> --------------------------------------</span></span> </span> Setup network configuration</span></span> </span> sysname [cirno]: … The following unfinished tasks are ready to be done:</span></span> tzsetup - choose time zone</span></span> bootsetup - configure hard disk to boot plan 9</span></span> </span> Task to do [tzsetup]:</span></span> </span> --------------------------------------</span></span> </span> Setup Time Zone</span></span> </span> Time Zone (Argentina, Asia_Bahrain, Asia_India, Asia_Jakarta, Asia_Jayapura, Asia_Makassar, Asia_Pontianak, Australia_ACT, Australia_Broken-Hill, Australia_LHI, Australia_NSW, Australia_North, Australia_Queensland, Australia_South, Australia_Sturt, Australia_Tasmania, Australia_Victoria, Australia_West, Australia_Yancowinna, Brazil_Acre, Brazil_DeNoronha, Brazil_East, Brazil_West, CET, Canada_Atlantic, Canada_Central, Canada_East-Saskatchewan, Canada_Eastern, Canada_Mountain, Canada_Newfoundland, Canada_Pacific, Canada_Yukon, Chile_Continental, Chile_EasterIsland, Cuba, EET, Egypt, GB-Eire, GMT, HST, Hongkong, Iceland, Iran, Israel, Jamaica, Japan, Libya, Mexico_BajaNorte, Mexico_BajaSur, Mexico_General, NZ, NZ_CHAT, Navajo, PRC, Poland, ROC, ROK, Singapore, Turkey, US_Alaska, US_Arizona, US_Central, US_East-Indiana, US_Eastern, US_Hawaii, US_Michigan, US_Mountain, US_Pacific, US_Yukon, Uruguay, W-SU, WET)[US_Eastern]: CET</span></span></code></pre> finishing up and making the user installation boot:</p> The following unfinished tasks are ready to be done:</span></span> bootsetup - configure hard disk to boot plan 9</span></span> </span> Task to do [bootsetup]:</span></span> …</span></span> If you use the Windows NT/2000/XP master boot record</span></span> or a master boot record from a Unix clone (e.g., LILO or</span></span> FreeBSD bootmgr), it is probably safe to continue using</span></span> that boot record rather than install the Plan 9 boot record.</span></span> </span> Install the Plan 9 master boot record (yes, no)[no default]: yes</span></span> Mark the Plan 9 partition active (yes, no)[no default]: yes</span></span> </span> The Plan 9 partition is now marked as active.</span></span> </span> --------------------------------------</span></span> …</span></span> Task to do [finish]:</span></span> </span> --------------------------------------</span></span> Congratulations; you've completed the install.</span></span> </span> </span> halting...hjfs: ending</span></span> /srv/cwfs.newfs.cmd...halted at Thu Aug 20 22:34:01 2026.</span></span> /srv/hjfs.cmd...</span></span> done halting</span></span> rebooting...</span></span> cpu0: exiting</span></span> cpu1: exiting</span></span> cpu3: exiting</span></span> cpu2: exiting</span></span> </span> </span> U-Boot 2025.01-3+deb13u1 (Jun 24 2026 - 09:07:15 +0000)</span></span> ….</span></span> bootargs is (tcp, tls, il, local!device)[local!/dev/sdF0/fs]</span></span></code></pre> it prompts to boot from the installation image again after the reboot, so let’s choose the disk with the system we just installed (on sdG0):</p> /dev/sdF0:</span></span> /dev/sdF0/data </span></span> /dev/sdF0/dos dos</span></span> /dev/sdF0/fs hjfs</span></span> /dev/sdF0/nvram</span></span> /dev/sdF0/plan9</span></span> /dev/sdG0:</span></span> /dev/sdG0/9fat dos</span></span> /dev/sdG0/data </span></span> /dev/sdG0/fscache cwfs64x</span></span> /dev/sdG0/fsworm </span></span> /dev/sdG0/nvram</span></span> /dev/sdG0/other </span></span> /dev/sdG0/plan9</span></span> bootargs is (tcp, tls, il, local!device)[local!/dev/sdF0/fs] local!/dev/sdG0/fscache</span></span> user[glenda]:</span></span> current fs is "main"</span></span> …</span></span> init: starting /bin/rc</span></span> % ls</span></span> acme.dump</span></span> bin</span></span> lib</span></span> tmp</span></span></code></pre> all successful! more or less.</p> boot partition hurdle</h2> unfortunately i couldn’t make it boot without the installation image as u-boot fails to find the FAT boot partition we made during [bootsetup] step. 9front makes that «9fat» partition that should theoretically work, but it doesn’t:</p> % qemu-system-aarch64 \</span></span> -M virt-11.0,gic-version=3,highmem-ecam=off,accel=hvf \</span></span> -cpu host -m 4G -smp 4 -bios u-boot.bin \</span></span> -drive file=9front.qcow2,if=none,id=disk \</span></span> -device virtio-blk-pci-non-transitional,drive=disk \</span></span> -display none -serial stdio</span></span> </span> U-Boot 2025.01-3+deb13u1 (Jun 24 2026 - 09:07:15 +0000)</span></span> </span> DRAM: 4 GiB</span></span> Core: 51 devices, 14 uclasses, devicetree: board</span></span> Flash: 64 MiB</span></span> Loading Environment from Flash... * Warning - bad CRC, using default environment</span></span> </span> In: serial,usbkbd</span></span> Out: serial,vidconsole</span></span> Err: serial,vidconsole</span></span> No USB controllers found</span></span> Net: eth0: virtio-net#32</span></span> </span> starting USB...</span></span> No USB controllers found</span></span> Hit any key to stop autoboot: 0</span></span> Scanning for bootflows in all bootdevs</span></span> Seq Method State Uclass Part Name Filename</span></span> --- ----------- ------ -------- ---- ------------------------ ----------------</span></span> Scanning global bootmeth 'efi_mgr':</span></span> Cannot persist EFI variables without system partition</span></span> Missing TPMv2 device for EFI_TCG_PROTOCOL</span></span> Missing RNG device for EFI_RNG_PROTOCOL</span></span> 0 efi_mgr ready (none) 0 <NULL></span></span> Booting bootflow '<NULL>' with efi_mgr</span></span> Loading Boot0000 'virtio 0' failed</span></span> EFI boot manager: Cannot load any image</span></span> Boot failed (err=-14)</span></span></code></pre> probably there could be a way to fix it by prepending a FAT partition to the system disk and reproducing what the installation’s «/dev/sdF0/dos» contains but it’s too challenging for me rn. maybe someone who actually knows what they’re doing and has more experience can do that or fix the installer itself.</p> i choose to just keep booting with both the installer disk and system disk then choosing the system disk on boot xD</p> booting with network and getting to GUI</h2> arm 9front doesn’t have graphics, but luckily we don’t need them where we’re going! instead we’ll use the amazing feature of the OS - native remote desktop. kind of. it’s not really a remote desktop in a way we understand the term now, more like a graphical terminal. anyway:</p> # get drawterm sources</span></span> git</span> clone git://git.9front.org/plan9front/drawterm</span></span> cd</span> drawterm</span></span> CONF</span>=</span>osx-cocoa</span> make</span> -j10</span></span></code></pre> now we need the OS to accept the connections. let’s add network device and forward the port that drawterm uses:</p> qemu-system-aarch64</span> \</span></span> -M virt-11.0,gic-version=3,highmem-ecam=off,accel=hvf</span> \</span></span> -cpu host -m 4G -smp</span> 4</span> -bios u-boot.bin</span> \</span></span> -drive file=9front.arm64.qcow2,if=none,id=boot</span> \</span></span> -device virtio-blk-pci-non-transitional,drive=boot,bootindex=</span>0</span> \</span></span> -drive file=9front.qcow2,if=none,id=disk</span> \</span></span> -device virtio-blk-pci-non-transitional,drive=disk</span> \</span></span> -display none -serial stdio</span> \</span></span> -netdev user,id=net0,hostfwd=tcp::17019-:17019</span> \</span></span> -device "virtio-net-pci-non-transitional,netdev=net0"</span></span></code></pre> then once booted:</p> % # start the network</span></span> % ip/ipconfig</span></span> % # set auth params</span></span> % echo 'key proto=dp9ik dom=9front user=glenda !password=pass' > /mnt/factotum/ctl</span></span> % # start listener socket</span></span> % aux/listen1 -t 'tcp!*!rcpu' /rc/bin/service/tcp17019 &</span></span> % listen started</span></span></code></pre> and on the host:</p> # enter «pass» when prompted</span></span> ./drawterm</span> -h localhost -a localhost -u glenda -x1</span></span></code></pre> you'll be greeted with rc shell, the same that we had in qemu serial. to get to actual gui type in:</p> cpu % rio</span></span></code></pre> this will give you the original 9front experience and now you can familiarize yourself with 9p protocol and mouse chording, and never look the same at the "modern" systems and their backwards ways. have fun exporing xD</p> </a> drawterm with rio desktop</figcaption> </figure> </div> next steps</h2> what i can't recommend enough is going through adventuresin9 channel on youtube. that dude has an amazing set of videos that can help anyone getting into the system quickly. especially:</p> introduction to user interface</a></li> acme text editor</a></li> graphical interface</a></li> customizing rio</a></li> </ul>



budapest august 2025
byte — Sun, 19 Apr 2026 00:00:00 +0000

  
    
        
            
        </a>
        
            OwO</figcaption>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
            ganz KCSV–7 my beloved</figcaption>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
            someone has a "green thumb"</figcaption>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
</div>


swiss february 2026
byte — Sun, 19 Apr 2026 00:00:00 +0000

  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
            it's their turf</figcaption>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
    </figure>
  
    
        
            
        </a>
        
            do you have games on your phone?</figcaption>
        
    </figure>
  
    
        
            
        </a>
        
            i'm one of you, i swear</figcaption>
        
    </figure>
  
</div>


Trans Rules Of Engagement
Florence Ashley — Fri, 11 Apr 2025 00:00:00 +0000
Strong communities make us all safer. As anti-trans movements gain in power and influence, holding space for each other through our flaws remains critical. Yet the very conditions that create our need for community care make it hard for us to care for each other. We are raw, wounded, traumatized, and hypervigilant. We make mistakes brought on by fear and hurt. We lash out at each other when we do wrong, often partaking in pile-ons facilitated by the synchronous nature of online interactions. Whether we realize it or not, we often exclude trans people from community when they need it most.</p>
I have lost count of the number of trans people I have seen cast out of online trans spaces for misdeeds both major and minor—far too often with my help. I sometimes find myself wondering where they are now and whether they are still alive. Because, as Kai Cheng Thom has taught us, social death often means real death for trans people. Trans communities are life-sustaining in a world that hates us so, so much. In a world that wants us dead. We have lost too many people not to stop and think about how we can foster life among each other.</p>
This goal I have for myself—that of fostering life—motivates the following principles and rules for engaging in online intra-community conflicts while preserving the life-sustaining spirit of our communities. Countless times have I failed to heed these principles and ignored these rules. This failure, which many of us share, is precisely why I now want to lay these principles and rules down on paper. If only as a reminder of my aspirations. The principles and rules are meant to be adopted for oneself, not imposed onto others. Their purpose is to foster productive engagement, not create even more conflict and rigidity. I hope that this will be a living document, and invite you to make your own version if you would like. Borrow what is useful, supplement with what is needed, alter what can be improved.</p>
Some, and perhaps all, of the principles I acknowledge are false, hence the need for a living document. Each of my suggested rules have exceptions. In setting them out, I am staking a claim as to the sort of myths and half-truths that are necessary to sustain life in a world that wants us dead. We must treat them as true if we wish to foster life-sustaining communities and survive the hellscape we belabor.</p>
Principles</h2>


We are all flawed, traumatized humans at the end of their rope. Many of our actions say more about the conditions we live under than who we are as people.</p>
</li>

No one is disposable. No one is unsalvageable.</p>
</li>

Life holds greater value than being right or comfortable. Hurt is preferable to death.</p>
</li>

No one should be deprived of community.</p>
</li>

Harm does not require further harm. Punishment does not equate protection or healing.</p>
</li>
</ol>
Rules</h2>


Do not depart from these rules, unless you have to.</p>
</li>

Morgan M. Page’s Rule: Try to avoid criticizing other trans people in public. The world does it enough already.</p>
</li>

Favor in person or private conversations: Addressing someone’s comments or actions in person or privately is typically more constructive and effective. It allows you to communicate more cogently and with more nuance problems in someone’s actions or words and because it is less likely to make them react defensively from a place of trauma or fear.</p>
</li>

Take your time: Few things require an immediate response. Responding while caught in a surge of thoughts and feelings is often unproductive. Ask yourself how much harm was done, versus how much we are reminded of an earlier harm. Ask whether your response is rooted in misperception or potential biases towards the person due to race, disability, gender, or other marginalized identities. Consider whether their words or actions reflect a different kind of thinking or communication style, a lack of access to education, or limited access to progressive communities and norms. You can respond tomorrow, once you have collected your thoughts, talked to others, and gained perspective.</p>
</li>

Don’t mob: Be aware of group dynamics. Ask yourself if you are connected to this person and in community with them. Avoid jumping into the fray when others are already criticizing the person. Do not invite others to join in and mob them. Withdraw if others join in, and kindly ask people to stay conscious of mobbing dynamics. Mobbing rapidly grows out of proportion.</p>
</li>

De-escalate: Focus on de-escalating conflicts. Ask what people mean or want, and why. Ask them for clarification or elaboration if needed. Ask yourself if you know enough about the context of the situation. Distinguish the action from the person, and acknowledge that it is normal to respond defensively or aggressively to public criticism and mobbing. People are traumatized, mentally ill, and are scared of losing the little social support they have. As a result, conflict can trigger a fight-or-flight response in both those who are criticized and who criticize, which leads to escalating conflict and ends in a loss of community. Dropping the conversation to return at a later date is preferable to escalation. Often, I find it best to limit myself to three replies in conversations that aren’t constructive.</p>
</li>

Respond proportionately: Responses to words and behaviours should be proportionate to their harm, and reflect a need for healing and protection rather than punishment. When we speak from a place of hurt, we can understandably but unfortunately forget the measure and impact of our response. Use language that reflects the nuances and gradations of harm rather than a coarse good and evil binary. Cutting all social support and community banishment are rarely a proportionate response, even for someone who doubles down and does not apologize. Responding proportionately is asking first and foremost what response sustains rather than dissolves life. Especially when it comes to words, it is better to under-react than to over-react.</p>
</li>

Ensure support for everyone: Check in on those who are criticized and those who criticize them. Remind people that we are all in this together, and that banishment is not how we work as a community. Everyone deserves to have their needs met. Do not shun or reproach people who offer support to those who were criticized or called out. Distinguish supporting a person from enabling their behavior.</p>
</li>

Hold space for people to grow: Allow space for people to be accountable, change, and move on from previous conflicts. Do not hold past behavior over people’s head, nor dig up past misdeeds to fuel present conflicts.</p>
</li>

Resolve conflict and harm as a community: We must ask how our communities enable and cause hurt and harm, and find ways to transform the conditions that create them. Holding accountable, problem-solving, and conflict resolution are functions that should be taken up by the collective, not isolated and unsupported individuals.</p>
</li>

Center those most hurt or harmed: Focus on supporting and empowering people who are hurt and harmed rather than on punishment. Ask what they need to be safe and integrated in our communities, while committing to support for everyone; what they need to repair their relationship to the person who hurt or harmed them. Focus your involvement on bringing people together, fostering dialogue and mutual understanding, and restoring a sense of community togetherness, rather than deciding who is right or wrong.♦</p>
</li>
</ol>