the goblin tinker's workshop logbook — found pages, taped in.
fig. 00 — crt, still humming, west window 4pm
entry vii — workbench still warm — coffee at 4 — antenna repaired — lichen patient
i.
on the fifth of nevermarch
found saturday by the loading dock — see fig. 01
the keycap was alone, lying face-down in a puddle of yesterday's rain. P. just the letter P. white plastic gone the colour of weak tea, with one corner moth-bitten as if a small thing had tried to read it. i took it home in my coat pocket. it was warm by the time i unlocked the workshop door, which is the kind of warmth you get only from solitary objects that have been thinking. i set it on the bench under the west window. the light at four o'clock turned it amber and i understood it was the first of seven.
ii.
thawday last, behind the bins
cassette, exploded — fig. 02 — side a only
a tdk d-90 in three pieces. the shell cracked clean along the seam, the felt pad still glued to the door, the tape ribbon spooled out and tangled in a tuft of moss. i wound it back by hand using a pencil, the way one was taught. side a survived. side b is somewhere in the alley still, probably in a starling's nest by now — they like the shine. i played side a on the recorder (fig. 04) once it was working again. fourteen minutes of someone humming the same four-bar phrase, then a cough, then silence. i suspect it was the previous goblin.
iii.
a coil that thought it was a snail
twelve feet of cat5 — fig. 03
i found it perfectly coiled at the bottom of a bucket, RJ-45 tucked into the centre like an antenna. it had no business being shaped that way. cables do not coil themselves. and yet. it had been raining for three days and the bucket was full of leaves and one earthworm, who was unbothered. i lifted it out and the coil held, a perfect spiral the diameter of a saucer. the seventh puzzle came in a bread bag, but the third arrived in a bucket and refused to be straightened. it now hangs on a nail above the bench, still spiralled, still patient.
iv.
the recorder with one wheel
tape recorder, half-open — fig. 04
the right capstan was missing entirely. someone had unscrewed it and left the screw in the empty socket — a small kindness. the motor still hummed when i fed it 9V, a low contented hum, like a cat that has eaten. i fashioned a replacement wheel from a wooden bead and a sliver of rubber band. it was wrong by 0.3mm but the tape did not seem to mind. fig. 02 played through it on the third try. i listened with the door closed and the kettle off. there is a particular silence that a workshop has when an old machine is working again, and it is the closest a goblin gets to prayer.
v.
the green tube — 5v, drifting
oscilloscope, working — fig. 05
the screen is the colour of bottle glass held against a candle. when powered, the trace draws a sine wave that drifts upward by a millimetre each minute, as if the workshop itself were tilting. it isn't. i checked with a level. the drift is real, internal, and probably indicates a failing capacitor in the timebase, but it has been failing for as long as i have owned the instrument and i have decided this is its character now. i no longer try to compensate. when the trace reaches the top of the graticule i simply turn the unit off and on, and we begin again, and again, and again.
vi.
a stack of disks bound in twine
3.5" floppies — fig. 06
eleven disks, bound horizontally with garden twine in a knot i did not invent. the topmost label, in a child's hand: PUZZLE MASTERS. the rest unlabelled. i untied the knot in the only sequence the twine permitted and the disks fell open like a hand of cards. eight of them spin. three click. of the eight, two contain readable data: a list of seven items in plain text, and a single bitmap of a snail with a soldering-iron shell. i have not told anyone about the bitmap because telling diminishes finds. but it is the patron-glyph of the workshop now, taped above the door.
vii.
the seventh puzzle came in a bread bag
pcb, salvaged — fig. 07
a paper bag from the bakery on glass street, folded twice, taped with masking tape, addressed in pencil to NO ONE IN PARTICULAR. inside: a single matchbox-sized PCB, rev 02, dated '94, with one capacitor visibly bulging in the way capacitors do when they have decided to retire. there is no enclosure. there is no documentation. the traces lead, eventually, to a single header pin labelled with a faded inkstamp: start. i have not connected power to it yet. i am saving it for a saturday when the workshop is warm and the kettle is full and i have nothing else to do but find out what the seventh puzzle wants.