"The old saying is: 'Trust is earned, not given.' (…) But in my experience trust rarely starts out at zero."
—Eufemia Krainski, humanitarian aid worker
#3L573 continued what it had already started out doing, scanning through all the more immediately useful spare parts the two drones had gathered or cut from the broken drones before. In this case this translated to manipulators and legs.
A few of them were various levels of workable, while a couple were in just as bad a shape as #3L573's old diagnostic. A disturbing fraction actually, which tripped the drone's algorithms. Its own manipulator had probably snapped due to a bad fall when whatever had disabled #3L573 had happened. Its chassis was dented on that side and the manipulator's mount had internal stress damage that seemed to corroborate this.
But these others? They were
also broken. Not in the same place necessarily, nor as badly, but they all looked as if the respective drones had all crushed their manipulators as they landed, or crashed onto them. It was unusable and notable and thereby
weird. Unless they were working or holding onto things, manipulators were folded into inlays in the chassis. Even if they weren't because they were holding things, drones could still fly somewhat straight and steady. They could flail wildly and still fly reasonably straight. Not that drones would, of course; it was wasteful and inefficient and only risked damaging in some way. It didn't matter much what would be damaged, the drone, the manipulator, or people. Management didn't like it when these kinds of damages happened and insisted on it avoiding them when possible.
Not that there was currently much of either around people or drones, but the point stood.
#3L573 went through the parts, scanning each one systematically. Considerable time passed as it did, all things considered. In turn manipulators and legs were thoroughly checked and sorted. The collection of usable parts was laid out on the recharge station, in reach of #3200, and the unusable went onto the ever growing [recycle] pile. #3L573 had taken its sibling's perimeter reconnaissance as a sign that it wouldn't start flying off randomly for now, trusting it not to scatter parts again. Plus they would be more readily accessible too.
In the end the systematic analysis yielded good results.
There were three general manipulators in working condition, while the remaining were only good for spare parts. After the experience with its new diagnostic arm, #3L573 checked their other manipulator's internal software as well and only two were showing similarly faulty software. The drone repeated the procedure from before, wiping and overwriting those packages with less buggy ones, turning two weird manipulators into functioning ones.
Only one of the maintenance manipulators was workable, and only after a similar overwrite of its systems package with one of the other three's. But it
was workable, which was enough. It didn't have the full range of motion, nor the full strength of a new manipulator, but this was still magnitudes better than having useless metal attached to the chassis.
Lastly were the legs. Well, more 'legs' than legs. These small manipulators were typically very strong for their size, capable of holding more than double the drone's weight on their own in standard gravity. Their intentional use was for a drone to attach and affix itself to a surface while it performed repairs with the proper manipulators. Actual transport drones and delivery bots got around on wheels or be turbines. In lieu of either these short and stubby 'legs' were a poor drone's substitute. They're like a caterpillar's legs, only much, much less numerous. But they
did attach quite well to even smooth surfaces. Usually. Their current insufficiency was due to two issue: One, they were old and weathered and well behind on their maintenance cycles; two, it was still really cold in the storage compartment. The legs' operational range only extended so far, and the current temperatures were well outside of habitat ranges. The gecko pads were so brittle they'd break after a few uses, a fact #3L573 had had the experience of itself when it had tried to clamber of the chair at the worktop and only managed a short stretch.
So while these legs weren't all broken and unusable as such—though some of them were and also landed on the [recycle] pile—they weren't generally notably better than #3L573's. They'd work fine for ground walking, but spidering up walls and ceiling wasn't doable unless the ambient temperature was dozens of Kelvin higher.
Maybe #3200 could make use of them; the drone had barely walked since its activation.
The other drone was still idling in its drone comb cell. Well, not for long. With a workable maintenance manipulator available #3L573 set its sibling to task. It was a bit bothersome to still have to tether through the recharge station, but doable for now.
#3200 flew over, visually inspecting the manipulators laid out and then grabbed the ones #3L573 had indicated. Usually it would have given the manipulator a cursory check before installing it, but without a functioning diagnostic manipulator of its own, there wasn't much it could do past a visual inspection. The replacement was a quiet affair, with little communications exchanged between the drones. Neither were talkative by habit or nature, which amounted to programming in the latter, and only one of them could develop the former anyway. So they only sent the occasional request for position changes and verification on internal reports and systems checks.
When it was done #3L573 ran the maintenance arm through various tests, noting how much the installed arm was different from the drone's expectations putting it under minor stresses, and adjusting its expectations as the results came in. The arm would serve, as predicted.
But #3200 wasn't done yet. There was still the general manipulator lying at the ready. The drone didn't pick up immediately; the currently mounted and unresponsive arm would have to be removed first. The first sign of there being something wrong was the difficulty the drone reported in getting the arm out of its socket. Once it was out the problem was obvious; the socket had been crushed and half pushed into the chassis. What was a originally a relatively minor fix turned into a major repair job. #3200 had to really go to town. It decoupled the connectors of the socket, then sliced through the bars connecting it to the chassis.
#3L573's risk assessment was jumping notably with every new cut applied and jerk shuddering through it. Too put it in other words: the drone was more than a bit worried about this development. The mounts were technically surface structures, but replacing them still required removing cowling and working on the chassis itself. The procedure was invasive, it weakened the overall structure, and it exposed more of the drone's innards than it was comfortable with.
But if it had to bear this to regain its functionality, then #3L573 would bear it. The worried drone still cranked up its inner heating element; there was no need to risk freezing its innards.
The repairs went smoothly, the broken mount was cut out and then… #3200 stopped. It looked around for something. Catching sight of it, the drone flew as close to the perimeter as it could, then halted. Even with outstretched manipulators it wouldn't be able to reach its goal. The stack of messages it kept pushing to #3L573 for tethering to the station changed a tiny bit.
#3L573 had continuously ignored its sibling's urge for communication with the Station, but now it snagged the useful bit of information out of the stack. It was in need of a mount for the manipulator to continue. The half-opened drone trotted over and scanned through the small mount of mounts they had gained by disassembling the broken drones. After having selected a few the drone returned to the recharger and plopped down. #3200 had watched this with what amounted to interest, and immediately picked up one of the mounts when it was clear its sibling didn't need it itself. Resuming the repair it sheared a bit more material off the internal struts of #3L573 and then welded the new mount into place.
After the repair was completed and manipulator was mounted, #3L573 ran the new general arm through a smaller gamut of test, and all returned positive. Or positive enough at least.
What should #3L573 do next?
[] Install/Replace: diagnostic manipulator in #3200
[] Install/Replace: general manipulator in #3200
[] Thorough check of #3200.
[] Install salvaged power banks in #3200
[] Explore: Storage level (have to climb stairs)
[] Diagnose: outlying lumps of stuff (& gather as appropriate)
[] Diagnose: spare parts
[] Construct mobile ramp (or step-ladder analogue) from available material
[] Do something else (what?)
Note:
Sorry about the delay. Part is due to what I said before, but another is simply dissatisfaction with how things are going. So far it seems as if this quest gets lost in micromanaging of events that could easily be addressed with a couple sentences or a few paragraphs at most. Sorry about that. I'll work on it.