| SUBMITTED: Stardate 96021.7 | FILED BY: LCDR H'ajah | SECURITY: Level 1 - Open |
SUMMARY: THE ENGINEERING DEPARTMENT HAS COMPLETED ITS CODE REVIEW OF THE DECOMPILED ORIGAMI BOT OPERATING SYSTEM.
KEY PERSONNEL:
LCDR H'ajah
LCDR Ophalesh (NPC)
LCDR Hallen, Dale (NPC)
LT Yohamdi (NPC)
LCDR Ophalesh (NPC)
LCDR Hallen, Dale (NPC)
LT Yohamdi (NPC)
NARRATIVE: COLLEAGUES, JOIN ME IN SINGING THE PRAISES OF THE COMPUTER SERVICES DIVISION, WHICH HAS COMPLETED ITS CODE REVIEW OF 30 MILLION LINES OF CODE DECOMPILED FROM THE ORIGAMI BOT INSTALL IMAGE CAPTURED LAST YEAR. THEIR FULL REPORT IS ATTACHED BELOW; I WILL SHARE ONLY THE MOST PERTINENT HIGHLIGHTS.
Hallen wrote:
The scouting mode, which appears to be the default setting, loops through the following conditional behaviors:
Ophalesh wrote:
The data structure is really interesting: highly compact and efficient, yet still extremely expressive considering the small byte size. Cracking it took the longest time, but once we had the UT trained properly we were blown away by how much instruction they managed to fit on such a tiny plate.
The problem, from our perspective, is that it's exquisitely bespoke for the hardware it runs on. To make any use of it ourselves, we'd basically have to adapt to their system architecture. Even then, it would take a ton of work to generalize the OS beyond the functionality of these oribots. This whole system was designed in a wind tunnel to do one thing very well and one thing only.
The problem, from our perspective, is that it's exquisitely bespoke for the hardware it runs on. To make any use of it ourselves, we'd basically have to adapt to their system architecture. Even then, it would take a ton of work to generalize the OS beyond the functionality of these oribots. This whole system was designed in a wind tunnel to do one thing very well and one thing only.
Hallen wrote:
Even after 10,000 hours of simulated stimulus, there are still large sections of the codebase that we've never seen active. Based on our review of the dark sector code, these seem to concern fabrication, harvesting, security/defense, and other protocols we've never been able to trigger. So far our behavior flowchart has close to half a million nodes and we're still far from done mapping.
Yohamdi wrote:
There's still about 6% of this code that the UT can't make heads or tails of. We're trying every heuristic in the book but without more context, these symbols are just not talking to us. They aren't consistent with the symbolic structures we see in the sections we've translated already, making it tough to extrapolate from one to the other. Basically we're stumped on what the hell they do.
Ophalesh wrote:
One thing that we're still working on is the memory storage format. We haven't yet cracked this encryption algorithm, but we can tell that the sectors devoted to memory are surprisingly small. That might be because the storage format is incredibly dense--we've had plenty of surprises so far, so we can't rule that out--or it might be because the little guys just don't retain much information.
RECOMMENDATION: WE WILL CONTINUE OUR STUDY OF THE OPERATING SYSTEM CODEBASE, WITH A PARTICULAR FOCUS GIVEN TO BREAKING THE MEMORY ENCODING. MEANWHILE, WE WOULD BE ABLE TO BETTER VERIFY OUR SIMULATION VALIDITY IF COMMAND AUTHORIZES THE FABRICATION AND IMPRINTING OF MORE ORIGAMI BOTS. A SMALL BATCH OF ONE THOUSAND WOULD SUFFICE FOR NOW.
//ATTACHMENT// ds13_eng_oribot_cr.rpt
Thanks to Katriel for providing the information that served as the basis for this report.