i think i need to make an interactive interface for (prime) group operations and i think the way you do that in C is to make an executable. that's such a cute idea it'll be like https://rosie-lang.org
3
Replies
0
Boosts
@jamiejennings i have decided to create a library/cli/framework to perform safe modern cryptography because i have waited a few years and nobody else is doing it. there are quite a few canards that academic cryptographers propagate endlessly that i feel you would no doubt be familiar with from the immensely uncreative status quo for text searching and matching tools.
@jamiejennings your work remains incredibly inspiring to me. i have (to my surprise) had reason to return to my theoretical investigations of parsing recently, largely as a result of major practical and theoretical flaws in checksum and encryption processes. in particular, it remains completely unheard of for cryptographic tools to make use of the fact that 99% of their use cases are fixed-length messages in process-local writable memory and not completely unstructured byte streams. it's absurd and unserious not to make use of that structure, particularly when it's not difficult to describe a very clear group-theoretical rationale for doing so (i don't believe in "information theory")
@jamiejennings most notably, when a checksum is used for software release integrity (which is itself applied on top of a separate archiving and/or compression protocol which is opaque and uncontrolled), the mathematical constructs employed are actively optimized to foil any forensic or explanatory capacity, should a file be modified, because checksums are designed for cryptographic signatures, which are architected to avoid leaking information about the message itself. but this obviously demonstrates cryptography's wartime origins continue to influence its modes of thinking--and it also demonstrates how bizarrely detached the filesystem is from the user and application code.
[filesystems are where i hope i can focus my graduate study]. this particular research i've done with robust checksum trees is i think so perfectly up your alley because it is very specifically derived from my domain expertise packaging software at big and small scale and it makes SUCH a great case study of how archival sciences and HCI are in fact a stronger basis than the field of academic cryptography to base robust ontologies around
