Artifact ID: | 7ec069676d243b0aef1fc0dae995dfaf342b30c6 |
---|---|
Page Name: | Experiment: mmmv_silkexec: spec Ideas 01 |
Date: | 2017-05-30 15:26:41 |
Original User: | martin_vahi |
Parent: | 27c71502aebd876e2b331446f1e1401adf8415a0 (diff) |
Next | da56669af3d21cbd1bbf539e49330c76c45fa58c |
This page is for a semi-random collection of ideas, how to improve/change/create the specification of the mmmv_silkexec.
General Observations
Due to the fact that applications work as a combination of application specific code and the dependencies of the application, in some cases some code is loaded dynamically, at runtime, a proper formal verification task gets prohibitively large pretty quickly. Even, if formal verification were possible, the behavior of software packages can depend on hardware, which varies between different users. If hardware is guaranteed to be the same for all users, for example, if all users use a specific version of the Raspberry Pi, then the fact that compile-time parameters depend on the previously installed packages can be countered by installing all packages of the package collection in the exact order that they were compiled and tested at test server, but even then the test server at least partly relies on the tests that have been supplied with the software package, by the software package creator, which may be a different person than the author of the packaged software. It is possible for the software packager to accompany the upstream deliverables with its own, additional, tests. Tests can include partial formal verification.
Hardware Related Observations
Social Aspects Related Observations