|
|
Log in / Subscribe / Register

The ABI status of filesystem formats

The ABI status of filesystem formats

Posted Oct 9, 2020 10:33 UTC (Fri) by roc (subscriber, #30627)
In reply to: The ABI status of filesystem formats by bangert
Parent article: The ABI status of filesystem formats

One nice thing about really formal specs (as opposed to careful English specs, which aren't really formal) is that for many specification languages you can find (or write) tools that leverage the specs to do useful work. For example, for a language like Alloy, fuzzing and model-checking tools can automatically generate testcases that are valid according to the spec and exercise all or most of the interesting edge cases (including many that the spec authors didn't think of).

You may also be able to prove useful consistency properties from the spec, e.g. that a "valid" filesystem doesn't have unaccounted-for blocks, etc.


to post comments

The ABI status of filesystem formats

Posted Oct 9, 2020 10:35 UTC (Fri) by roc (subscriber, #30627) [Link]

What I meant to add was: if you write formal specifications the right way, you can use these tools to get more value out of the spec, which tilts the cost/benefit tradeoff towards formal specifications.


Copyright © 2026, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds