[bitcoin-dev] Formal specification of Miniscript in Alloy

[bitcoin-dev] Formal specification of Miniscript in Alloy

[Dmitry Petukhov]

I have created a formal specification of Miniscript [1] using
the specification language of Alloy analyzer [2]

Link: https://github.com/dgpv/miniscript-alloy-spec

Possible uses for the spec:

- Implementing Miniscript libraries, as additional reference that might
  be easier to navigate than prose spec

- Generating test cases for implementations, although currently this
  will be a manual process due to the tools limitation (can be overcome
  with GUI automation)

- Checking the implementation against the spec, by writing a program
  that would generate Alloy .als files from the data structures of the
  implementation, and then checking these files in Alloy

- Extending or amending Miniscript, if the need arise. Having
  extenstions and changes checked (with bounds) against a spec should
  help catch inconsistencies

- Exploring the properties of Miniscript

If you have an interest in Miniscript, please consider looking at the
spec and share your ideas.

The spec may contain mistakes, as it was not yet checked against any
implementation, it was only checked for consistency using its own
predicates, with the scope of up to 8 nodes.

If you notice a mistake or inconsistency, please submit an issue on
github (or communicate this in other ways)

[1] http://bitcoin.sipa.be/miniscript/
[2] https://alloytools.org/

Re: [bitcoin-dev] Formal specification of Miniscript in Alloy

[Clark Moody]

[-- Attachment #1: Type: text/plain, Size: 1911 bytes --]

Thanks for this! I can't comment on the correctness of your implementation,
but I really appreciate the idea and effort.

By chance, did you come across any other spec definitions in alternate
formal grammars?


-Clark


On Wed, Nov 25, 2020 at 5:35 AM Dmitry Petukhov via bitcoin-dev <
bitcoin-dev@lists.linuxfoundation.org> wrote:

> I have created a formal specification of Miniscript [1] using
> the specification language of Alloy analyzer [2]
>
> Link: https://github.com/dgpv/miniscript-alloy-spec
>
> Possible uses for the spec:
>
> - Implementing Miniscript libraries, as additional reference that might
>   be easier to navigate than prose spec
>
> - Generating test cases for implementations, although currently this
>   will be a manual process due to the tools limitation (can be overcome
>   with GUI automation)
>
> - Checking the implementation against the spec, by writing a program
>   that would generate Alloy .als files from the data structures of the
>   implementation, and then checking these files in Alloy
>
> - Extending or amending Miniscript, if the need arise. Having
>   extenstions and changes checked (with bounds) against a spec should
>   help catch inconsistencies
>
> - Exploring the properties of Miniscript
>
> If you have an interest in Miniscript, please consider looking at the
> spec and share your ideas.
>
> The spec may contain mistakes, as it was not yet checked against any
> implementation, it was only checked for consistency using its own
> predicates, with the scope of up to 8 nodes.
>
> If you notice a mistake or inconsistency, please submit an issue on
> github (or communicate this in other ways)
>
> [1] http://bitcoin.sipa.be/miniscript/
> [2] https://alloytools.org/
> _______________________________________________
> bitcoin-dev mailing list
> bitcoin-dev@lists.linuxfoundation.org
> https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
>

[-- Attachment #2: Type: text/html, Size: 3244 bytes --]

Re: [bitcoin-dev] Formal specification of Miniscript in Alloy

[Dmitry Petukhov]

В Wed, 25 Nov 2020 09:43:10 -0600
Clark Moody <clark@clarkmoody.com> wrote:

> Thanks for this! I can't comment on the correctness of your
> implementation, but I really appreciate the idea and effort.
> 
> By chance, did you come across any other spec definitions in alternate
> formal grammars?

As far as I know, this is the first formal spec for Miniscript.

I also considered using the K framework [1] for the task (it would
also give the parser out of the box), but at the time it did not have
the reference documentation, only tutorials on their site. It seems that
they have published the new site, with more documentation.

I am satisfied with how Alloy spec turned out, though - in my opinion,
the node definitions in the spec are very readable.

[1] https://kframework.org/


> On Wed, Nov 25, 2020 at 5:35 AM Dmitry Petukhov via bitcoin-dev <
> bitcoin-dev@lists.linuxfoundation.org> wrote:
> 
> > I have created a formal specification of Miniscript [1] using
> > the specification language of Alloy analyzer [2]
> >
> > Link: https://github.com/dgpv/miniscript-alloy-spec
> >
> > Possible uses for the spec:
> >
> > - Implementing Miniscript libraries, as additional reference that
> > might be easier to navigate than prose spec
> >
> > - Generating test cases for implementations, although currently this
> >   will be a manual process due to the tools limitation (can be
> > overcome with GUI automation)
> >
> > - Checking the implementation against the spec, by writing a program
> >   that would generate Alloy .als files from the data structures of
> > the implementation, and then checking these files in Alloy
> >
> > - Extending or amending Miniscript, if the need arise. Having
> >   extenstions and changes checked (with bounds) against a spec
> > should help catch inconsistencies
> >
> > - Exploring the properties of Miniscript
> >
> > If you have an interest in Miniscript, please consider looking at
> > the spec and share your ideas.
> >
> > The spec may contain mistakes, as it was not yet checked against any
> > implementation, it was only checked for consistency using its own
> > predicates, with the scope of up to 8 nodes.
> >
> > If you notice a mistake or inconsistency, please submit an issue on
> > github (or communicate this in other ways)
> >
> > [1] http://bitcoin.sipa.be/miniscript/
> > [2] https://alloytools.org/
> > _______________________________________________
> > bitcoin-dev mailing list
> > bitcoin-dev@lists.linuxfoundation.org
> > https://lists.linuxfoundation.org/mailman/listinfo/bitcoin-dev
> >