W3C home > Mailing lists > Public > public-schemaorg@w3.org > April 2017

Re: Version Control

From: Timothy Holborn <timothy.holborn@gmail.com>
Date: Tue, 11 Apr 2017 18:24:32 +0000
Message-ID: <CAM1Sok2R7MEJXUEY5Uu8AEqurtkXg8aCsNz=h3xrt16bMOBOkg@mail.gmail.com>
To: Dan Brickley <danbri@google.com>, "schema.org Mailing List" <public-schemaorg@w3.org>
On Wed, 12 Apr 2017 at 03:05 Dan Brickley <danbri@google.com> wrote:

> On 11 April 2017 at 18:46, Timothy Holborn <timothy.holborn@gmail.com>
> wrote:
> > How is version control handled?
>
> We have numbered releases (with occasional minor hot fixes in between)
> - see http://schema.org/docs/releases.html. For fine-grained detail
> there are github commit IDs. If you want a document to indicate a
> dated release that it uses, you can use
> http://schema.org/schemaVersion . At Github we could (and will) make
> better use of tags for releases etc.
>

Sounds like a plan.

Re: VCWG work (verifiable claims) it seems kinda important the
intepretation is retained; or that the signed-instrument be able to be
updated?

Perhaps via: https://www.w3.org/TR/ldn/ ??

dependency graph;
1. gen signed Verifiable Claim ("VC") (incorporating RDF + Version ID)  -->
What's the relationship thereafter with LDN/LDP?

2. definition is updated, mechanism (ie: LDN) used to provide the option to
update VC with new definition? --> interesting workflow considerations...


> Dan
>
> p.s. Regarding signing, in the foaf project ~15 years ago


yup. guru. i know. ;)


> (and
> somewhere in http://lists.foaf-project.org/pipermail/foaf-dev/ I
> guess) we used to carefully GPG sign the schema/spec with each
> release, but it never really found much use so we stopped doing it.
> Even when you fix some things rigidly with crypto, there are always
> squishy human aspects where you'll find semantic drift e.g. reference
> by description ("President Bush") or definitions whose community
> understanding shifts even if the formal definition text doesn't (e.g.
> schoolHomepage, which evokes a slightly different idea amongst EN-US
> vs EN-UK speakers).
>

yup.  working on a tool concept around 'RDF sentences'
https://lists.w3.org/Archives/Public/public-schema-gen/2017Mar/0001.html  |
but not much to report yet (just finished www2017 event work); still early
stage concept.  "foaf" context concepts are kinda temporal imho, or
'squishy', as you put it.  some consideration was made about the potential
to cg the concepts therein; somewhat moreover relating to owl / sparql
scripts, perhaps packaged in ldp containers?

version control (other than using github) seems like an emerging issue.
Was thinking maybe it's a problem for the LDP group, do you have any
thoughts on the matter?

packaging the definitions may bloat 'blockchain' methods, yet it's kinda
important the context is retained within a 'trust' instrument (as to
maintain the means for intended interpretations).  Herein/hereafter,
decentralised discovery considerations.  I'll be constructing a document
with snippets from various docs produced over time to a more concise
overview of thoughts in the area, taking into consideration some of the
other lifecycle dependencies in a manner i'm hoping will lead to
progressive outcomes...

The RDF sentences concept - seems like bit of an (interesting) mission.

Tim.H.

>
>
> > Within the Verifiable Claims works[1] ontology is used in a verifiable
> > claim, which is then signed[2]; yet the question becomes, what happens if
> > the ontological definition changes?  in theory, this may change the
> concept
> > signed as a 'claim' by changing the definition of the URIs involved.?
> >
> > CONCEPT:  Add version control somehow.
> >
> > not sure how.
> >
> > Tim.H.
> >
> > [1]
> >
> https://docs.google.com/presentation/d/1pFGC1G7CbizUuvbmjECfnNRL4fZk9QLxG8d3nehgwNU/edit#slide=id.p
> > [2] http://json-ld.org/playground/
>
>
Received on Tuesday, 11 April 2017 18:25:17 UTC

This archive was generated by hypermail 2.4.0 : Friday, 17 January 2020 17:12:35 UTC