If a relationship and the relationship portion of a link is identical, then you don't need different terms -- links are a wrapper around relationships, and you can call them "relationship" and "secure relationship". If relationship isn't a proper subset of link, then the distinction will need to be articulated. I'd assume that the same artifact reference and the same integrity mechanism is needed regardless of whether you put it in an annotation or somewhere else.
"Reference" vs. "contain" doesn't sound like a problem - the hash of an artifact and it's creator/provenance is the same whether it is referenced or copied. I'm not sure about the meaning of "inherit" - I'd assume that it means a "fragment within the current document" - i.e. the artifact specifier is a reference within the parent/container rather than a fully-qualified reference to a different document. Regards, Dave On Wed, Apr 29, 2020 at 4:43 PM Santiago Torres Arias <[email protected]> wrote: > Hi Everyone, > > I finally got myself some time to collect some initial topics regarding > the linking profile for 3.0. I was thinking of adding individual > threads, but I think that for now maybe we want to make sure we are > bringing the priority topics forward. Here are the things I was thinking > of discussing: > > 1. The name. I recall integration being suggested by Nisha, and having > some pushback on linking (as it is reminiscent of http hyperlinking). > I'm not too against integration profile, I like "artifact flow" but > that's a mouthful. Ideas welcome! > > 2. The existing overlap with the current relationship property. It feels > they fill similar roles, yet there are nuances. Here's what I could > gather: > > - Relationships are intended to document *how* artifacts connect to each > other. I may be missing a little context here so I'd appreciate to > have more ideas around this. > > - ~links~ are intended to provide artifact flow integrity (i.e., to > ensure these relationships are integral). They also focus on > authenticating creators and consumers > > 3. If this overlap is taken care of this properly means: > - not-links could be "secure relationships", which opens the can of > worm of inheritance vs reference vs contains > > - not-links could be a completely different animal (e.g., an > annotation) and as such we can take the conversation in another > direction. > > I think these are the overarching discussions to have. I believe that > probably figuring out 2 will set the tone for 3, so we can probably > start there :) > > I hope this helps stirring discussion! > > Cheers! > -Santiago > > > > -=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#3882): https://lists.spdx.org/g/Spdx-tech/message/3882 Mute This Topic: https://lists.spdx.org/mt/73362126/21656 Group Owner: [email protected] Unsubscribe: https://lists.spdx.org/g/Spdx-tech/unsub [[email protected]] -=-=-=-=-=-=-=-=-=-=-=-
