The reason the artifact âtypeâ keyword is optional in 1.3 is not to support refinement, but rather because the 1.3 spec assumes that based on the file extension or the mime type of the artifact file, the TOSCA Artifact Type can be inferred
by checking the collection of Artifact Types in the Simple Profile and selecting the one that supports that file extension and/or mime type. This is spelled out in Section 3.6.7.2.1 of the Simple Profile v1.3 spec.
Since we no longer include a standard profile in Version 2.0, this approach is no longer workable and the Version 2.0 spec must be updated to reflect that type is a mandatory keyword for artifact definitions.
Chris
From: Tal Liron <tliron@redhat.com>
Sent: Thursday, February 3, 2022 11:32 AM
To: Chris Lauwers <lauwers@ubicity.com>
Cc: tosca@lists.oasis-open.org
Subject: Re: [tosca] "type" keywords should be conditional, not mandatory
The artifact definition supports a short notation, by which
only the "file" keyword is provided. This is clearly a situation in which we assume "type" is conditional and thus can be calculated. So, are we assuming a refinement is possible only in node types, by which we inherit the artifact definition from the parent
node type? And we are assuming that in the case of node templates it's always a replacement, meaning that this short notation can't be used there?
See, I think it's not so trivial and worth revisiting.
You are correct about the type keyword being optional in refinements. In fact, in refinements *ALL* keywords must be optional, under the assumption that refinements refine already valid definitions, which already have all the mandatory
keywords. The Version 2.0 spec states this in Section 5.1, but we could probably improve on the language to make this even more clear.
Let's use the word "conditional" rather than "optional" (as we've been using until now) because "optional" implies it's up to the user, which is not true. In some cases the keyword is absolutely mandatory.
That said, you chose the one example where this rule doesnât hold, since artifacts are treated differently from most other entities in TOSCA. Artifacts donât have definitions and
associated assignments. Instead, they are just directly associated with (or attached to) either a node type or on a node template. If a derived type defines an artifact with the same name as an artifact on a base type, then that artifact just replaces (or
overrides) the artifact in the base type. It doesnât refine it, it just redefines it. This is spelled out in Section 5.3.7.2.3 of the Version 2 specification.
I remember us discussing this but I don't remember us finalizing it, exactly because of this compilation. :) How would you know if the intent is to refine or not? If the designer is using the same artifact name as in node type, doesn't
it mean that they intend to refine it? I agree it's a complex case, but the discussion of the conditionality of "type" might be a great way to tease out the complications.
I would like us to put it on the TODO list to be revisited. (Or maybe we'll include it in the discussion about what to do with artifacts generally.)
|