Initial claims ontology #804
Conversation
There was a problem hiding this comment.
partiallySupports should be an attribute of PROPERTY_TYPE, not of PROPERTY, since that determination is made for the PROPERTY_TYPE. We need to think of PROPERTY_TYPE as a first class ENTITY, rather that just an "enumeration". We will still have predefined identifiers in SADL for PROPERTY_TYPE for he enumeration.
I'm expecting the PROPERTY_TYPEs to exist in the ontology before any OBJECTIVES are ever added. Would it make sense for OBJECTIVES to point to PROPERTY_TYPE, perhaps? |
|
Eric,
I was just going to modify this comment based upon discussion with Vatsan this morning. The "partiallySupports" attribute has been in the PROPERTY class (and not used) long before the new approach with CLAIMS was proposed. I agree with you it doesn't belong in PROPERTY_TYPE, but neither does it belong in PROPERTY or OBJECTIVE. Our understanding is that it belongs in CLAIM, which can support specific OBJECTIVE(s) based upon the PROPERTY(s) and THEORY the claim uses. E.g., one can write claims to support different objectives using overlapping sets of properties.
Devesh
From: Eric Mertens ***@***.***>
Sent: Friday, October 28, 2022 12:05 PM
To: ge-high-assurance/RACK ***@***.***>
Cc: Bhatt, Devesh (AdvTech) ***@***.***>; Comment ***@***.***>
Subject: [External] Re: [ge-high-assurance/RACK] Initial claims ontology (PR #804)
WARNING: This message has originated from an External Source. This may be a phishing email that can result in unauthorized access to Honeywell systems. Please use proper judgment and caution when opening attachments, clicking links or responding.
partiallySupports should be an attribute of PROPERTY_TYPE, not of PROPERTY, since that determination is made for the PROPERTY_TYPE. We need to think of PROPERTY_TYPE as a first class ENTITY, rather that just an "enumeration". We will still have predefined identifiers in SADL for PROPERTY_TYPE for he enumeration.
I'm expecting the PROPERTY_TYPEs to exist in the ontology before any OBJECTIVES are ever added. Would it make sense for OBJECTIVES to point to PROPERTY_TYPE, perhaps?
-
Reply to this email directly, view it on GitHub<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fge-high-assurance%2FRACK%2Fpull%2F804%23issuecomment-1295241482&data=05%7C01%7Cdevesh.bhatt%40honeywell.com%7C7c5a3099cc244c97cee108dab9068535%7C96ece5269c7d48b08daf8b93c90a5d18%7C0%7C0%7C638025735032017829%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=qa3DtwslqDKXCNa6EzpYJXF72J3S%2F7By9BcxpRCJnr4%3D&reserved=0>, or unsubscribe<https://nam12.safelinks.protection.outlook.com/?url=https%3A%2F%2Fgithub.com%2Fnotifications%2Funsubscribe-auth%2FADJXVB2F2O2R4Z5QWBUI6OLWFQBS7ANCNFSM6AAAAAARFT4VFU&data=05%7C01%7Cdevesh.bhatt%40honeywell.com%7C7c5a3099cc244c97cee108dab9068535%7C96ece5269c7d48b08daf8b93c90a5d18%7C0%7C0%7C638025735032017829%7CUnknown%7CTWFpbGZsb3d8eyJWIjoiMC4wLjAwMDAiLCJQIjoiV2luMzIiLCJBTiI6Ik1haWwiLCJXVCI6Mn0%3D%7C3000%7C%7C%7C&sdata=6o8sCrgRtbFGy1GWaLe95POCFuethcbZwUoZTR314ls%3D&reserved=0>.
You are receiving this because you commented.Message ID: ***@***.******@***.***>>
|
|
All of the claims types have been incorporated into the enumeration variant CSVs. I recommend using the [...] menu to select View File to get a pretty table view in the PR changed files tab. We've instantiated the claims diagram with example bits from each of the contributing teams claims information. Note that the "source" column in these CSV files are simply for reference as we iterate on this design. They are not going to be loaded into RACK itself.
|
|
@bhatt111 we've moved |
|
@bhatt111 I've incorporated your changes to the SRI overlay |
|
I just noticed that with the new claim enumeration types, we now have CWEs added as both PROPERTY_TYPES and MITRE-CWE. Is this really what we want? Should we remove one of them? |
|
@kityansiu I agree that it's not great that the CWEs are duplicated. Maybe another encoding would just be to have a single "CWE present in module" property and require the claim to use "addresses" to point the the actual MITRE_CWE instance. |
|
Extending what Eric said, another way (we are handling an analogous situation in the SRI team) is to have "CWE NOT present in module" as a PROPERTY_TYPE and then create individual instances of PROPERTY that would reference a particular CWE using “mitigates” or similar attribute. |
No description provided.