[To Discuss] Discuss PDF VCs

Question

[To Discuss] Discuss PDF VCs

kimdhamilton opened this issue 4 years ago · comments

Kim Hamilton Duffy commented 4 years ago

Discuss use of PDF as the verifiable credential wrapper, involving PDF with embedded JSON-LD/RDF

Nate Otto · Answer 1 · Mon Oct 19 2020 23:58:47 GMT+0800 (China Standard Time)

Goal: We want something that can be consumed by a machine and by a human.

What is the relationship between the machine-readable data proposed to be embedded in the PDF and the data displayed to humans? There seems to be a strong potential for there to be different information stored in the JSON-LD layer vs in the human-readable layer. What would be the mechanism to prevent the human-readable and machine-readable data from diverging significantly? (because if they diverge, we're not getting the stated benefit of having both human and machine-readable data in one file)
"It is two separate pieces of data, they are just tightly bound to one another"

Brazilian Transcripts use case. There was an XML grammar that they define. "That data becomes displayed as well as machine readable."
- Take a chunk of XML, put it inside a PDF/A-3 format
There was a slide showing CSV embedded in a PDF. How is the example CSV connected to the data shown in the visual display? Is it actually processed in producing the visual presentation of the table in the PDF? It seems no, that there is separate data that is linked from attachments to document sections.
If we try to solve the above problem and get the credential data itself to be visually displayed instead of copying it into a visual display, how do you avoid the necessity of putting extra junk only intended to control the visual display into the credential data?
Does combining PDFs break the signatures?
With PDFs, you can optionally declare a "relationship" string to describe the purpose for which attached data is embedded in the PDF.
You can also define a "schema" for data.
It is possible to search inside attachments.

Kostas Karasavvas · Answer 2 · Wed Oct 21 2020 22:44:36 GMT+0800 (China Standard Time)

I am interested in this as well !

Victoria xDemic · Answer 3 · Tue Nov 10 2020 00:08:20 GMT+0800 (China Standard Time)

Good Morning Kim, What is the meeting ID? The zoom ID on calendar invite doesn’t work. Thanks Sincerely, Victoria Feng xDemic Inc. Founder | CEO Email: victoria@xDemic.com Linkedin: https://www.linkedin.com/in/fengvf/ Web: www.xdemic.com 1250 Broadway | New York | NY 10001

…

On Nov 8, 2020, at 22:24, Kim (Hamilton) Duffy ***@***.***> wrote: Discuss use of PDF as the verifiable credential wrapper, involving PDF with embedded JSON-LD/RDF — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#4>, or unsubscribe <https://github.com/notifications/unsubscribe-auth/APZYBA7S42724XWY4IXMV7TSO5OFFANCNFSM4TOY2SYA>.

Kerri Lemoie, PhD · Answer 4 · Sat Oct 16 2021 02:51:22 GMT+0800 (China Standard Time)

@OR13 do you have suggestions on how to move this topic forward? It doesn't seem specific to EDU and it was mentioned that this topic is of interest to you. Related: #3

Orie Steele · Answer 5 · Tue Oct 19 2021 02:28:30 GMT+0800 (China Standard Time)

I've seen VC embedded in PDFs a view ways.

CBOR-LD / optical encoding via QR code (of content)
optical encoding via QR code (of link to content)
embedding JSON directly
encoding JSON as a table / decoding JSON again.

I would avoid attempting to embed a VC in PDF using PDF internals and I would stick to portable solutions (things that work for images generally).

Kerri Lemoie, PhD · Answer 6 · Tue Dec 07 2021 06:25:39 GMT+0800 (China Standard Time)

Labeling to review at next meeting to gather feedback on OR13's answer.

Kerri Lemoie, PhD · Answer 7 · Thu Feb 17 2022 05:45:04 GMT+0800 (China Standard Time)

The task force suggested that we reach back out to Leonard Rosenthol at Adobe to update us on this topic.

Discussed at these calls: