[POC] Usage of DAM references in word document

[npeltier]

hey!

spinning off here a discussion around an initiative wrt our DAM "relationship". So far we don't have any in the SP / Milo setup, and it may feel great for some, but still bloat docs pretty highly.
Idea many times discussed in different places would be to use image reference that would be hosted "somewhere".

This should considerably shrink doc size, speed up all doc based processes, and open the gate for managing assets in a central place, potentially bringing some resource & operation efficiency.

Pitfalls are already known too:

• editing will be yet less wysiwyg,
• milo would have to be in charge for generating required DOM on its precious time,

but we'd like to measure it with numbers :) hence following ticket
https://jira.corp.adobe.com/browse/MWPW-157395

please let us know what you think here :)

[thedoc31]

This is a good point to bring up. Sharepoint bloat is definitely a concern, especially with duplication for localization. It seems to me like there are a few options we can investigate:

1. Create a new Sharepoint site dedicated to asset management
2. Brand Portal (Gen Studio?)
3. Retain or continue using AEM Assets / Dynamic Media farms
4. Structured S3 bucket

And like you said, any such solution we'd need to test:

1. Page performance penalty
2. Content Security Policy complexity
3. How would it be implemented in Akamai?
4. Can edge workers support this?
5. How would authors / studio get the images into the solution
6. How would authors determine what to put into the word doc / metadata?

[auniverseaway]

Aside from my typical boilerplate of, "DA already solves this entire problem." there's a few other considerations:

1. WebP w/ appropriate fallback formats will also need to be solved for. This will require a server-side component. Unless...
2. We should try to lean on Helix to ingest the image to get same origin for free. I'm suspect this will work if the source image needs to change because they would then be a different hash.
3. If you need to change docs in bulk to point to a different asset, you have not solved the problems inherent with MS Graph and DOCX.
4. Care will need to be taken that 301s are not in the pipeline for rendering. You can get a human readable URL from a Helix ingested image, but it will 301 to the media hash.
5. If alt text is a priority for first pass SEO, this is a non-starter. First-pass SEO does not use JS so all these images will look like links and will not have alt text.

Some other thoughts
GWP have historically used the DAM as a dumping ground for ~90% of their content. Meaning: little to no curation. The particular problem illustrated above represents a small fraction of the content. Take this example DAM search of Shantanu...

There's little curation to some of our most sacred imagery and I do not see this changing any time soon. People want to create and upload what is in front of them without having to muck around in looking for existing assets.

Do I think having a nicely curated and central DAM is a worthy goal? Yes. Do I think it's practical given our history of how we treat content? It's probably aspirational at best.

I do think there's some interesting real-world use cases here, though: FPO (for position only) images pre-event is a good example of one.