Consider updating Dataset guidance to mention Google's description field length limit

[amoeba]

At DataONE, we insert schema.org-/SOSO-compliant JSON-LD into all of our dataset landing pages. The value we set for the description field comes directly from the original metadata record and is used unmodified. We use Google's Search Console to monitor how our Dataset records are being indexed and recently noticed a few records which failed validation with an error message of:

Invalid string length in field "description"

It turned out that these records had unusually long descriptions (e.g., ~6000 characters) and Google's guidelines have something to say about the length of this field:

The summary must be between 50 and 5000 characters long.

We haven't validated that this is the cause of our validation problem above but we thought maybe folks here would be interested in including a note about Google's guidance above. We'll be testing a fix soon and will report back here with the result.

Has anyone run into this? Is it worth putting into the SOSO Dataset guidance doc? Is this limit (either the low end, 50, or the high end, 5000) problematic for SOSO's use cases and, if so, would it be worth bringing this up with folks at Google?

[smrgeoinfo]

I think it would be useful to mention these constraints. I don't think the string length limitation should be a problem. If you need more than 5000 characters, we might have a convention to add additional information in the Dataset/abstract... Using any other of the many possible Text fields on CreativeWork would create problems for clients harvesting and indexing the metadata.

[mbjones]

+1 on adding a statement to the guidance doc -- it might trip quite a few people up. About 50% of the DataONE metadata records that we get directly from providers fail this check.