search

w3c / wpub

W3C Web Publications
https://w3c.github.io/wpub/
Other
78 stars 19 forks source link

Added a text whereby using the embedded manifest is RECOMMENDED. #412

Closed iherman closed 5 years ago

iherman commented 5 years ago

This change is per https://www.w3.org/publishing/groups/publ-wg/Meetings/Minutes/2019/2019-03-18-pwg#resolution2

Fix #327

Preview | Diff

mattgarrish commented 5 years ago

Generally looks fine to me. I just made a couple of minor editorial tweaks to put the recommended practice first (emphasize what we want people to do) and to shorten the note.

iherman commented 5 years ago

Embedding is the preferred option as search engines will only process schema.org metadata in JSON-LD format when it is embedded in an HTML page.

May not be o.k. This makes a general statement on all search engines, which we cannot be sure is the case. Well, we know that this is what Google does (ie, does not), but on long term we cannot be sure and maybe, say, Yandex is more flexible.

What about a will->may, like:

Embedding is the preferred option as search engines may only process schema.org metadata in JSON-LD format when it is embedded in an HTML page.

mattgarrish commented 5 years ago

I actually took "may" out to avoid an rfc keyword in a note. :)

What about "typically" or "usually" or something along those lines rather than "may"?

iherman commented 5 years ago

Is it a problem with 'may', as opposed to 'MAY'? I do not think so... let alone the fact that this is a note, ie, informal anyway...

mattgarrish commented 5 years ago

Well, yes, we should avoid keywords in informative content. Relying on uppercase to determine what is a keyword isn't terribly accessible.

It also adds confusion about whether we intended to make a normative statement, IMO, even if in this case it would be outside the scope of the spec.

In these cases, "might" can usually be dropped in instead with no impact on the meaning.

iherman commented 5 years ago

So... can we use

Embedding is the preferred option as search engines might only process schema.org metadata in JSON-LD format when it is embedded in an HTML page.

I made a new commit accordingly.