Remove EmbeddedRenderer section.

[dmitrizagidulin]

Removes the EmbeddedRenderer method section, per the group consensus on the previous several calls (due to Phone Home concerns).

---

Preview | Diff

[iherman]

Didn'd we also say that there is no reason to keep the concept of a TemplateRenderer anymore as a result of this change?

[isaackps]

Hi,

I have missed the point of removing this embeddedRenderer, can i find out what is the main reason for doing so?

[msporny]

@isaackps wrote:

I have missed the point of removing this embeddedRenderer, can i find out what is the main reason for doing so?

Hey Isaac, yes, the main reason is that the HTML template renderer provides the same functionality without the phone-home concerns and information leakage concerns. We expect that when we attempt to do the horizontal review at W3C (which we hope to kick off in the next several weeks), that the privacy and security reviews will not allow this mechanism to proceed, thus blocking the entire specification from proceeding.

We can document it as a mechanism that has been deployed, but I think the WG is likely to determine that the approach shouldn't be used since the HTML-based render method is equivalent without the privacy/phone-home concerns. That said, I hope that we have provided an alternative mechanism that your implementations could upgrade to (there is nothing to say that the rendering must happen on the client, but we wanted to make sure it was possible to do so w/o a server and in a completely sandboxed, no networking environment).

Did that answer your question, @isaackps?

[msporny]

@iherman wrote:

Didn'd we also say that there is no reason to keep the concept of a TemplateRenderer anymore as a result of this change?

Hmm, I may be misunderstanding, but no, we still need TemplateRenderMethod for card, nfc, and html.

[iherman]

@iherman wrote:

Didn'd we also say that there is no reason to keep the concept of a TemplateRenderer anymore as a result of this change?

Hmm, I may be misunderstanding, but no, we still need TemplateRenderMethod for card, nfc, and html.

...and I'm not convinced that we need an extra class layer between RenderMethod and the one you mention.

[isaackps]

@isaackps wrote:

I have missed the point of removing this embeddedRenderer, can i find out what is the main reason for doing so?

Hey Isaac, yes, the main reason is that the HTML template renderer provides the same functionality without the phone-home concerns and information leakage concerns. We expect that when we attempt to do the horizontal review at W3C (which we hope to kick off in the next several weeks), that the privacy and security reviews will not allow this mechanism to proceed, thus blocking the entire specification from proceeding.

We can document it as a mechanism that has been deployed, but I think the WG is likely to determine that the approach shouldn't be used since the HTML-based render method is equivalent without the privacy/phone-home concerns. That said, I hope that we have provided an alternative mechanism that your implementations could upgrade to (there is nothing to say that the rendering must happen on the client, but we wanted to make sure it was possible to do so w/o a server and in a completely sandboxed, no networking environment).

Did that answer your question, @isaackps?

@msporny Thanks for the explanation, understand about the phone-home concerns and will look into the equivalent HTML-based render method. 👍

[msporny]

@isaackps wrote:
@msporny Thanks for the explanation, understand about the phone-home concerns and will look into the equivalent HTML-based render method. 👍

Great, thank you @isaackps -- and more specifically, if it doesn't work for your needs, please let the group know as quickly as you can so we can fix it. We are trying to get into feature freeze for this specification within the next several weeks.

[dmitrizagidulin]

Thanks @isaackps! (merging this PR)

@iherman we can discuss the class hierarchy on today's call.