DAG of documents

[quicwg]

DAG of documents

[martinthomson]

I considered doing this, but I think that the loss of utility is pretty acute. The solution I have in place will work for future submissions reasonably well (modulo some tweaking).

[MikeBishop]

Do you mean the loss of ease for submission, or that the documents themselves would suffer? Arguably, if we want these to be modular, we should be defining more than one of each pluggable "thing" to help ensure they are. But since we aren't doing that, forcing ourselves to define the interfaces (or for now, setting out space for the interfaces) seems like the next best way of ensuring we achieve that goal.

Of course, it might force some delay between version submissions while we wait for the new XML to publish; I agree with you on the ease-of-submission issue.

[MikeBishop]

Tagging this as "needs discussion," not for the editorial aspect of it, but because this reflects the "Does Transport depend on TLS?" discussion. If Transport depends on TLS, we can't have a DAG.

[martinthomson]

Clearing the label. I think that the transport depends on TLS, but we're trying to keep a clear delineation between the two.

[martinthomson]

@MikeBishop, I'd really like to close this with no action. Do you want to make a case for doing this?

[martinthomson]

In the absence of a response. I'm going to make an executive decision here.