- From: Dan Brickley <danbri@google.com>
- Date: Mon, 9 Oct 2017 15:53:04 +0100
- To: Ed Summers <ehs@pobox.com>
- Cc: Hans Polak <info@polak.es>, "schema.org Mailing List" <public-schemaorg@w3.org>
- Message-ID: <CAK-qy=5WWdA4PgKWMU5tfNm3WFvWckyN6=YxhJ7wo_YJG7wR6Q@mail.gmail.com>
On 9 Oct 2017 15:21, "Ed Summers" <ehs@pobox.com> wrote: > On Oct 9, 2017, at 9:21 AM, Dan Brickley <danbri@google.com> wrote: > > To stress Charles' point once more, there are 100s of people on this list. I fear none of them will get much out of speculation on the internal architecture of Google's ( / YouTube's) Web infrastructure, and those of us with some modest knowledge of it can't talk much about it anyway. That said I strongly suspect the particular issue under discussion is simply a bug. Out of scope for this group's real focus. It seems to me that talking about implementations of scheme.org is an important part of talking about schema.org -- if not, I'm not quite sure what I'm doing on this list. There's truth in that for sure, but there has to be a limit. Speculating on internal details of Google's implementation (and general Web infrastructure) is almost always painfully inaccurate, and doesn't help encourage participation from eg Microsoft, Yandex et al here. There are for sure lessons from our implementation at Google that are worth sharing. One is that it is painful that the rdfa/microdata/jsonld trinity don't have out of the box support in many toolkits, esp JS. Also that the interplay of web components with structured data formats seems worthy of more attention. And that remote fetching of json-ld schemas may make it an awkward format for using in IoT due to privacy/security issues. Happy to talk about any of this and more... but can't help with the more speculative "how Google works" stuff for hopefully obvious reasons. Dan //Ed
Received on Monday, 9 October 2017 14:53:34 UTC