- From: Michael[tm] Smith <mike@w3.org>
- Date: Thu, 5 May 2016 13:56:56 +0900
- To: Adrian Bateman <adrianba@microsoft.com>
- Cc: Zach Koch <zkoch@google.com>, Payments WG <public-payments-wg@w3.org>
- Message-ID: <20160505045656.GI15261@sideshowbarker.net>
Adrian Bateman <adrianba@microsoft.com>, 2016-05-04 15:39 +0000: > Archived-At: <http://www.w3.org/mid/CY1PR03MB14058E6396A639F21647300BD37B0@CY1PR03MB1405.namprd03.prod.outlook.com> ... > · For now, we are only discussing these strings as identifiers. In > the future, though, we will no doubt discuss what resources the URLs > might point to. If we use the relative URL approach that I proposed in > Option 1a then this potentially puts a lot of network load on whoever > hosts the base URL. > > This was a problem in the past when W3C hosted DTDs and XML namespace > schema (in the past, Microsoft’s network was regularly rate limited to > w3.org because of errant software running on machines behind our proxies > that was frequently downloading schema definitions). Yeah, if the WG decision ends up being that we want to use URLs for this, then as a group we are also also going to need to consider what domain names to use for them other than w3.org domains. Because if the group proposes to use w3.org URLs, I will be obligated to personally strongly recommend to the W3C Director to not allow that, and I am confident that the W3C systems team will strongly support that recommendation. See https://www.w3.org/blog/systeam/2008/02/08/w3c_s_excessive_dtd_traffic/ for the reason why. I can imagine that others might suggest ways we could mitigate a similar problem in this case, but I think the response is going to be that the best way to avoid is to not use w3.org URLs for anything like this to begin with. —Mike -- Michael[tm] Smith https://people.w3.org/mike
Received on Thursday, 5 May 2016 04:59:39 UTC