Details:
http://www.w3.org/blog/systeam/2008/02/08/w3c_s_excessive_dtd_traffic
Personally, if I had to serve this traffic, I'd put a highly-concurrent
Web server (e.g., Lighttpd) in place and build a one-second+ delay into
serving those responses. Fine for human use, but will motivate software
authors to find out why things suddenly got so slow...
Beyond that, question: how many IETF specs use a namespace URI as the
primary means of identifying the format/protocol? As opposed to a media
type. My suspicion is that the number is fairly small. If that's the
case, I wonder if there's a need for these URIs to be resolvable.
Cheers,
On 19/02/2008, at 6:01 AM, Julian Reschke wrote:
Lisa Dusseault wrote:
Hi,
I'm adding IANA and Thomas Narten (author of IANA considerations
docs) to this thread to get early IANA feedback about this idea. It
may well be a good idea, but we definitely want to think it through:
- A registry for HTTP URLs in the IANA domain
... including a HTML or XML document hosted by IANA on its Website
at the location specified in the registered URL
... which means that end-users and developers will use these URIs,
perhaps very frequently
What do we know about the W3 experience hosting namespace URIs with
actual documents at the URIs? I've been hearing rumours lately about
something like this causing expensive levels of traffic, but I can't
recall the details.
...
That was about DTDs (fetched by XML parsers), not namespace documents.
BR, Julian
--
Mark Nottingham http://www.mnot.net/