- From: Tobie Langel <tobie@w3.org>
- Date: Tue, 25 Jun 2013 12:37:45 +0200
- To: Robin Berjon <robin@w3.org>
- Cc: spec-prod@w3.org, Marcos Caceres <w3c@marcosc.com>
On Tuesday, June 25, 2013 at 12:12 PM, Robin Berjon wrote: > On 25/06/2013 12:09 , Tobie Langel wrote: > > On Tuesday, June 25, 2013 at 11:58 AM, Robin Berjon wrote: > > > The only thing that I care about from the ReSpec point of view is > > > performance. If you're only adding a few fields, it probably > > > doesn't matter, but if you're adding enough that the output might > > > start being noticeably slower to load and parse, then I'd rather > > > the API that ReSpec uses stayed lean and mean. > > > > > > Most requests are clocking well above a single packet, so I don't > > think even doubling the size of the output would have a significant > > perf impact. > > > That WFM. Incidentally, I wonder if gzip might help here. I should take > the time to look at what kind of performance impact the current usage > has to see if it's worth optimising or not (I suspect not, but where > performance is concerned gut feelings are almost always wrong :). Shipped: https://github.com/tobie/specref/commit/50d39cc7 --tobie
Received on Tuesday, 25 June 2013 10:37:54 UTC