- From: Kjetil Kjernsmo <kjetil@kjernsmo.net>
- Date: Wed, 28 Mar 2012 15:21:14 +0200
- To: Jonathan A Rees <rees@mumble.net>
- Cc: LOD List <public-lod@w3.org>
On Wednesday 28. March 2012 08.52.12 Jonathan A Rees wrote: > The reason I prepared the survey is because the same ideas keep coming > up over and over again in this context, and, understandably, quite > often people coming into the discussion aren't aware of what has come > before in its ten-year history. I am aware of the critisism that it is hard to deploy. However, what I not aware of is work around *why* typical web hosters find it hard to deploy. Based on my experience, this is mostly because custom code is hard to keep track of, and the security implications of having a lot of custom code are pretty scary to most hosters. In that case, if you want to modify headers and status codes, you need a colo-service, which is expensive and you have to support it yourself, instead of a much cheaper service that is supported by a hosting provider. Thus, what is needed is not to dumb things down, but to find out how the required code goes from being custom to being something hosters are happy to support. I.e., why is phpmyadmin not a custom solution? Howcome memcached can be used by just plugging into existing infrastructure? Why did nginx overtake MS IIS on active sites in just a few years? That's the questions we should be asking, once that is understood, a 303 or a 209 or whatever is a triviality. We need to understand how to get code into hoster's solutions. Personally, I think getting it into Debian is something important *I* can help doing, even if it is rather far from what I should be doing professionally... :-) Kjetil
Received on Wednesday, 28 March 2012 13:22:30 UTC