- From: David Sheets <sheets@alum.mit.edu>
- Date: Fri, 10 Oct 2014 13:25:09 +0100
- To: Anne van Kesteren <annevk@annevk.nl>
- Cc: Yehuda Katz <wycats@gmail.com>, Sam Ruby <rubys@intertwingly.net>, "www-tag@w3.org" <www-tag@w3.org>
On Fri, Oct 10, 2014 at 11:03 AM, Anne van Kesteren <annevk@annevk.nl> wrote: > On Fri, Oct 10, 2014 at 11:47 AM, David Sheets <sheets@alum.mit.edu> wrote: >> If you insist on calling a set of browsers the web platform, please >> include web servers, deployed software, and development libraries >> which also implement the standards you seek to achieve. >> >> I say this because there are been shockingly little discussion on this >> list regarding convergence of libraries or deployed systems. > > It is exactly the URLs that servers emit, e.g. through Location, that > browsers need to deal with. We also know that curl/wget want to parse > URLs the same way as browsers precisely because of this phenomenon. > Yes, they are all intertwined and the parsing details leak. I have no doubt that software exists which wants to consume broken URLs. In terms of the technical architecture of the Web, what widespread effect do you foresee occurring from this normative expansion of the space of URLs? Do you have plans to relate the New URL to the previous standards? If I have a distributed system which needs to process URLs, should I upgrade all of it to match browsers because browsers tolerate behavior observed in 0.00001% of content? What happens when someone reads about New URLs and then tries to give my system a URL only recently described by the specification?
Received on Friday, 10 October 2014 12:28:14 UTC