- From: Karl Dubost <karl@la-grange.net>
- Date: Sun, 4 Aug 2013 11:02:18 -0400
- To: Liron Shapira <liron@quixey.com>
- Cc: Tim Berners-Lee <timbl@w3.org>, contact <contact@appurl.org>, Larry Masinter <masinter@adobe.com>, Yves Lafon <ylafon@w3.org>, TAG List <www-tag@w3.org>, Leigh Klotz <klotz@quixey.com>
- Message-Id: <BEA4564E-96FB-43AD-9962-8A88208A3F5A@la-grange.net>
Hi Liron, Liron Shapira [2013-08-03T16:30]: > No, we want DNS to take over the role of making up URL schemes for apps. We want apps to use the system that exists today and works. We think the industry needs this as a stopgap measure until http-triggered app launching is universally supported. Hmm… One of the worries I have is that in your documentation you are proposing a Well-Known-Location, see my previous email [1] and you might want to think about /.well-known/ RFC [2] So I went to try and see what was happening so far and took the list documented on Apps directory [3] and I noticed that in fact the schema file was hosted at appurl. Hmmm… So are there using how the doc is recommending it. http://appurl.org/app/ebay.com/appurl.json Let's see. → http GET http://appurl.org/apps/ | egrep -e '<a href="/app/http://(.*)/">([^<].*)</a>' | sed -e 's,<a href="/app/\(.*\)">.*,\1appurl.json,' | sed 's/^ *//g' http://ebay.com/appurl.json http://echofon.com/appurl.json http://evernote.com/appurl.json http://www.flixster.com/appurl.json http://foursquare.com/appurl.json http://grid6.us/appurl.json http://www.imdb.com/appurl.json http://instagram.com/appurl.json http://pandora.com/appurl.json http://www.panogramic.com/appurl.json http://skype.com/appurl.json http://open.spotify.com/appurl.json http://twitter.com/appurl.json http://www.wolframalpha.com/appurl.json http://yelp.com/appurl.json Which gives as a result that only two had a 200 OK. ebay 404 echofon 301 -> 404 evernote 404 flixter 302 -> 404 4sq 301 -> 404 grid6 200 imdb 404 instagram 404 pandora 404 panogramic 200 skype 301 -> 301 -> 404 spotify 404 twitter 301 -> 404 wolfram 404 yelp 404 On those two Grid6 and Panogramic, only one seems to be independent. Grid6 is owned by your company. → whois grid6.us | grep "Registrant Organization" Registrant Organization: Quixey → whois panogramic.com | grep "Registrant Organization" Registrant Organization: Nixter inc So from this list only two 200 OK (with the well-known location issue already mentioned). The rest is not implementing it on their own servers. Why it worries me? Because when you start to deploy things on the Web in public, they become part of the legacy of the Web. There is no eraser on the Web. Once it is deployed, it will stay. Developers will start using http://appurl.org/app/ebay.com/appurl.json instead of http://ebay.com/.well-known/appurl.json It will become a "fait accompli" if Quixey is successful. Suggestions: 1. Do not make appurl.org/app something special with a json file 2. Change the documentation to use /.well-known/ rfc 5785 3. In the case you *really* want appurl.org/app link, make it a permanent redirect to the real one and ONLY for those which exist already. Do NOT create one for those who do not have. [1]: http://lists.w3.org/Archives/Public/www-tag/2013Aug/0031.html [2]: http://tools.ietf.org/html/rfc5785 [3]: http://appurl.org/apps/ -- Karl Dubost http://www.la-grange.net/karl/
Received on Sunday, 4 August 2013 15:02:26 UTC