W3C home > Mailing lists > Public > whatwg@whatwg.org > November 2014

Re: [whatwg] New approach to activities/intents

From: Roger Hågensen <rescator@emsai.net>
Date: Mon, 10 Nov 2014 00:30:09 +0100
Message-ID: <545FF901.9080506@emsai.net>
To: whatwg@lists.whatwg.org
On 2014-11-07 20:01, Nils Dagsson Moskopp wrote:
> Roger Hågensen <rescator@emsai.net> writes:
>
>> A link element in the header, maybe call it <link rel="share"
>> href="http://example.com/article/12345/" />
>> or <link rel="share" /> if the current url (or the canonical url link if
>> present) should be used, although I guess in a way rel="share" will
>> probably replace the need to use rel="canonical" in the long run.
> I do not understand. Why should one invent a rel value (“share”) that
> conveys the same semantics as an already existing one (“canonical”) ?
>

Three reasonings:
1. HTTP (301) redirects are advised over rel=canonical, Matt Cutts at 
Google has posted about that in the past as far as I recall. And it 
makes sense as the bots don't need to parse the page to get the 
canonical url.
2. Bookmarking should be of the current page the user has displayed, if 
they bookmark the page and a different url is bookmarked I'd consider 
that undesired behaviour (in the eyes of the user) unless a UI informs 
them or gives them an option.
3. rel=share has already has been "invented" though I'd hardly call 5 
letters an invention.

rel=share also shows clear intent.

A bookmark may be user specific or private to that user.
A canonical (or HTTP 301) indicate to the browser or bot that the page 
is "over there" and not here.
A share is intended to to be, well shared.

It semantically makes sense, at least to me.
rel=bookmark and rel=canonical and a rel=share are all hints.

A search engine for exasmple if it sees a rel=share link that is 
different from say the canonical url (either via HTTP 301 or current 
page or rel=)canonical) should probably ignore it as such a share link 
may have a share tracking url with a reference ID in it.

Also, rel=share is in the wild, I had a url to a list of rel= 
occurrences on the web but ironically I did not bookmark i/note it down. 
While it was low on the list, it was there.

Anyway, this is one place where the rel=share idea is mentioned. 
https://wiki.mozilla.org/Mobile/Archive/Sharing

There is also a rel="share-information" floating around out there, but 
the search engines aren't making it easy for me to search this stuff 
(I'm probably using the wrong syntax/markup). But I found it referenced 
here https://code.google.com/p/huddle-apis/wiki/AuditTrail

There is a rel=share example use on page 5 of 
https://tools.ietf.org/id/draft-jones-appsawg-webfinger-00.txt
Used exactly as how I described it.

Here is a example of rel="share-link" being used 
https://github.com/engineyard/chat/blob/master/views/index.jade

And rel="share" is used in an example here 
https://code.google.com/p/huddle-apis/wiki/Folder#Response
And stated specifically here 
https://code.google.com/p/huddle-apis/wiki/Folder#Sharing_a_folder

As I see it there share is not the same as bookmark or canonical.
There may be some overlap with rel=share and a normal link though (if 
rel=share is used outside the html head).


-- 
Roger "Rescator" Hågensen.
Freelancer - http://www.EmSai.net/
Received on Sunday, 9 November 2014 23:30:38 UTC

This archive was generated by hypermail 2.3.1 : Monday, 13 April 2015 23:09:32 UTC