Re: Using "Punning" to Answer httpRange-14

[continuation of the discussion that followed after
http://lists.w3.org/Archives/Public/public-rww/2012May/0020.html and
is now no probably more relevant to public-opentag than to public-rww
ML]

On Tue, May 15, 2012 at 3:56 PM, Kingsley Idehen <kidehen@openlinksw.com> wrote:
> On 5/15/12 9:37 AM, Michiel de Jong wrote:
>>>
>>> >  The point isn't about making anything obligatory. The point is about
>>> > leaving
>>> >  the AWWW as it is. It handles the problem well. The fact that people
>>> > get it
>>> >  wrong doesn't mean you break the system.
>>
>> OK, I disagree with that, but at least i understand your standpoint
>> now. Thanks for taking the extra effort of explaining it in some more
>> detail!
>
>
> How can you disagree with a route to the destination you seek?

so to recap, there are 4 types of flavoured links:
- subject-sense, object-sense e.g. a "friendOf" link
- subject-sense, object-content e.g. a "homepage" link
- subject-content, object-sense e.g. an "author" link
- subject-content, object-content e.g. a "summary" link

At current, most people (including most web page authors, including
bloggers, frontend developers, API designers, etcetera) have never
thought about this and include either flavoured or unflavoured links
without thinking about it. Sometimes there will be hashes in the URLs,
sometimes not. Rarely will URLs resolve to a 303. People interpret the
flavour of the link by looking at it, and maybe looking up what the
spec says about it.

Now there are two recommendations that aim to make it easier to
publish and consume flavoured links, namely 303s and hash-uri-rule.
Note that both recommendations would only help to distinguish
'friendOf' links from 'homepage' links and 'author' links from
'summary' links. They are schemes to determine whether a link is
object-sense or object-content. They say nothing about whether they're
subject-sense or subject-content. Also, they only work for links that
are marked up with flavours, not for <span>s and <div>s etcetera.

Anyway, these recommendations exists and both gained some supporters,
so that means we now have three options:

A * stop talking about 303s and the hash-uri-rule, help flavoured
links win bigger acceptation by describing what publisher already do,
and how consumers deal with that, namely punning. Example:
http://support.google.com/webmasters/bin/answer.py?hl=en&answer=1408986

B * recommend that publishers should ideally use the hash-uri-rule
(and maybe 303s) and tell consumers to consider this when interpreting
flavoured links.

C * dictate that publishers should use the hash-uri-rule (and maybe
303s) and tell consumers to consider the rest of the web 'outside the
linked data realm'. Then, we will always know exactly whether a link
is object-sense or object-content, although we would still use punning
to work out whether a link is subject-sense or subject-content.

I myself will choose option A, and i would like to ask the w3c to
choose either option A or option B.

Kingsley replied saying we should choose option C. I told him i
disagreed, and that's where we were. so continuing from there:

> Please
> explain to me the point of disagreement if you don't concur with the
> comments that follow:
>
> You are seeking unobtrusive introduction of new Web realms (systems or
> dimensions) that don't force the hands of Web users. There is nothing about
> the 303 heuristic that forces the hands of existing Web users that are
> minting identifiers for Web Resources ever day.
>
> A problem arises when we conflate the roles and overall profiles of:
>
> 1. Web User
> 2. Web Client Application Developer
> 3. Web Server Application Developer.

a blogger who puts an 'author' link into their blog so that it shows
up nicely in google, is also a web user.

what you're saying is that since the people who develop applications
are smart, it is OK if we force a strict and complicated system on
them. But then you create an "elite web" which only the enlightened
followers of w3c's technical discussions have access to.

I want no elite. I want the web for everybody. Normal people who add
some semantic markup to make their gardening blog link to the Latin
names of plants. And they shouldn't need to hire an elite "application
developer" from some consultancy firm to write a blog. It's just a
blog! :) Simple html, with semantic markup. Writing on the web should
be easy and simple.

So i am asking you to revise that stance: when you publish data with
your high-end professional applications, by all means, be strict.
Always put the # in the right place. But when you consume content that
you retrieve from the World Wild Web, do everyone a favour and be
accommodating when maybe there's a # too many of too few.

You can show a warning, that's fine. Just don't block this
not-totally-w3c-compliant 'web pleb' content altogether.

There is so much wealth of content that you're missing out on
otherwise. All the bloggers who are following Google's instructions
are putting 'author'-flavoured links into their blogs without any
hashes in there. Please don't step on these people with your strict
rules.


Cheers,
Michiel.

Received on Tuesday, 15 May 2012 15:09:57 UTC