gmane.education.gcs-pcs

Re: Re: what happened to Wikipedia COinS?

Andy Mabbett — 2012-11-15T22:12:14

I'm not a proponent of the sudden removal of COinS, which I opposed (I
had a hand in introducing COinS to many templates on Wikipedia)

I am a proponent of a citation microformat, which could sit alongside
COinS and, perhaps, /eventually/ replace it.

If you could put your concerns online, perhaps in a blog post, I'll
bring it to the attention of those concerned and see if we can't get
the removal reversed; or accelerate the microformat proposal.

A.

On 15 November 2012 21:53, Godmar Back <godmar-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

Re: Re: what happened to Wikipedia COinS?

Godmar Back — 2012-11-15T22:01:29

PS: I just remembered that even though we haven't officially upgraded to
LibX 2.0, we're already collecting statistics via Google analytics. From
August 1st to Nov 14, LibX users have clicked on a hyperlinked COinS 5,461
times. The vast majority of which was, it stands to reason, from Wikipedia
(though we don't track the page a user is on when they click on a COinS).
And that's for only 10 editions who have a significant LibX 2.0 user base -
once we flip the "upgrade everyone" switch, it's likely to increase by an
order of magnitude.

So, there's real usefulness even while it's preliminary.

 - Godmar



On Thu, Nov 15, 2012 at 4:53 PM, Godmar Back <godmar-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

Re: Re: what happened to Wikipedia COinS?

Godmar Back — 2012-11-15T21:53:36

Andy,

you seem to be a proponent - it now seems they (who, really?) has thrown
out the child with the bathwater. Why didn't they at least AJAX the data in
on demand (if page load times were the issue?)

Really unfortunate, especially since they (Wikipedia) didn't even know how
widely their COinS were used.  Certainly, they added real value to users of
LibX, even with their current (limited) implementation. Many newspaper
articles that were inaccessible with the direct link on the Wikipedia page
became accessible just via the OpenURL resolver.

The big irony is this: I was just able to recruit a student to refine COinS
support in LibX using the now available Summon API/Widget service to
direct-link to a resource. We're having our 2nd meeting, and wanted to
start with an analysis of the real-world COinS quality provided by
Wikipedia to see which techniques to use to provide the user with direct
access. Then they're gone.

I'm not sure if we'll invest in 2001-style scraping of metadata without any
easily disce

Re: Re: what happened to Wikipedia COinS?

Andy Mabbett — 2012-11-15T20:46:34

Yes; what more would you like to know?

See also:

http://en.wikipedia.org/wiki/Wikipedia_talk:WikiProject_Microformats#Proposal:_citation_microformat


On 15 November 2012 20:37, Godmar Back <godmar-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

Re: what happened to Wikipedia COinS?

Godmar Back — 2012-11-15T20:37:59

This happened:
http://en.wikipedia.org/wiki/Template_talk:Citation/core#Removing_COINS_metadata

They removed it because their implementation was too slow :-(

It doesn't appear that they included any of the microformats, either.

Does anyone have any perspective/insider knowledge on this?


On Thu, Nov 15, 2012 at 2:35 PM, Godmar Back <godmar-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

what happened to Wikipedia COinS?

Godmar Back — 2012-11-15T19:35:00

Hi,

does anyone know what happened to Wikipedia's COinS?

They seem to have been removed (!?)

 - Godmar

COinS implementors. Please tell us if you want your COinS rendered!

Godmar Back — 2012-08-09T15:58:52

LibX renders COinS, typically as icons that provide links to OpenURL
resolvers. We're currently making them a lot smarter by using services
such Summon to tell an interacting user more about the availability of
the referred-to item.

We are regularly getting requests from LibX maintainers and users to
suppress COinS on certain pages where the COinS is placed in a way
that disrupts the layout of the page.

vufind is a big offender.

The most prominent COinS processors are probably Zotero & LibX. Zotero
doesn't render them and thus doesn't care where they are. We do.

The spec says:

"Since an important use of this metadata will be to allow processing
agents to make OpenURL hyperlinks for users in libraries (latent
OpenURL), the method must allow the metadata to be placed any where in
HTML that a link might appear."

Proposal:

As a COinS provider, please provide us with guidance if you want your
COinS rendered or not.

If you don't, simply place it in an invisible container or add
style="display: none" and we

Re: Revisiting the choice of for unAPI?

schuyler — 2011-04-07T15:12:28

I don't have experience myself, but WAIARIA seems to be pretty widely
supported (at least partially):
http://www.iheni.com/screen-reader-testing/

Though as I'm reading more, it seems most modern screen readers ignore
things that are display:none anyway, so if anything, it's overkill.

Re: Re: Revisiting the choice of for unAPI?

Galen Charlton — 2011-04-07T13:26:48

Hi,

On Wed, Apr 6, 2011 at 3:42 PM, schuyler <schuyler1d-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

Interesting, and would certainly be trivial to implement.  Since
WAI-ARIA only recently became a candidate recommendation, do you know
if enough assistive technology recognizes aria-hidden to make a
practical difference at the moment?

Regards,

Galen

Re: Revisiting the choice of for unAPI?

Galen Charlton — 2011-04-05T18:36:16

Hi,

On Tue, Apr 5, 2011 at 2:26 PM, Mike Rylander <mrylander-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

It looks like that the value-title variant [1] of the value-class
pattern would give us a way to avoid the accessibility problem while
not inflicting opaque strings meant for machine processing on humans.

[1] http://microformats.org/wiki/value-class-pattern#Parsing_value_from_a_title_attribute

Regards,

Galen

Re: Revisiting the choice of for unAPI?

Galen Charlton — 2011-04-05T20:21:40

Hi,

On Tue, Apr 5, 2011 at 3:49 PM, Mike Rylander <mrylander-Re5JQEeQqe8AvxtiuMwx3w< at >public.gmane.org> wrote:

The difference is that <abbr> has particular semantics that are used
by screen readers; in particular, the contents of the title attribute
are read out by screen readers.  The value-title pattern uses empty
spans that are, according to the link, are hidden both from visual
display and screen readers.

Regards,

Galen

Re: Revisiting the choice of for unAPI?

schuyler — 2011-04-06T19:42:34

MediaThread (http://ccnmtl.columbia.edu/mediathread) searches for
unAPI information in its bookmarklet, parsing it on sites like WGBH's
OpenVault (e.g. http://openvault.wgbh.org/catalog/org.wgbh.mla:MLA001106).

The more ways we have to search for an unAPI id, the more burdensome
the spec becomes.  The ABBR element is nice because it's rare on a
page (unlike SPANs).

First, to solve the accessibility issue clinically, perhaps we can
just recommend adding attribute < at >aria-hidden="true"
http://www.w3.org/TR/wai-aria/states_and_properties#aria-hidden

<abbr class="unapi-id" aria-hidden="true" title="abc-news-videosource:
9468a"></abbr>

If we're going to add new ways to reference unAPI IDs in HTML, I'd
suggest linking it up with the HTML5 MicroData work which solves any
accessibility issues and will have javascript apis and new tooling
connected to it.
http://www.whatwg.org/specs/web-apps/current-work/multipage/links.html#microdata

The new way would be something like:

<span itemscope itemprop="unapi-id" itemty

Re: Revisiting the choice of for unAPI?

Mike Rylander — 2011-04-05T21:17:58

That's the point I missed.  Thanks.  Supporting both, for backward
compat, would not bother me at all.  I suspect that many unAPI
consumers actually just look for any element with a class of
'unapi-id' anyway.

Re: Revisiting the choice of for unAPI?

Mike Rylander — 2011-04-05T19:49:43

But, isn't that exactly what unAPI specifies today in an <abbr>?

Re: Revisiting the choice of for unAPI?

Mike Rylander — 2011-04-05T18:26:37

Evergreen (http://evergreen-ils.org/) uses unAPI extensively, so we
wouldn't want to see the abbr-title pattern removed, but I don't see
any reason to avoid aligning the spec with the value-class pattern a
bit in terms of supported implementation.

Note, though, that the value for unAPI is a tag: URI, not human-usable
data, so it may be less applicable to unAPI than, say, hCard.  My
$0.02 ...

Re: Revisiting the choice of for unAPI?

Jonathan M. Lane — 2011-04-05T17:46:32

Hi Dan,

On Mon, Apr 4, 2011 at 01:29, Dan Chudnov <dchud-63aXycvo3TyHXe+LvDLADg< at >public.gmane.org> wrote:

Exactly right.


I am not using it myself, but I work closely with the technical
library at my institutional library, who are actively maintaining some
unAPI access points on some of their services and systems.

While unAPI may be past it's prime (I am not certain if unAPI has been
replaced by something newer/more appealing), updating the standard to
recommend the use of a more accessibility-friendly and semantically
correct microformat would be beneficial for anyone still using unAPI
or those considering implementing it.

Any reasons why it may not be worth updating the standard to use the
value-class-pattern microformat?

Jonathan

Re: Revisiting the choice of for unAPI?

Dan Chudnov — 2011-04-04T04:29:00


Thanks for bringing this up.  Sorry for the delayed response.

I haven't been tuned in to microformats in a while.  Is this the pattern you're referring to?

  http://microformats.org/wiki/value-class-pattern

I'm not sure there's a lot to gain from updating unAPI at this point.  Are you using it?

  -Dan

Re: Revisiting the choice of for unAPI?

Jonathan M. Lane — 2011-03-21T16:38:59

I think it might be time to revise unAPI to encourage (preferably mandate, 
but that's probably unrealistic) the use of the value class pattern<http://microformats.org/wiki/value-class-pattern>, 
which is a proven microformat pattern that does not have the same accessibility 
issues<http://microformats.org/wiki/accessibility-issues#abbr-design-pattern>as the ABBR 
design pattern <http://microformats.org/wiki/abbr-design-pattern>.

Re: unapi example?

Dan Chudnov — 2010-04-21T14:30:25


You're right, the site no longer works that way.  I disabled the wordpress backend a while ago to make the site static and should have thought more about the impact of breaking those links.

I see that it's confusing to have these broken links in the spec, but it isn't clear to me whether the spec is, er, "prominent"? "successful"? "important"? (etc.) enough to warrant keeping a working and up-to-date php app running on the backend.

What do you think?  -Dan

unapi example?

Jodi Schneider — 2010-04-19T17:16:49

Hi,

I'm trying to understand the unAPI spec:
http://unapi.info/specs/

The example (under 'complete example') at that page says:
"Compose the unAPI base URL, weblog entry identifier, and the oai_dc
FORMAT by visiting http://unapi.info/news/unapi.php?id=http://unapi.info/news/archives/9&format=oai_dc
to see a Dublin Core record for the entry in the OAI-PMH schema for
Dublin Core data."

However, http://unapi.info/news/unapi.php gives a 404.

Should this be updated? And, by the way, any suggestion for a short
complete example of the markup required?
-Jodi