Fwd: Changes vs. new element

And another post that should have gone to the list, defending my
desire to provide recommendations for providing structured data around
holdings.

Apologies, everyone, for not paying closer attention to my To: header :/

---------- Forwarded message ----------
From: Dan Scott <denials@gmail.com>
Date: Thu, Aug 1, 2013 at 2:55 PM
Subject: Re: Changes vs. new element
To: Tom Morris <tfmorris@gmail.com>


On Thu, Aug 1, 2013 at 2:27 PM, Tom Morris <tfmorris@gmail.com> wrote:
> On Thu, Aug 1, 2013 at 1:54 PM, Dan Scott <denials@gmail.com> wrote:
>>
>>
>> > For availability location there is http://schema.org/availableAtOrFrom.
>>
>> Supporting my need to put together a clear wiki-page version of my
>> original proposal[1], I actually had mapped "shelving location" to
>> http://schema.org/availableAtOrFrom.
>
>
> Unless by "shelving location" you mean West Branch vs South Branch vs the
> library systems's web site for e-loans, I don't see how anyone except a
> librarian (ie *not* the target for this markup) could possibly care.  Does
> BestBuy include that the DVD is in aisle 6 or aisle 9 in their offers?  I
> bet not.

When I go to a large store like Home Depot or Best Buy to pick up an
item that I've researched online, I often wish that they did include
that in their online information. Because otherwise I frequently have
to either wander the aisles guessing at where it might be located
(possible motivation for the store: get me to buy extra stuff while
I'm wandering), or track down a sales associate who can point me in
the right direction. It's inefficient. (Come to think of it, I'm sure
I have run across sites that offer aisle information before... hmm.)
So yes, I do believe that someone (other than a librarian) who values
their time could care.

>> > There is also http://schema.org/inventoryLevel which could be used.
>>
>> Also, I tossed around the idea of inventoryLevel, but if we're mapping
>> individual Offers to individual (e.g. barcoded) items, that will
>> always either be 1 or 0. Not sure how useful that is? That's why I
>> went with AggregateOffer to provide per-branch counts.
>
>
> Again, why would a consumer care what the bar code on their book says?  Do
> they really say "I want copy 2 because my friend told me the pages are
> cleaner?"

If you've ever purchased a used book, yes, that's a possibility.
Copies signed by the author are another potential use case. Or
large-print copies versus paperbacks.

This is all very system dependent, but assuming that in this effort
we're not going to fix all library systems and cataloguing practices
so that we have perfect metadata input & output as part of this
effort, sometimes people _are_ going to want the thing that is only
identifiable by call number "FIC KIN LARGE PRINT", possibly down to
the barcode level...

Hey, on that note,
http://www.abebooks.com/servlet/BookDetailsPL?bi=4213019283 uses
schema.org via RDFa Lite to mark up their search results and item
details. Product + Book & Offers. Nifty!

> As manufacturers start putting RFID chips in packages of razor
> blades, is CVS going to have a separate offer for each serial number?  Of
> course not.  The serial number is important to them to help track
> "shrinkage" (ie theft), but the consumer could care less as long as there
> are at least as many packages on the shelf as they want to buy.

Agreed. That's a completely different use case, though, isn't it?

> Rather than starting from the premise that a library is a completely unique
> beast, I'd start from the consumer's point of view that it's an institution
> providing goods and services just like hundreds of other institutions that
> they deal with in their daily lives.

You'll note that I have largely been trying to take the tack you've
suggested by reusing the existing classes and properties that
schema.org already offers. Please don't try to cast me as a
navel-gazing librarian, that's not going to be productive.

If we have the information at hand, what's the harm in marking it up,
or providing suggested mappings? In the best-case scenario, as Thad
suggests, it gets integrated in interesting ways down the road. In the
worst-case scenario, it gets ignored. Doesn't seem too terrible to me.

Received on Tuesday, 6 August 2013 17:15:49 UTC