Re: Framing Use Cases (was: Annotating Bibliographic Information)

If I create an annotation (public or private) on an amazon page for a book that simply says “I own this” (perhaps as a tag)… then to me that’s a perfectly valid annotation.   Isn’t that the equivalent of a card catalog entry?  Perhaps the LoC annotates the canonical copy of an EPUB with a holding record that says they have a physical copy of the book located in their archive, when it was acquired.  I might discover that annotation, and know that I can go see the physical copy— perhaps it’s a rare book, or an illuminated manuscript.  That would be helpful to me, and I would be glad they did that.  The addition of additional information about the condition of the cover doesn’t somehow magically transform the record into an annotation to me, the notion that a copy exists in some physical space is enough by itself.

I don’t think annotations have to augment the original source in a way that “enhances” it semantically or provides contextual information.  That is certainly one use of annotation— but not the exclusive one.  

A similar example is if I tag/recognize a song I’m listening to with Shazam (Soundhound, et al).  That act simply says “I listened to this”.  It doesn’t mean I "liked it", or "have it", but it’s still a useful act (to me) and I should probably be able to export my Shazam “annotations" so I can reuse them in other ways.  (Like creating a playlist).  It’s a shame that these apps we use are not open today, but perhaps in the future they will be, and perhaps the annotation data model can serve that use case.  I don’t see why not.

Another one would be to recreate the Pocket app w/ a browser button that simply creates a page level annotation, privately under my annotation account, with a tag of my choice— “SaveForLater”— for instance.  Since I’m annotating lots of other articles with “real annotations" having these bookmarks added to my stream is useful, and allows me to search and filter through one interface for different kinds of things.  

TL;DR: The data model is agnostic to semantics.  In my mind, there’s no more validity to an annotation that adds to the meaning of it’s target from another that doesn’t.  Introducing that would require that you and I and everyone else agree what is “meaningful", which is ... difficult.

On January 7, 2015 at 11:11:44 AM, Benjamin Young (bigbluehat@hypothes.is) wrote:

Hey Ray (et al),

First, thanks for contributing these use cases! You're one of the first (and sadly only...) so far, and it's appreciated. :)

Second, I'll admit to being confused by the use of "annotation" (broadly) for things like Cover Art and Holding records. However, the paper you sent around contained this valuable description:
In the BIBFRAME view, the purpose of an annotation is to:

express an opinion, for example a review of a book;
attach institution specific information, for example holdings of a book; or
contribute enhancements to a resource description, for example cover art or summary descriptions.

The key words that stood out to me are below (per list item):
1. "express an opinion"
2. "institution specific"
3. "contribute"

Your paper made it clearer (at least to me) that the actors in the stories were "sitting outside" the canonical data sources and / or were generally creating data elsewhere first, then (perhaps) contributing it.

I am still a bit puzzled by the choice of the word "annotation" for #2 and #3 above, and perhaps you can clarify it further. I certainly qualify as a "library layman." :)

To me (at least), an annotation "equation" goes something like...
original resource + annotation = a unique and greater whole

In my head (at least), I don't feel like a card in a card catalog is an annotation--simple a "statement of record" for that catalog.

However, if that same card had the sentence, "The cover's torn...sorry" scrawled on it, that note from the librarian would certainly be an annotation--but I'm not sure I'd consider the card itself to be an annotation...just a "database entry."

Perhaps you could clarify why the word "annotation" was chosen for use with holdings (#2), and that would further clear things up.

Thanks again for providing the use cases, Ray. I hope some of my musings are helpful. :)

Take care,
Benjamin

P.S.: I now have a strong desire to go find an old library card catalog somewhere and breathe in the mustiness.... :)
--
Developer Advocate
http://hypothes.is/

On Mon, Dec 29, 2014 at 10:16 AM, Denenberg, Ray <rden@loc.gov> wrote:
Doug - I responded in part last week and promised to get back to you on the second part. Sorry for the delay  (holidays got in the way).

Briefly, no, I do not agree with your  "exemplar theory" as the basis for all use cases.  That's not to say I'm right and you're wrong, only to say I think we need others to weigh in.

I do not think it is useful to try to cast all annotations in terms of (as you put it) "here's a bunch of people  in the wonderful future world of web annotations! "  because the essence of the use case becomes lost in the details. I think in many cases a succinct statement of the use case or story  is more useful.   That said, I think it is very useful to have some of the use cases expanded as you describe, but not all (or even most) of them.

But again, I think we need other opinions on this.

Ray

> -----Original Message-----
> From: Doug Schepers [mailto:schepers@w3.org]
> Sent: Friday, December 19, 2014 7:43 PM
> To: Denenberg, Ray; 'Web Annotation'
> Subject: Framing Use Cases (was: Annotating Bibliographic Information)
>
> Hi, Ray, folks–
>
> TL;DR:
>
> Let's describe our use cases in a way that highlights the most common uses
> of Web Annotations.
>
>
> Long-winded version:
>
> I took the liberty of converting your .doc file into HTML format
> (attached); this is the preferred format for W3C, for obvious reasons,
> not least of which is that it can be viewed in email clients and our
> online archives without opening it in a separate app. I hope this helps
> the conversation.
>
> I took the further liberty of adding a couple of links that I found
> helpful in reading it; I didn't know before what BIBFRAME was, so I
> linked to the Wikipedia article on that; I also added a link to the
> BIBFRAME Annotation Model whitepaper, because I was confused about the
> way you were using the word "annotation". I hope my links help others on
> this list.
>
> As I said on the last telcon, if the deliverables of this WG help your
> use case, that's great; if the data model or other specs can be tweaked
> to be more helpful to your particular need, that seems like a good idea.
>
> The web has become the overwhelming success it is today because people
> took the basic building blocks provided to them, and used them for
> creative purposes unforeseen. Few people would have thought at the
> beginning of the Web, "hey, let's take this document format and use it
> to build full-powered applications." But that's what we did, and that's
> what makes it great.
>
> So, if you see some of the basic building blocks of Web Annotations and
> think, "hey, we can use that to build a distributed bibliographic
> reference system that allows 'class inheritance' or 'subtyping' of
> bibliographic entries to add information (like whether our library has a
> copy of this book)", then that's useful, especially if it means that
> some of that content can be directly exposed through the web more
> easily. If we don't have to do anything special to meet your use case,
> and it just works out of the box, even better!
>
> For me, however, that's not an exemplar of a Web Annotation. It's a
> specialized use with some overlap. Your technical terminology uses the
> word "annotation", but I think you mean it in a slightly different sense
> than what I'd call an annotation (and more like what I'd call
> "inheritance").
>
> (For those not familiar with exemplar theory: if I ask you to name a
> type of bird, you're likely to say "pigeon" or "starling" or "hawk" or
> "duck"; you're less likely to say "ostrich", and even less likely still
> to say "penguin". These are all fine birds, but the latter 2 have fewer
> features in common with other birds. Members of a category that have
> more features in common with other members of that category are known
> as
> "exemplars".)
>
> If an annotation only has a link selection, and no body (e.g., a
> highlight), is it an annotation? How about if the body of the annotation
> is simply a link to another lengthy resource, or there's no body but two
> link selections? How about if the body of the annotation is a link to an
> image or video, which is then rendered inline in the annotation viewer?
> Yes, those are all annotations, but they aren't exemplars, in my opinion.
>
> Why is this relevant? When we're collecting use cases, we're not just
> making a list of all possible uses for a technology. We're communicating
> an aspirational goal for our desired outcome to a wide community of
> potential stakeholders, in an effort to get them involved because they
> see relevance to what they're doing; so you might think that we want to
> cast the net as broadly as possible. But we're also trying to convince
> them that this effort is worth investing resources in, and that the odds
> of success are high, which means that we are clear on our goals and
> priorities, and that we are focused on a set of smaller gains that lead
> toward broader wins.
>
> With that in mind, my preference would be for our use cases and
> requirements to be framed in terms of those exemplars that the broadest
> audience is likely to relate to. It might be as simple as casting the
> actors. So, when I think of your cover art example, I might say:
>
> Anna is reading a short story on her ebook reader, and the main
> character reminds her of a drawing by her friend; she annotates an
> instance of the character's name with a link to the drawing online. Ben
> works at a library, and has gotten permission to add the short story to
> their ebook collection; he wants to find cover art for it, so he
> searches an online annotation service for annotations on the short story
> that include images, and finds Anna's annotation. Ben obtains permission
> from the artist to use the image, and publishes the short story with its
> new cover art.
>
> (I might have added some social aspect to it, like, "Ben replies to
> Anna's annotation asking for the source, and Anna connects him to her
> friend.")
>
> This still covers your use case, but it does so in a narrative that
> emphasizes different aspects of the desired ecosystem; it's got the
> distributed aspect, an end user reading and annotating a selection in an
> ebook, online annotation services, linking to online image services,
> social media, search and discovery, all wrapped up in a story, with
> characters who have motivations. It feels less like "there's a
> collection of data in a database, we don't care how the data got there,
> and we searched the database", and more like, "here's a bunch of people
> in the wonderful future world of web annotations! Huzzah!" (And hey,
> maybe Anna works at a library, too, so this might be the very same story
> you told.) It includes an exemplar act of annotation, something that
> could only happen with web annotations, rather than a story that could
> substitute "Google image search" for "annotation" (of course, Google
> could index annotations to add relevance to their image search...).
>
> You might think this is trivial, but I think it makes a real difference
> in the story we're trying to tell people about this new thing called
> "Web Annotations", which they are trying to distinguish from other
> technologies.
>
> Does that seem reasonable to you?
>
> Regards-
> -Doug
>
> On 12/19/14 10:50 AM, Denenberg, Ray wrote:
> > At the recent call I promised to elaborate on the use cases I had
> > suggested, including cover art, and to try to explain why I think that
> > cover art really is an annotation use case.   In order to do that I need
> > to provide background on some of the thinking within the
> > library/bibliographic community about annotations (specifically the
> > thinking with the BIBFRAME project, http://www.loc.gov/bibframe/).  So I
> > have prepared a (roughly) two-and-a-half page discussion paper that
> > tries to provide background, in “layman” terms (i.e. for those not
> > familiar with library/bibliographic terminology).  The paper is attached
> > and I hope you will take the time to read it and to comment.
> >
> > (Note: I am not sure if this is the proper way to contribute a paper; if
> > not, let me know how.)
> >
> > Ray
> >
>