Notes of some things discussed today (feel free to supplement):

– supporting WorldBrain’s Memex use case (see relevant thread); a good
start may be to have a corpus of documents and annotations for those to
create a test suite with.
– relatedly, the idea came up to help our users anchor annotations
created by Hypothesis[0] or H-derived applications. This could make it
easier for such applications to start using Apache Annotator, without
worrying that their existing annotations will no longer anchor as
before. We could approach this either in usage examples, or in our code
base itself (perhaps we accept an ‘H compatibility mode’ flag?);
possibly by wrapping the actual Hypothesis anchoring code with our
asynchronous API.
– generally, we lack an easy way to describe a target selection in
multiple ways, and conversely an easy way to resolve (anchor) a bunch of
alternative selectors (using some strategy of e.g. the order in which to
try apply them).
– for approximate text quote anchoring, we may want to look into Robert
Knight’s work again[1].
– we could easily add support for anchoring an XPathSelector[2] using
modern browsers’ built-in XPathEvaluator[3]. Conversely, i.e. describing
selections as XPaths, is not as obvious to solve as there are many ways
to describe the same target.

[0] https://hypothes.is/
[1] https://github.com/robertknight/approx-string-match-js/
[2] https://www.w3.org/TR/2017/REC-annotation-model-20170223/#xpath-selector
[3] https://developer.mozilla.org/en-US/docs/Web/API/XPathEvaluator

On 30/04/2020 15:51, Benjamin Young wrote:
> Woot! Should be there today. And hopefully in time. ;-P
> ________________________________
> From: Gerben <ger...@treora.com.INVALID>
> Sent: Thursday, April 30, 2020 9:08:54 AM
> To: dev@annotator.incubator.apache.org <dev@annotator.incubator.apache.org>
> Subject: Re: Weekly calls/chats?
>
> Reminder: next call is in two hours, 15:00 UTC in
> <https://meet.jit.si/apache-annotator>

Reply via email to