- From: Herbert Van de Sompel <hvdsomp@gmail.com>
- Date: Mon, 26 Oct 2015 14:06:08 -0600
- To: Bernadette Farias Lóscio <bfl@cin.ufpe.br>
- Cc: Erik Wilde <dret@berkeley.edu>, Jeremy Tandy <jeremy.tandy@gmail.com>, Phil Archer <phila@w3.org>, "public-dwbp-wg@w3.org" <public-dwbp-wg@w3.org>, Herbert Van de Sompel <hvdsomp@gmail.com>
hi all, I provided examples to Bernadette related to Memento and versioning. It's all hypermedia/REST/HATEOAS along the lines of what Erik promotes. But I think there are very real opportunities in the document to explicitly promote the (best) practice of providing typed links, eg in HTTP responses. For example, Example 8 shows metadata that describes a dataset. But it does not show how that description can be discovered by a client. This can be done using a typed link the the HTTP Link header provided when accessing the dataset URI. Assume: * http://example.org/dataset-001 is the dataset URI * http://example.org/metadata/dataset-001is the URI of metadata that describes the dataset In that case: => Discover metadata via dataset curl -i -H http://example.org/dataset-001 HTTP/1.1 200 OK Link: <http://example.org/metadata/dataset-001>; rel=“describedby”; type="some mime type", <http://example.org/metadata/dataset-001>; rel=“describedby”; type="some other mime type", => Discover dataset via metadata: curl -i -H http://example.org/metadata/dataset-001 HTTP/1.1 200 OK Link: <http://example.org/dataset-001>; rel=“describes” cheers herbert On Fri, Oct 23, 2015 at 6:25 AM, Bernadette Farias Lóscio <bfl@cin.ufpe.br> wrote: > Hi all, > > Thanks a lot for the great discussion! > > I'd like to make two comments: the first one about versioning and the other > one about the links. > > As we discussed before in our WG, it is not easy to define when to create a > new dataset or a new version of an existing one. In the DWBP document we > give some ideas about this, but the choice is up to the publisher [3]. On > the other hand, we have BP to tell how to do to deal with versioning. In the > current version of the document, we have two BP: Best Practice 8: Provide > versioning information and BP 9: Provide version history. There are some > vocabularies that we can use to describe the version of a dataset as well as > the relationship between dataset versions. However, I couldn't find terms to > describe the modifications of one dataset with respect to its previous > version. In BP9, we argue that this information should be available. I was > wondering if we can use the Web Annotations Model to describe this [1]. Does > it make sense to you? Do you have other suggestions? > > Concerning the links, I agree that it is too ambitious to go into the > details of hypermedia. However, as Phil proposed, we will have BP saying > that resources (datasets and data items) should have unique identifiers. We > also have BP related with how to provide access to datasets [2]. We already > have a BP that tells to Follow REST principles when designing APIs. Maybe, > this BP may be improved to cover some aspects related with Eric's concerns. > What do you think? > > Cheers, > Bernadette > > [1] http://www.w3.org/TR/annotation-model/ > [2] http://w3c.github.io/dwbp/bp.html#dataAccess > [3] http://w3c.github.io/dwbp/bp.html#dataVersioning > > 2015-10-16 17:06 GMT-03:00 Erik Wilde <dret@berkeley.edu>: >> >> hello jeremy. >> >> On 2015-10-09 23:12, Jeremy Tandy wrote: >>> >>> Now, Eric's point [9] is that there is a "difference between 'web data >>> only' and the 'web of hypermedia-driven services'" and that "'webby >>> data' is a necessary but not sufficient condition to have hypermedia. >>> [which requiresproviding navigational affordances to get things done >>> with that data." >>> I see that in the vast majority of cases, the data is accessed via a >>> service end-point ... even if it is a trivial HTTP Get. But there are >>> cases where (as I said in point #3 above) that you simply want to use >>> URIs as identifiers. This clearly is not hypermedia. I wonder if there >>> are two levels of requirements here? At this point, I'm unable to unpick >>> this distinction further, but I'm sure it will be relevant in the >> >> >> thanks for pointing this out, and i absolutely agree that there are two >> levels at play here. one level that's needed for "just data" is the ability >> to identify resources, and this is what RDF is all about. and generally >> speaking, non-RDF models may want to use similar policies for identifying >> (some of their) resources via URI. doing this means that identifiers are >> context-free and thus globally usable, which is a very valuable thing when >> you start to combine data of various origins. >> >> however, links are a fundamentally different issue. they are not >> identifiers of data resources, but instead identifiers of hypermedia >> resources which provide support for interactions. in order to be useful, >> links must be typed >> (https://github.com/dret/hyperpedia/blob/master/concepts.md#link-relation-type), >> so that clients can decide which links to follow to accomplish their >> application goals. >> >> as phil suggested, maybe it is too ambitious for DWBP to go into the >> details of hypermedia, and why and how it matters. but as i said earlier, it >> seems to me that in order to make data available on the web, most often >> there are services associated with it. even if often just rather simple ones >> such as the ability to page through a collection, or to filter it. providing >> guidance on how to do this in a webby way would be very valuable. >> >> thanks and cheers, >> >> >> dret. >> >> -- >> erik wilde | mailto:dret@berkeley.edu - tel:+1-510-2061079 | >> | UC Berkeley - School of Information (ISchool) | >> | http://dret.net/netdret http://twitter.com/dret | >> > > > > -- > Bernadette Farias Lóscio > Centro de Informática > Universidade Federal de Pernambuco - UFPE, Brazil > ---------------------------------------------------------------------------- -- Herbert Van de Sompel Digital Library Research & Prototyping Los Alamos National Laboratory, Research Library http://public.lanl.gov/herbertv/ ==
Received on Monday, 26 October 2015 20:06:37 UTC