Does any software currently look for links with type application/json? I could see that being useful, but only if we also specify the "schema" for the returned JSON object. i.e. instead of a link of type application/json a link something more like application/book+json where the returned JSON object has a known structure. This would be similar to how RSS feeds can be indicated using a link of type application/rss+xml. If the link were just application/xml how would you know it was RSS? Open Library already has its own JSON "book" schema but a metadata format is not so useful if it's just used on one site! - mang > The use case is that software that's looking at that page can easily see > that there's a json representation of it available, and where to get it. > This is not for raw human use, it's for software use. > > Examples of software that might be 'looking' at an OpenLibrary item > detail page include: > * a browser plug-in like LibX, looking over the user's shoulder as he > browses the web. > * A web spider of some kind > * An OAI-PMH client of an OAI-PMH feed that includes OpenLibrary pages. > (feed provided by OL/IA itself, or an OAI-PMH feed provided by a third > party but indexing OL content). > * Many more uses that we can't predict right now, but which people will > come up with when the architecture is there. > > The link that Ed suggests is to advertise in a standard > software-understandable way "There is a JSON representation of this > resource available." A button that can be seen by humans may or may not > be a good idea, depending on if you think humans are interested in > clicking to get a JSON representation (I doubt it). The link Ed > suggests is for a different purpose, for automated discovery of the json > representation of the resource represented by the url. > > Jonathan > > Karen Coyle wrote: >> Ed Summers wrote: >>> I guess the main thing I wanted to >>> communicate is that you could simply add: >>> >>> <link rel="alternate" type="application/json" >>> href="http://openlibrary.org/api/get?key=/b/{open-library-id}" /> >>> >>> to the <head> element in OpenLibrary HTML pages for books, and that >>> would go a long way to making machine readable data for books >>> discoverable by web clients. >>> >>> >>> >> Ed, the first thing that comes to my mind when I see this is: button. >> Unless folks will be blindly crawling the OL* I don't know how they'll >> get to a particular page to execute this code, except by being a >> person searching and getting the web page. (If they are using the >> search API they get back a list of IDs from which they'd create one or >> more 'get' commands like this one.) A download button on the page >> would make sense, but mainly if it downloaded into a usable format >> (EndNote, MARC). All this to say that I don't get what the use case is >> for this particular bit of code -- but I'm assuming you had one in >> mind. Please do tell! >> >> * If anyone wants the whole OL database, json dumps are available: >> http://openlibrary.org/dev/docs/jsondump >> > > -- > Jonathan Rochkind > Digital Services Software Engineer > The Sheridan Libraries > Johns Hopkins University > 410.516.8886 > rochkind (at) jhu.edu >