- From: elf Pavlik <perpetual-tripper@wwelves.org>
- Date: Wed, 14 Oct 2015 18:01:05 +0200
- To: James M Snell <jasnell@gmail.com>
- CC: Harry Halpin <hhalpin@w3.org>, "public-socialweb@w3.org" <public-socialweb@w3.org>
On 10/14/2015 05:46 PM, James M Snell wrote: > The short answer is: nothing breaks. Implementations that support the > `application/activity+json` media type will understand that AS 2.0 is > a JSON-LD based syntax and will use JSON-LD mechanisms to process > them. The implementations that intentionally choose not to use JSON-LD > mechanisms to process are given sufficient warning that interop issues > could arise from that decision. I have impression that by saying "To have functioning interoperability you need to process it as JSON-LD, but you can consider it as 'plain JSON' if you don't mind interop issues to arise" we may mislead people that they really can treat it as 'plain JSON'. I also have hard time to understand why Harry while suggesting that most people will ignore the JSON-LD aspect, doesn't mention that interop issues could arise if one really does it... > > Also,tThere is absolutely nothing stopping an implementation from > using the `application/ld+json` media type when transmitting Activity > Streams 2.0 data if they have good reason to do so. The > `application/activity+json` media type is more specific, yes, but it > *does not break anything*. > > Also, quick correction on Melvin's post: the open issue on github is > *not* about changing the media type to `application/activity+json`. > It's about requiring the use of the `application/ld+json` media type > with an additional profile parameter. I've seen absolutely no reason > to require the `application/ld+json` media type and after implementing > AS 2.0 support in a few applications, I see absolutely nothing that > breaks or doesn't work by using the `application/activity+json` media > type. > > I'd very much like to just put the whole media type issue to rest as > it really is a red herring. I've heard many times that it "breaks" > things without seeing any actual evidence that things actually do get > broken. > > - James > > On Wed, Oct 14, 2015 at 8:28 AM, elf Pavlik > <perpetual-tripper@wwelves.org> wrote: > [snip] >>> >>> Can you point to what tooling breaks? >>> >>> The larger issue with some of the RDF-centric approaches is that while >>> we can recommend Link HTTP headers and MIME types, most tooling that I >>> know of ignores both of these, and would also not expand JSON-LD (since >>> most tooling is JSON-centric) >> >> Systems which integrate AS2.0 data from multiple sources need at minimum >> expand CURIEs to distinguish properties using full URIs >> >> http://www.w3.org/TR/activitystreams-core/#aggregation-of-extensions >> >> Also our charter says: >> >> "A transfer syntax for social data such as activities (such as status >> updates) should include at least the ability to describe the data using >> *URIs* in an extensible manner, time-stamping, and should include a >> serialization compatible with Javascript (JSON) and possibly JSON-LD." >> >> >> >
Received on Wednesday, 14 October 2015 16:01:08 UTC