I guess in that early stage it seems OK. Compared to a JSR it probably is no further than EDR stage right now.
Where say you have a particular state of the data set and either Java or JS client, tagging them in a consistent way could make most sense. That was the main reason for any tags other than /release you'd find. Especially when there's something to demonstrate it is often a good idea. On Wed, Aug 5, 2015 at 6:34 PM, Reza Naghibi <r...@naghibi.com> wrote: > While your point is valid, it doesn't change the fact that a breaking > change is being introduced into the specification. > > On Wed, Aug 5, 2015 at 12:28 PM, Werner Keil <werner.k...@gmail.com> > wrote: > > > That's why occasional snapshots would also be nice here, even if they're > > like javascript-client-2.0-alpha ;-) > > > > When ApacheCon is closer I'll have a look, but not yet. > > > > On Wed, Aug 5, 2015 at 6:25 PM, Reza Naghibi <re...@apache.org> wrote: > > > > > Small chance... but just incase someone is working or looking at the > > > attribute spec, its going to change. Im currently putting together the > > > javascript client [1] and given that JSON is a first class citizen of > > > javascript, I want to make sure the JSON spec is perfect. Right now > > > attributes are this: > > > > > > "attributes": [ { Attribute }, { Attribute }, ... ] > > > > > > This will change to this: > > > > > > "attributes": > > > { > > > "patternId_1": { Attribute }, > > > "patternId_2": { Attribute }, > > > ... > > > } > > > > > > This will much optimize the structure for javascript and likely other > > > clients. > > > > > > [1] > > > > https://svn.apache.org/repos/asf/devicemap/trunk/clients/2.0/javascript/ > > > > > >