Ben Sigelman

@tahini thank you for opening this issue. Regarding `component`, there are two concepts already, one of which is somewhat hidden. 1. As mentioned already, there is the [component](https://github.com/opentracing/specification/blob/master/semantic_conventions.md#span-tags-table) tag... I...

@felixbarny sorry for the insane delay in responding to this. I spent some time on my email filters over the weekend so that I'm not missing so much stuff I...

@felixbarny I continue to think about this off and on. The more I think about `"component"`, the less I like it... the idea there was definitely to identify the package...

@felixbarny I'm not proposing that `kind=cs` would be interpreted any differently... there would simply be more Spans with `kind` annotations. I would expect the brave-opentracing bridge to simply ignore `kind`...

@lookfwd sorry for the delay, I missed this the first time around :-/ There is huge variation on this front. I have seen plenty of "real" production environments with relatively...

(cc @tedsuo @objectiser @ivantopo @felixbarny @mabn @sjoerdtalsma @yurishkuro @cwe1ss @pavolloffay @wu-sheng)

Having described the problem, my vote is to do nothing for now. My experience has been that tags are almost always added right when a Span is started or right...

(very sorry for falling off the map PR/Issue-wise) I've fully caught up on this. It's still my opinion that in *almost* all scenarios where one is setting tags, one is...

If we're going to make a change, my favorite option is still something like @mabn's [comment](https://github.com/opentracing/opentracing-java/issues/134#issuecomment-303083932): ``` tracer.withActiveSpan(span -> span.setTag(...)); ``` This has the added benefit of not-executing any of...

@yurishkuro re your concern `#2`: I suppose we could put the implementation for the helper in the `util` module (as a static method that takes a Tracer and closure as...