Why do resultsSerializers require a stream vs returning value?

[gdbassett]

I feel like I'm missing some obvious reason why JSONResultSerializer wouldn't just return the stringified json object (r

rdflib/rdflib/plugins/sparql/results/jsonresults.py

Line 50 in be13b3a

r = json.dumps(res, allow_nan=False, ensure_ascii=False)

) vs requiring a stream. It seems simply leaving stream empty (and defaulting to None) would be a simple way to do this.

Is there some over riding reason to require the results be steamed out?

(Also is there any good examples of using results parsers and serializers? I've had substantial issues finding them.)

[aucampia]

Hard to say why, we should fix it though.