is it possible to crudely in-vitro translate?

if you had synthetic GFP RNA, could you mix it with a crude cell
extract and see a glow? what about a doing
chromatography/gel-separation on the proteins and mix each one with
the supernatant of a centrifuged cell extract, then mixing with RNA to
look for a glow. in the first case you hope to get lucky, in the
second you would be trying to find the ribosomes, and add in some
amino acids and tRNAs.

To find T7 RNA polymerase from your extract(gotta be in the strain),
you would get this plasmid with a T7 driven GFP:
http://www.addgene.org/pgvec1?f=c&identifier=24387&atqx=(t7%20AND%20gfp)%20AND%20vectype:Bacterial%20Expression&cmd=findpl
(Plasmid 24387: pHL32)

To find normal RNA polymerase from a cell extract, you would get a
"normal" GFP plasmid:
http://www.addgene.org/pgvec1?f=c&identifier=17823&atqx=gfp%20AND%20vectype:Bacterial%20Expression&cmd=findpl
(Plasmid 17823: pBAD*RFPEC2)
http://www.nature.com/nbt/journal/v14/n3/abs/nbt0396-315.html

then feed protein isolates to the plasmid, then feed each of those
samples to your confirmed-to-work translation system, and you find the
glowing sample... that sample had RNA polymerase in it.

So theoretically this logic could yield both a transcription and
translation system, all starting from a GFP RNA, and a GFP DNA.

I can't wait to try this!