Thanks for the input!
-I probably could simplify the lines a little more than than 99%. It seems that there is an enormous removal of nodes going from raw .shp import to 99%, and the gains decrease considerably after that. That being said, it's probably just my hyper-accuracy GIS brain speaking through, and decreasing a little more past 99% would have virtually no effect on a 1:80k map.
-The press that I'm working with does a PDF X4 workflow by default. My X4 docs that run through their RIP are the issue and take ~13 hours to RIP per plate. When they run them through as PDF1.3/X-1a they only take 4 hours per plate (acceptable), but I had issues because then the black was knocked out. With the normal X4 processing (which the press had used previously) the black was auto-overprinted--as it should be. I didnt have the black objects set to overprint in the .ai doc as this hadn't been an issue in the past. My understanding is that PDF1.3 uses a transparency flattener and makes some sort of rasterized interpretation of the transparency.
I'll try to place a PDF version of the file instead of the .ai file in indesign to see if that helps anything. One would think that it wouldn't make a difference as indesign appears to only read the imbedded pdf within the .ai doc anyway, but I don't know for sure.
Matthew- I wonder if the double placement is an issue as well. I haven't submitted files yet without the double pdf placement. My sheer lack of knowledge of how PDF's and RIP's interact is frustrating. (aside: does anyone know how to render rasterized separations (plus spot channels) without owning full-on pre-press software?) If you have virtually the entire map rendered invisible via indesign pdf placement layer overrides, and only a single pdf layer (labels around the edge) visible, does the RIP still try to process all of the pdf layers for that frame/object, even though it's just saying, "oh yeah, don't show this...and the next 100 layers and 2 billion nodes.."?
I hadn't thought about producing a a mega-raster from indesign as a file to supply to the press. We're talking a 150lpi/2400dpi CMYK image i'm thinking? I'm not sure how to do this and retain the spot color channels on the contour lines and keep the black overprinted. I really like the crispness that you get from spot color contour lines and having the black all be overprinted on a single plate. But, i'd be interested in seeing some sampls of maps created from this method. Does anyone use this method?
Sam - I think you might be on to something with the postscript objects. When i save the pdf out of indesign with the X4/1.4 PDF version the filesize is ~180MB, and a PDF1.3/X-1a doc is ~550MB. This additional info might be because instead of just a partially transparent stroke, it's calculating a new appearance wherever that partially transparent stroke interacts with anything else (which is a lot) and writing that info to each of the separations? I typically render the pdf in photoshop to do test prints on plotters, and my machine (mac 18GB of ram) easily renders the PDF1.4 doc, but crashes whenever I try to render the PDF1.3 version. Would there be a way to do the overprint instead of opacity trick on say, an orange line built with two process plates? Setting my Orange line to 100% opacity and overprint definitely gives a similar effect, but it's quite a bit 'muddier' as it is multiplied on the shaded relief image.I don't think there is another way to keep it bright without knocking out the other plates?
Hans - I had not thought about gaining any size reduction by running it through Acrobat. I'll give that a go!