macPT,
The advantage of plotting your drawings on vellum, signing the plots and making blueprints of them, is that you can verify the source of any drawings people are working with. I like blue/white print machines because you can only print from the vellum original, and you cannot modify the blueprint. Xeroxes are evil.
You want to issue drawings to vendors in a portable format, and you do not want them modified. You want to be able to visit your vendors and verify that they are working with your approved, unmodified drawings.
Am I right?
Solutions:
1.] Issue blueprints as noted above. This does not stop you from issuing DWG/DXFs, but these are unofficial. The blueprint is official and always correct, regardless of what is on the DWG/DXF file.
2.] Write your files out as PDF. To an engineering/manufacturing shop, this is a read-only format, although I would not trust graphics shops. It will probably be difficult to get large prints off of the PDF, and there will be no way to extract part geometry off of it.
3.] Write your files out as DWG/DXF and record the exact size (to the byte) of the file, and the date. It will be hard for your vendor to modify the file without changing the size and date. It will be harder if you issue in DXF rather than DWG.
4.] Attach a digital checksum to your files. There must be technology for this. If somebody modifies something, the checksum changes. This is a more reliable check than just recording the file size and date
Consider what happens when you visit your vendor and see that someone has plotted a copy of your electronic file. How do you verify that the file is an exact copy of what you sent? I can load a read-only file, change a dimension and then print. All I cannot do is save.
Consider how you apply limit tolerances to your scale CAD models. Eg. 16.15/16.05DIA. I draw my parts to exact nominal size, in this case, 16mm, and set the tolerances to +0.15.+0.05. Another designer might arbitrarily use MMC, setting the lower limits for his holes and the upper limits for his shafts. In this case, the model would be 16.05mm, and the tolerances would be +0.10/0. Modeling the hole and shaft at LMC is a dumb idea, that will not stop everyone from doing it. Another approach is to model at 16.10mm, and set the tolerances to +/-0.05. How to you interpret a dimension like 23.72/23.61DIA from the model?
What this all shows us is that model dimensions do not provide a reliable description of dimensions with accurate tolerances. Those blueprints look better and better to me.
JHG