Don’t blame me, I’m only a format!

Wow, this RS274X data is distorting my design! Look at all those self-intersecting polygons that are causing my copper planes to distort! Format problem or implementation problem? Well, you say that if all CAD/CAM tools in the flow worked to exactly the same coordinate system, with exactly the same floating point resolution, the problem would not exist. No more copper-plane wrong-flood-fill-misery! Much more likely that the whole world will agree to start speaking Esperanto as from midnight tonight…

Isn’t it annoying when your design gets corrupted like that? But is the problem in the format, or the CAD software that output the data, or is it the fault of the CAM software that input the data? Where are the technical support engineers to help you sort that one out?
Or what about arcs in Gerber data? Does your CAM system break arcs into many straight segments because it cannot handle the mathematics of an arc in its CAM functionality? Maybe there should not be any arc-command in Gerber at all, since it generally causes trouble somewhere. I think we should legislate about this; take it to the United Nations!

What about the old favourite of filling copper surfaces with thousands of drawn vectors? Safer than using self-intersecting polygons, or is it…? Depends on whether the Gerber output processor in the CAD system understands how to make the vectors overlap so that gaps will not open up when making input into another CAM system with a different database resolution. But will it at the same time cause the plane to be oversized? Better put production on hold…

The same thing applies to most formats, though I can say that ODB++ is inherently immune to the self-intersecting polygon problem. Well, I would say that, wouldn’t I!

But what a lot of nitty-gritty problems there can be in making an effective data flow across multiple CAD, DFM, CAM tools! In the majority of cases it is not the format’s fault (though I think I will need professional counseling before I stop banging on about self-intersecting polygons…). No, it is all about implementation – how the format is implemented in the CAD/CAM software. That is where something more than just a format specification document is needed. Solutions-providers must cooperate to ensure best-practice implementation of a format so that the end-users  (designers, process engineers, DFM experts), in other words, the people who pay for all this, can get their work done accurately, efficiently and quickly without having to make print-outs and use other formats in parallel “just in case”. All that leads to a fog of F.U.D. (Fear, Uncertainty and Doubt) as to which data represents “the truth” about the product-design that is being worked on.
What are the opinions on this? Is it the format, or the implementation? Where is the problem, and who should fix it? And – who pays?