Tue Jul 02, 2013 12:12 am
Clayton and Dale,
The problem I'm running into is that new constraints have been added to the workflow that conflict with my previous methods. These cause the the Mac OS to become quite testy. Worse, even when I override it, I'm never quite sure if I'm burning my bridges and creating other problems down the line. That is why I'd rather be able to get into the structure and extract copies, instead of renaming and moving, and otherwise screwing things up profoundly. I'm not against change and adapting to new structures, especially at the start of projects such as Creator, but I've already got a large body of legacy files and folders that are all under active development.
Dale's characterization of the problem as a single serial workflow contrasted with parallel projects is extremely helpful, by the way. I'm someone who needs to use Creator on many projects at once. The beginner likes to deal with a single project. In trying to help teachers, for example, I'll need many projects, where each project represents one beginner student.
Yes, the concept of a factory file is excellent for a controlled production setting where all the variables are nailed down and unchangeable, cast in stone or more likely plastic in this case. Introducing changes made in the default profiles may, however, introduce undesirable changes into the engineered process. It just depends what the effect of those particular changes are. Such changes just might have the opposite effect of the intended stability.
In a development setting one almost never wishes to exactly replay a previous build. The next build is the last build with a few tweaks, e.g., a slightly different stl with the same profile as before; or, the same stl with a few profile tweaks for the new feedstock; or just a slight tweak for the new bed material; etc, etc, etc. So, with the bundling function, development also requires an unbundling function. This isn't a dissolution of the previous factory file, but a way to peek into it and extract pieces to use in the next build with some incremental changes. The previous factory file is still there as a security blanket just in case a roll-back is needed. Compaction may complicate the unbundling tool. It may be better to have an separate compaction tool that can be applied to a decomposable factory file when it is ready to be preserved for production. At this stage in 3D Printing, we are almost always in development mode, very seldom in production mode.
(I spent many years engineering early software configuration management tools, e.g., UNIX make, DSEE (Apollo/HP) ancestor of ClearCase, ancestor of Rational, and others. 3D printing problems are pretty simple by comparison, but still not easy.) One last comment. The parameters of FFF 3D printing have many interactions. The web of interactions sometimes depends on the design of each type of printer. It would be almost impossible to come up with disjoint sets of independent parameters. It is better to just try to save the whole set of parameters of each build, brute force.
I'm willing to help you kick around some of these ideas if it will help. However, my personal agenda is somewhat different.
Best,
Rich Thall