# The following was supposedly scribed by # Bruno Postle # on Tuesday 24 August 2004 10:20 am: >[snip hub and node model] > >> In this scenario, how are the rhizopod-format databases affected by a >> format-incompatible upgrade to the 3D relational modeling program? =A0Th= ey >> aren't. > >That's correct so long as applications are always importing and >exporting to and from the format rather than accessing the data >directly. I think you're missing my point here. Your concerns are valid, but I think= =20 the "version per-entity" issue mostly addresses a hub which is more=20 complicated than rhizopod. The 3D modeling program would *never* address the rhizopod data directly=20 (unless it plans to do it in a very simple way.) Also, you should not see= =20 the rhizopod<->sturgeon connector as simply an import/export application. = It=20 should rather be seen as a daemon which does a real-time translation betwee= n=20 the two. The hub-node model comes from import/export concerns, but when the connecto= rs=20 are between persistent databases, they don't have to be as dumb as when the= y=20 hook to dwg/ps/svg files. The point of the hub-is-a-node principle is that you have to make sacrifice= s=20 and difficult decisions to "talk-down" to simpler formats. Wouldn't you=20 rather have those issues handled by a connector than have to deal with them= =20 directly in your code? So, the applications address the data in the hub which is most suited to th= eir=20 task. Thus, the 3D relational modeling program would NOT cause=20 version-incompatible changes to show up in the rhizopod directory. All applications in the scenario *are* addressing the data directly, but th= ey=20 get to see it in the shape that is best suited to them. =2D-Eric =2D-=20 "Matter will be damaged in direct proportion to its value." --Murphy's Constant