Experts within Toyota on TPS will always want to drive in the direction of the ideal of one-piece flow. They believe in this quite passionately. In a Toyota assembly plant this looks like a super long continuous flow line. The plastics plant look like a process island of molding machines though there is a clear flow of raw materials to finished bumpers that are built in sequence to the assembly line. The body shop is mostly flow lines as is paint. The stamping plant is another set of process lines to build up a major stamping which is based on replenishing a supermarket of stamped parts pulled by welding. Now with the newer set-parts system (kitting) and minomi (eliminating containers) is getting closer to flowing stamped parts a car at a time to welding.
Many Toyota suppliers are building smaller subassemblies like brake systems for different customers based on different designs and it makes sense to have in parallel multi=process cells fed by manufacturing processes like stamping or injection molding. So you will see cells, individual processes, and various ways of pulling product to the final assembly line.
The desired future state of material and information flow for any operation will be different, as represented by a value stream map, depending on the operation. But in the early days the big hitter that gave big results was tearing up the factory to make cells and the immediate results were stunning. As those companies matured they learned there was a lot more to TPS then building cells, like standard work, parts presentation, pull systems, daily kaizen, developing people and culture, etc.
I agree that if we walk through hundreds of plants we still will see most are organized as process islands and have not gotten as far as having well designed cells or multi-product, multi-process flow lines. The lean movement hit in waves of activity, often then followed by waves of inactivity as financial or other pressures hit the firms. Of those who put in a lot of cells, some continued to evolve and others never got farther then the physicals of the cell with little real improvement beyond that. There is no question that there is much to be done although I have seen plants without cells that are farther advanced culturally in continuous improvement then plants with cells that on closer examination are really lean facades.