Clay Leeds wrote:

Does the idea that there would be intermediate results mean that a "human" could determine which is the best to perform the final layout? I'm thinking of a system similar to how some OCR programs enable the user to contribute to the process of recognition when the OCR program has problems determining a word or character. (FYI: OCR=Optical Character Recognition--used in scanning text-based documents which are converted to text for archiving, indexing, etc.).

If so, could the implementation offer some way of "saving" the best method? I would think it would work like a userconfig file.

I meant to add my reasoning. Since one can assume that generating multiple iterations will be processor intensive, if there existed some sort of "config" file identifying clues as to how the hz algorithm should proceed for a particular file, it would shorten the processing time, as well as ensure that the output was always the same.


I don't see it this way. The multiple iterations occur within the rendering of a particular document. I saw the human intervention as a way of interactively deciding on one layout among a number of possibilities which the processor was unable easily to distinguish, not as some asynchronous event. While I can see the possibility of "learning" about the scoring of layout possibilities from such interactive input, I can't see that sets of rendering decisons need to persist.

"Lord, to whom shall we go?"

To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, email: [EMAIL PROTECTED]

Reply via email to