Adding a version number after the fact could break existing robots. So that is not an acceptable design. It violates the robustness principle.
This would need to be a separate file, probably "robots2.txt". The proposal assumes a time-based robot with centralized control. I seriously doubt that any of the WWW-wide search engines still use robots like that. The rate directives, in particular, might not be implementable by any of the high-volume robots. Has there been a recent survey of robots.txt usage and correctness? Last I heard, it was pretty consistent at about 5% of sites. As long as 95% of sites don't use robots.txt at all, it seems odd to make it more complicated. wunder -- Walter Underwood Principal Architect Verity Ultraseek _______________________________________________ Robots mailing list [EMAIL PROTECTED] http://www.mccmedia.com/mailman/listinfo/robots