Re: A serious detail point
> No, but you have stated too simple a version of the problem. Assume
> instead that you have 5,000 links to spans in one target document,
> intermixed in logical order of presentation with 5,000 links to
> 300 other documents.
> Above all, though, let's stop supporting positions by appeal to what
> will be most common. This cannot be known before XML is deployed,
> and it should not be a consideration in writing the spec, unless a
> full scenarios document is developed.
Sorry if I wasn't clear -- I'm not hypothesising about what will be
common, I'm referring to documents I already have, with a private link
syntax, which exhibit the properties referred to in my earlier