Diwaker Gupta wrote: >> Looking closely at document2html (in common) I have my doubts about >> the usefulness of translating every id-attribute into an html-anchor >> (<a name="..."/>) element. Especially when I start using ids for CSS, >> it means that this drastically increases the volume of the final page >> by creating lots of anchors that nobody needs. > This also breaks HTML compliance in many cases because in many places the > value of the id attribute is constructed using the contents of <title> tags, > for instance. IIRC, a URLencoder() is run on the truncated string to generate > the ID attribute, and this often contains characters (such as +) that are not > allowed by XHTML in id attributes. Good point though a different problem really. Would you consider fixing the way the IDs are generated when there are no ID-attributes present. Is there another algorithm we can use that will not give us forbidden chars in to ID? >> Wouldn't it make more sense to limit this anchor-processing of >> id-attributes to a few structural elements like headings and just copy >> the id without processing for all other elements? > +1. What do you mean by copying the id for all other elements? I mean that an ID-element present in a source element would be copied into the resulting html-element without any further processing. So that I transform <figure id="bla"> would result in something like <img id="bla"> and can be used in CSS-styling of the output. -- Ferdinand Soethe
