Last modified: 2011-03-13 18:05:38 UTC
To use <pagelist> the name of the index page must match the name of the djvu file, which is a great limitation. There should be a parameter 'file', which allows to specify the name of djvu file.
This limitation has a purpose. It preserves index pages from becoming a mess.
I don't understand what you mean by mess. Are there technical difficulties? If not, please, bear in mind that this extension is used on many different projects, so this behaviour should be at least configurable. For example on Polish Wikisource names of index pages does not have to match the name of files.
A given page of the "page" namespace should have a unique index page, because it belongs in a single book. A page should never be referred in several indexes. For djvu files the best way to ensure this is to use the same name for the index and for the pages. For older indexes that do not use "pagelist", the unicity of the index is unfortunately not ensured. Note that another method to ensure unicity of the index would be complicated to implement and more difficult to understand for users.
Nothing is more complicated for new users to understand, than the introduction of two new namespaces, Index: and Page: and the need for a separate set of pages for transclusion. It should be possible to start a new book/volume/document on a single page, which acts as its front page and at the same time performs the role of the Index: page. This page should be named after the title of the volume. Then all the pages of the volume should appear as subpages, e.g. My_book/123 If chapters need their own subpages, they can be named with something more than just a number, e.g. /chapter_17 From such a subpage, it is obvious which the parent page is, just by cutting away the slash and the number. Thus, no need for an Index: namespace and no need for a Page: namespace.