Bowerbird wrote:
jon said:
How is that? Annotations can be linked to the text using the markup as "hooks" (e.g., using XPointer.) The more markup there is, the more hooks to latch onto.
please show me -- and the original poster -- an implementation that actually works, now.
Simply add an 'id' to any XHTML tag, and you can link to it using any web application. Fragment identifiers are used in many millions of web pages, if not billions. The first step to an "annotation application" (which you are driving at) is to have the underlying standards (the "hooks") worked out so as to easily allow linking by the annotation application. This is already worked out in the XML world (e.g., XPointer.) That's what I meant. In LibraryCity we are now working on such annotation and related social networking applications (e.g., blogs, wikis) to digital texts using XPointer (which includes simple 'id' links.) Since the W3C standards are open and universal, others can likewise build their own application -- no need to invent a new hooking mechanism.
Pointing to 'id' ("fragment identifiers") is the most robust and can survive various types of document edits. In plain text systems, where annotations have to hook to the content itself (rather than markup which is separate from the content), it is more difficult to prevent link breakage.
you are trying to make us believe that the text changes and the markup doesn't.
Markup can change, but in the case of using 'id', the document maintainers will be careful to keep 'id's undisturbed as much as possible during document edits. This actually allows *major* changes to documents and yet keep existing links unbroken. Can this be done with plain text? Not as easily (it's not impossible, but requires some sort of mapping system, or a knowledge of all the known externally-generated links into the original text document.) For example, an author may issue some work, and later revise it by rewording paragraphs, adding new paragraphs, new chapters, etc. If they are careful, they can assure integrity of existing links into the updated Work by assuring the 'id's are properly preserved and placed where they should be. Here's an example: ====================================================================== [First Edition of a work] <p id="1234">First paragraph.</p> <p id="1235">Second paragraph.</p> [Second Edition] <p id="1234">First paragraph with some minor edits.</p> <p id="4567">Inserted whole new paragraph.</p> <p id="1235">Second paragraph with some minor edits.</p> ====================================================================== If I have an external annotation which points to the content of the second paragraph in the First Edition: id="1235", then in the Second Edition, the link will remain unbroken even if a new paragraph was inserted before it *and* the content in that paragraph was revised but not enough to be topically different. PSWG discussed the issues of interpublication linking for over a month for the next generation Open eBook Publication Structure, where we wanted to enable robust interpublication linking, annotation, etc., into OEBPS Publications. There's a *lot* of subtle and not-so-subtle issues involved, some of which I've outlined in prior messages to gutvol-d and TeBC a while back. The original proposal to allow external annotation of digital texts (like PG texts) may seem like a new idea to many PGers here, but it's been something several of us have considered for quite a while (I was thinking of it back in 2000 for Yomu.) It's not new to me. I've even mentioned it here a few times, but not so explicitly (because we had not yet publicly announced LibraryCity.)
what you've done, though, is merely specified that there is markup which _cannot_ change (the "fragment identifiers"), so as to assure link-permanence. if i were to specify content that can not change, i can guarantee link-permanence as well.
Of course, one can come up with a scheme to link into plain text documents by character counting, paragraph counting, or a number of other methods (including Ted Nelson's Project Xanadu approach.) It is entirely possible someone has even come up with an IETF/RFC or something else covering this. Have you researched to see what others have already proposed for such a standard? (For Ted Nelson's Xanadu, refer to http://www.xanadu.com/ )
and in almost all cases, we're more likely to have text-invariance than to have markup-invariance. (but this is beside the point, since it's easy enough to specify invariance of text and markup. it is also very easy to show link breakage in cases of variance.)
Ah, but with XHTML 'id' (and now XML xmlid), it is possible to do significant text amendment and preserve existing links based on 'id'. Of course, as noted above link preservation can be achieved for plain text emendments, but it appears to be much messier, especially at the authoring level. But then someone smart and motivated (like Bowerbird) may come up with a clever way to make this work for plain text documents. Jon