Actually, our team has produced "ScolastiX", an application very close to the
thing you describe here. If you can read French, I wrote a brief article about
it, a couple of years ago:
The basic aim of the project was to help a team of experts (scattered
through Europe) to annotate a corpus of texts (in a structured manner,
i.e. with predefined patterns), without having to make the scholars go
through a demanding learning process. With ScolastiX scholars can
annotate a text according to various predefined "approaches", through
very simple forms (as every Internet user is now capable of filling in a
It works roughly like this:
- when a text is added to the corpus (plain text format, possibly
including tags for italics), it is transformed into an internal XML
format, very basic: each word and each sentence is tagged with "w" and "s"
elements respectively, and each of these elements receives a unique identifier
- the scholars working on the corpus see a text on the left side of the
screen (basic internal XML presented through a basic XSL), and a form on
the right side (for our project for example, we have forms to add the
scheme of a text, the /exempla/, the /distinctiones/ and other
rhetorical figures, etc.). To add an annotation, they fill in the form,
and attach the annotation to the text by clicking on the beginning and
ending word or sentence of the concerned span of text (the unique
identifiers of these elements are used here). The annotations are then
stored in a database (we use MySQL).
- when the work is done, it is possible to export the annotated text in
any DTD / Schema: different export rules enable the administrator to
create different XML documents from the same source, merginf the annotations and
the text. Basically, when a text is exported, a kind of parser examines each
element in the basic internal XML file, and if an annotation has been added
with, as beginning or ending the unique identifier of this element, then it
triggers a procedure that differs according to the export rules (which tags must
be added for this type of annotation, where, which attributes, etc.)
Our team has used ScolastiX for some years now, for our ongoing work on the
sermons of Jacques de Voragine, and it works well so far! :) Yet, there hasn't
been much development on the software itself in the last two years: the latest
currently downloadable version (http://www.sermones.net/scolastix/) is
unfortunately not compatible with the latest versions of PHP, so it would mainly
need to be tidied up a bit (our team uses a previous version). But there might
be some news soon as another team might use ScolastiX for its own annotation
projects and has proposed some help.
I'd be very happy to discuss it further with all of you interested!
Selon Tim Finney <[log in to unmask]>:
> If you want others to be able to write graffiti on a base text then you
> need to add referencing hooks to it. If the base text is fluid (i.e. has
> a variant textual tradition) then that adds another layer of complexity.
> Perhaps a good starting point would be to mark up every word of an
> authoritative base text with a <w id="someID"> element.
> Once the referencing hooks are there then you can add stand off markup.
> It would be a good thing to have a tool that does this: one person could
> add graffiti and another could say "Show me all the comments on this
> phrase by person XYZ". Actually, the tool could be responsible for
> adding the hooks to the selected base text. Just thinking aloud.
> Tim Finney
> On Sun, 2008-08-17 at 20:10 +0200, Peter Boot wrote:
> > Eric Lease Morgan schreef:
> > > Moreover, what if some sort of tool, widget, or system were created
> > > that allowed anybody to add commentary to texts in the form of TEI
> > > mark-up. Do you think this would be feasible? Useful?
> > This is certainly interesting. It is something that many people have
> > experimented with, but I don't think there are at present really
> > satisfactory tools for doing something like this. One of the things
> > you'll have to decide on is whether to store your annotations in the
> > text, somewhere in the back of your document, or in external documents.
> > Another issue is how you want to present the text to be annotated to the
> > user: in raw XML, the user may have trouble orienting himself, so you
> > may want to apply a (css? xslt?) stylesheet (that should be sufficiently
> > generic to cover most of the annotated files). But this requires your
> > widget to interact with a browser or other application that knows how to
> > render xml using a stylesheet. Another question is whether potential
> > users will trust your system sufficiently to save their work
> > (annotations represent work) in the system.
> > Any experimentation in this area is useful, I believe.
> > Peter