On Fri, Nov 19, 2010 at 7:30 AM, FORT, Karen <span dir="ltr"><<a href="mailto:Karen.FORT@inist.fr" target="_blank">Karen.FORT@inist.fr</a>></span> wrote:<br>
<blockquote style="margin: 0pt 0pt 0pt 0.8ex; border-left: 1px solid rgb(204, 204, 204); padding-left: 1ex;" class="gmail_quote">
As for NLP as such, the question seems more difficult to answer, but for
now, I'd tend to think, like you do, that Bird & Liberman were the
first to introduce it in the field, don't you think so?<br></blockquote><br>It certainly predates the Bird & Liberman reference around 1999/2000 that Piotr mentions. <br><br>Although this was a bit before my time, I believe that the term "standoff markup" or "standoff annotation" was introduced by the DARPA TIPSTER program that started in 1991. This scheme stores the document in one file and various annotation layers in other files. These annotations may be produced manually or automatically.<br>
<br>I am not sure who was responsible or the general idea or the specific encodings used in TIPSTER -- <a href="http://www.fas.org/irp/program/process/tipster.htm" target="_blank">http://www.fas.org/irp/program/process/tipster.htm</a> suggests that standardization began in 1994, but the ideas may have been in use in individual TIPSTER projects before that. <br>
<br>Hamish Cunningham and the folks behind the GATE architecture may know more; GATE implements the DARPA scheme, I believe.<br><br>Thompson & McKelvie (1997) is well-cited for explaining how to encode standoff markup in SGML or XML: <a href="http://www.ltg.ed.ac.uk/%7Eht/sgmleu97.html" target="_blank">http://www.ltg.ed.ac.uk/~ht/sgmleu97.html</a><br>
<br>cheers, jason<br><br><br>