{
    "paper_id": "W98-0118",
    "header": {
        "generated_with": "S2ORC 1.0.0",
        "date_generated": "2023-01-19T06:04:24.060127Z"
    },
    "title": "A Standar~ Representation Framework for TAG",
    "authors": [
        {
            "first": "Issac",
            "middle": [
                "L I P N"
            ],
            "last": "Fabrice",
            "suffix": "",
            "affiliation": {},
            "email": "fabrice.issac@lipn.univ-parisl3.fr"
        },
        {
            "first": "Institut",
            "middle": [],
            "last": "Galilee",
            "suffix": "",
            "affiliation": {},
            "email": ""
        },
        {
            "first": "Av",
            "middle": [
                "J E"
            ],
            "last": "Clement",
            "suffix": "",
            "affiliation": {},
            "email": ""
        }
    ],
    "year": "",
    "venue": null,
    "identifiers": {},
    "abstract": "We present in this paper a markup language suitable for representing a tree adjoining grammar. Using a uniform way to represent TAG, the development of tools, e.g. parser/recognizer, editor, ... , could be clone to the benefit of the entire TAG community.",
    "pdf_parse": {
        "paper_id": "W98-0118",
        "_pdf_hash": "",
        "abstract": [
            {
                "text": "We present in this paper a markup language suitable for representing a tree adjoining grammar. Using a uniform way to represent TAG, the development of tools, e.g. parser/recognizer, editor, ... , could be clone to the benefit of the entire TAG community.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Abstract",
                "sec_num": null
            }
        ],
        "body_text": [
            {
                "text": "Our work consists of proposing a framework dedicated to designing an XTAG-like standard environment. We present in this paper a markup language suitable for representing a tree adjoining grammar. The TAG formalism is used in plenty of works all around the world. However it is difficult to exchange syntactic data ( i. e. a grammar or a piece of grammar written in the TAG formalism) as well as computer tools based on these eyntactic data. Ueing a uniform way to represent TAG, the development of tools, e.g. parser/recognizer, editor, ... , could be clone to the benefit of the entire TAG community. Note that all kinds ofTAG (LTAG, MCTAG, ... ) can be represented in our language. We choose SGML as descriptive language, as seen as briefly in the first section. in the aecond section, we provide an overview of the structure of a TAG document, followed in the third section by an example.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            },
            {
                "text": "We describe our language in SGML(Goldfarb90; Herwijnen95) (Standard Generalized Markup Language). SGML is itself a metalanguage. SGML is an efficient tool to describe classes of documents because (i) it is an ISO specification 1 , thus, a standard 1 1so 8879:1986. (ii) a lot of tools can be used to edit, verify or exploit SGML based documents 2 \u2022 SGML is a meta-Janguage which allows specification through a Document Type Definition (DTD) :",
                "cite_spans": [
                    {
                        "start": 254,
                        "end": 264,
                        "text": "8879:1986.",
                        "ref_id": null
                    }
                ],
                "ref_spans": [],
                "eq_spans": [],
                "section": "SGML",
                "sec_num": "1"
            },
            {
                "text": "\u2022 a set of markups;",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "SGML",
                "sec_num": "1"
            },
            {
                "text": "\u2022 how these markups can be combined.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "SGML",
                "sec_num": "1"
            },
            {
                "text": "In our case, the dass of documents is the set of TAG grammars. The most popular DTD is HTML which is used as a norm for data representation on the Web but there are other projects, notably the TEI project. The Text Encoding Initiative (TEI) (SMB94) is an international project to develop guidelines for the preparation and interchange of electronic texts.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "SGML",
                "sec_num": "1"
            },
            {
                "text": "The TEI proposes recommendations for feature structure markup (LS95) which can be used to represent any feature structure, including TAG. However, we think the markup set defined is not specific enough to be easily treated.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "SGML",
                "sec_num": "1"
            },
            {
                "text": "First a good TAG document is preceded by a prologue which indicates the TAG DTD version:",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Structure of a TAG document",
                "sec_num": "2"
            },
            {
                "text": "< !DOCTYPE DTO PUBLIC \"DTD TAG 0. 2\">",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Structure of a TAG document",
                "sec_num": "2"
            },
            {
                "text": "The whole document is enclosed by the <tag> and </tag> markup. lt is composed of a header and a bod!/, The header, enclosed by the <tagheader> and </tagheader> markup, contains information about _ the document itself: title, date of creation, name of the creator, origin of the data, type of data.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Structure of a TAG document",
                "sec_num": "2"
            },
            {
                "text": "The body, enclosed by the <ts> (Tree Set) and </ts> markup, forms the usable part of the document. A tree set is a list of tree families ( <tf>), elementary trees (<et>) or parsed trees ( <pt>). A tree contains only one node ( a node is indicate by <n>): the root node. The node markup can then be used recursively (a <n> can contain a <n>) to define the tree. A node contains some markups: <val> (value) is the tag of the node (i.e. category for elementary and derived trees or tree name for a derived tree); <fs> (feature structure) for FB-TAG.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Structure of a TAG document",
                "sec_num": "2"
            },
            {
                "text": "The tree of the figure 1 indicates the relationships between markups 4 .",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "Structure of a TAG document",
                "sec_num": "2"
            },
            {
                "text": "We give below a simple example. Let us suppose we have a tree with the features associated to the nodes : ",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "An example",
                "sec_num": "3"
            },
            {
                "text": "V",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "An example",
                "sec_num": "3"
            },
            {
                "text": "r l_{tltlo t 1 f_{tPCDATA} 1 !__dato t 1 UtPCPATl) 1 , __ cnator a 1 UtPCDAU} 1 l__orig\u2022 a ( L(tPCDATl) 1 l __ tJpO?) LCIPCDATll __ t&+) l_{tf 1 Ut+) __ et 1 (_(n) l_(nl, 1 L (IRCDATA) 1 , __ fa?. 1 l_(f+) 1 LO 1 I f ( _ (BNPTY} 1 1 I LIRCD1Tl) 1 l __ n+} ( 1-(U) 1 1 __ pt)+ L{atring, 1 UtPCDAU} 1 !__tt, \u2022\u2022\u2022 (_.dU} l.(n) \u2022 \"",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "An example",
                "sec_num": "3"
            },
            {
                "text": "Finally we will mention the way we can use such a description. lt is obvious that an SGML document, as this one, is not supposed to be directly understandable/readable to humans. lt is, in fad, used as input/output to computer tools. For instance, if developers follow these guidelines, the three steps of a parser -grammar generation, tree elimination, parsing -could be built by three different people.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "conclusion",
                "sec_num": "4"
            },
            {
                "text": "The final environment will contain ;",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "conclusion",
                "sec_num": "4"
            },
            {
                "text": "\u2022 a graphical TAG editor, in order to create or to modify TAG grammars;",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "conclusion",
                "sec_num": "4"
            },
            {
                "text": "\u2022 tools for parsing (generation, disambiguation, parsing);",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "conclusion",
                "sec_num": "4"
            },
            {
                "text": "\u2022 miscellaneous tools (for instance a OOE;X or HTML transduction of TAG trees).",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "conclusion",
                "sec_num": "4"
            },
            {
                "text": "The aim of this paper is not to give a final (nor complete) version of a language providing descriptions of TAG, but rather ad as a starting point. Actually, I think a data interchange norm can't be established by only one person. Tbat's why 1 wish the community take a part in the development of this norm.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "conclusion",
                "sec_num": "4"
            },
            {
                "text": "~For instance the sgllllql tool can extract part of document in relation to a query on the tags.3 As HTML, in fact it is a classical way to describe a SGML docurnent.",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "",
                "sec_num": null
            }
        ],
        "back_matter": [
            {
                "text": "Kllail : fabric\u2022.iaaacClipn.univ-paria13.fr ",
                "cite_spans": [],
                "ref_spans": [],
                "eq_spans": [],
                "section": "annex",
                "sec_num": null
            }
        ],
        "bib_entries": {
            "BIBREF0": {
                "ref_id": "b0",
                "title": "The SGML handbook. Oxford",
                "authors": [
                    {
                        "first": "_(",
                        "middle": [],
                        "last": "Goldfarb",
                        "suffix": ""
                    },
                    {
                        "first": "F",
                        "middle": [],
                        "last": "Charles",
                        "suffix": ""
                    }
                ],
                "year": 1990,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Goldfarb _(Charles F.). -The SGML handbook. Oxford, Clarendon Press, 1990.",
                "links": null
            },
            "BIBREF1": {
                "ref_id": "b1",
                "title": "Eric Van). -SGML pratique",
                "authors": [
                    {
                        "first": "",
                        "middle": [],
                        "last": "Herwijnen",
                        "suffix": ""
                    }
                ],
                "year": 1995,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Herwijnen (Eric Van). -SGML pratique. -Paris, International Thomson Publishing France, 1995.",
                "links": null
            },
            "BIBREF2": {
                "ref_id": "b2",
                "title": "). -A rationale for the tei recommendations for featurestructure markup",
                "authors": [],
                "year": 1995,
                "venue": "Computers and the humanities",
                "volume": "",
                "issue": "",
                "pages": "191--209",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Langendoen (D. Terence) et Simons (Gary F.). -A rationale for the tei recommendations for feature- structure markup . . i\": Computers and the hu- manities, pp. 191-209. -I<luwer Academic Pub- lishers, 1995.",
                "links": null
            },
            "BIBREF3": {
                "ref_id": "b3",
                "title": "et Burnard (Lou). -Guidelines for Electronic Text Encoding and Interchange {TE! PS). -Text Encoding Initiative",
                "authors": [
                    {
                        "first": "",
                        "middle": [],
                        "last": "Sperberg-Mcqueen",
                        "suffix": ""
                    }
                ],
                "year": 1994,
                "venue": "",
                "volume": "",
                "issue": "",
                "pages": "",
                "other_ids": {},
                "num": null,
                "urls": [],
                "raw_text": "Sperberg-McQueen (C.M.) et Burnard (Lou). - Guidelines for Electronic Text Encoding and In- terchange {TE! PS). -Text Encoding Initiative, 1994.",
                "links": null
            }
        },
        "ref_entries": {
            "FIGREF0": {
                "type_str": "figure",
                "num": null,
                "text": "anchor NP_O.t:<nU111>=VP.t:<nu.m> HP_O:<pers>aVP.t:<pers> S.b:<mode>=VP.t:<mode> VP.b:<mode>=V.b:<mode> VP.b:<nU111>=V.b:<num> VP.b:<pers>=V.b:<pers> NP_O.t:<vh>=-S.b:<iny>a-The SGML result is the following: <!DOCTYPE DTD PUBLIC \"DTD TAG 0.2> <tag lang=french> ~Note that thls tree is automaticaly generated with a",
                "uris": null
            },
            "TABREF0": {
                "text": "<tf> (tree family) encloses a tree family. The attribute (narne) indicates the name of the family. A tree family is composed of a list of trees (markup <t>).<et> (elementary trees) encloses elementary trees. The attribute (name) indicates the name of the tree. As a family tree, elementary trees are composed of a list of trees (possibly one).",
                "num": null,
                "content": "<table><tr><td>&lt;pt&gt; (parsed tree) encloses a parsed tree ( i. e. a de-</td></tr><tr><td>rived tree). Three markups are used to describe</td></tr><tr><td>(i) the string recognized by the tree ( &lt;string&gt;)</td></tr><tr><td>(ii) the tree itself ( &lt;t&gt;) and a set of derivation</td></tr><tr><td>trees (DT).</td></tr></table>",
                "type_str": "table",
                "html": null
            }
        }
    }
}