===============================================================================
Date: 

Paper #: 

Title: 

Reviewer:

Reviewer Signature (please do not sign anywhere else):


===============================================================================

                COMPUTATIONAL LINGUISTICS

              Referee Report to the Editor

		Julia Hirschberg
		Editor, Computational Linguistics
		AT&T Bell Laboratories
		Room 2D-450
		600 Mountain Avenue
		Murray Hill, NJ 07974

		acl@research.att.com

		908-582-7496 (office)
		908-582-7308 (FAX)

	[When completed, please send this review by email OR by
	 surface mail to the  editor, Julia Hirschberg, at the
	 above email or regular mail address.  An electronic verions
	 of this review form is available by anonymous ftp from
	 cs.columbia.edu and is located in acl-l/Styfiles/clreview.]

________________________________________________________________________


			Guidelines for Reviewers

Here are some suggestions on the kind of information that will be
useful in making a decision on a manuscript: In addition to detailed
comments on the paper, a candid assessment of the paper's strengths
and weaknesses (in particular, its degree of originality) in such
areas as formalism, implementation, linguistic theory and data (where
appropriate) would be helpful.  Is the paper sufficiently novel and
interesting in at least one of these areas to warrant publication?
Does it present a significant new departure from existing theory and
practice, is the formalism it contains consistent, the implementation
sound, the linguistic data well chosen and insightfully analysed?

In addition to comments on originality, please also consider how well
presented the material is.  These comments can be particularly helpful
if authors are asked to revise and resubmit a manuscript.  Here are
some possible criteria:

well-structured:	clearly organized, has an initial overview,
			a final summary, structure is logical.
well-written:		good prose, reasonable sentence length,
                  	grammar at native-speaker level.
well-motivated:   	identifies important problem(s) to tackle,
                  	other analytical options are discussed/rejected.
focussed:         	clear statement of goals, goals carried out,
                  	results evaluated in terms of goals.
situated:         	notes relevant earlier work, relates
                  	own proposals to that work.
exemplified:      	formalism/impl given detailed applications,
                  	applications are intuitive and well-explained.
accessible:       	technical terminology explained on first use,
                  	self-contained, careful development of new concepts.
evaluated:	  	evaluation methodology is explained clearly, novel
		  	techniques are adequately explained and justified,
		  	results are described in sufficient detail.


			Summary Information for the Editor:

Briefly, what is this paper about?










What are its major strengths and weaknesses?








What do you think its major contribution to the field of computational
linguistics would be, if it were published?










			Summary evaluation for the Editor:

This paper:

1) [ ]  is acceptable more or less as it stands: [explain]


2) [ ]  is basically acceptable, but needs some changes before publication.

        The paper is:

	    [Delete items that do not apply (email) or check all that do.] 

        - - too long (list sections)
        - - too short (explain what is missing)
        - - badly expressed or organized
        - - lacking in references to or awareness of related work
        - - too vague or woolly
        - - other (explain)

3) [ ]  reports interesting work, but is not suitable for publication
        in its present form because it is:

	    [Delete all that do not apply (email) or check all that do.] 

        - - a premature progress report
        - - too vague and imprecise
        - - not sufficiently general
	- - not sufficiently original
        - - other (explain)

    Should the author be encouraged to revise and resubmit?
         [ ] yes     [ ] no

4) [ ]  is unsuitable for publication, and should be rejected.
        The work it reports is:

	    [Delete all that do not apply (email) or check all that do.] 

        - - not suitable for CL (a more relevant journal would be: ...)
	- - already published in another journal or book 
		(publication details: ...)
        - - incompetent or naive
        - - other (explain: ...)



Additional comments for the editor (only):
----------------------------------------

	[Please attach additional comments for the editor here. (The
	form and these comments will not be sent to the author, but
	parts may be summarized.)  Please send comments to be sent
	to the author in the section provided below.  Please do *not*
	give any indication of your identity below this line.]




===============================================================================

____________o/____________o/____________o/____________o/____________o/_________
            0\            0\            0\            0\            0\         

===============================================================================


Comments to be sent to the author:
---------------------------------

	[If you are sending this in email, these comments should be
	 ascii text without formatting commands (e.g., no latex source
	 please.]

Paper #:

Title: 




===============================================================================