The page limit is
Please do not be confused by the page numbers in the preproceedings. Regardless of the prior size of your article, you can extend you contribution within the limits.
If you follow the following instructions you willget a result where the PDF contains also the table of contents and the index of authors. They are not counted in to the total number of pages in the contribution.
Because Springer will use Computer Modern instead of Times Roman font family that we unfortunately advised to use earlier, the produced text may grow slightly when you follow the instructions below. We apologize for our wrong information.
These instructions are based on the following sources:
You should send the material through email as an attachment document. Send the package to the address "fsmnlp(a)ling.helsinki.fi" and use as the Subject: "FSMNLP: PROCNNNN: submit" where NNNN is your paper ID. For example, the headers of my submission email looks like this:
To: fsmnlp(a)ling.helsinki.fi
Subject: FSMNLP: PROC9257: submit
ID code | Paper title | Copyright Form | Slides Cached |
---|---|---|---|
9001 | Characterizations of Regularity | OK | |
9002 | Finnish Optimality-Theoretic Prosody | ||
9003 | Partitioning Multitape Transducers | OK | YES |
9013 | Squeezing the Infinite to Finite | OK | |
9024 | A Novel Approach to Computer-Assisted Translation based on Finite-State Transducers | OK | |
9034 | Finite-State Registered Automata and Their Uses in Natural Languages | OK | YES |
9046 | TAGH: A Comlete Morphology for German based on Weighted Finite-State Automata | OK | YES |
9058 | Klex: A Finite-State Transducer Lexicon of Korean | OK | |
9069 | Longest-Match Pattern Matching with Weighted Finite-State Automata | OK | YES |
9077 | Finite-State Syllabification | OK | |
9087 | Algorithms for Minimum Risk Chunking | OK | |
9099 | Collapsing $\epsilon$-Loops in Weighted Finite-State Machines | ? | |
9109 | WFSM Auto-intersection and Join Algorithms | OK | YES |
9121 | Further Results on Syntactic Ambiguity of Internal Contextual Grammars | YES | |
9133 | Error-Driven Learning with Bracketed Constraints | OK | |
9145 | Parsing with Lexicalized Probabilistic Recursive Transition Networks | OK | YES |
9156 | Integrating a POS Tagger and a Chunker Implemented as Weighted Finite State Machines | OK | YES |
9168 | Modelling the Semantics of Calendar Expressions as Extended Regular Expressions | OK | |
9180 | Using Finite State Technology in a Tool for Linguistic Exploration | ||
9191 | Applying a Finite Automata Acquisition Algorithm to Nambed Entity Recognition | OK | |
9202 | Principles, Implementation Strategies and Evaluation of a Corpus Query System | OK | |
9214 | On Compact Storage Models for Gazetteers | OK | YES |
9226 | German Compound Analysis with \textit{wfsc} | OK | YES |
9234 | Scaling an Irish FST Morphology Engine for Use on Unrestricted Text | OK | OKYES |
9246 | Improving Inter-letter Communication in Cascaded Finite-State Partial Parsers | ||
9257 | Pivotal Synchronization Expressions: A Formalism for Alignments in Regular String Relations | ? | YES |
9269 | A Complete FS Model for Amharic Morphographemics | ||
9271 | Tagging with Delayed Disambiguation | OK | |
9273 | A New Algorithm for Unsupervised Induction of Concatenative Morphology | OK | |
9275 | Morphological Parsing of Tone: an Experiment with Two-Level Morphology on the Ha Language | ||
9277 | Describing Verbs in Disjoining Writing Systems | ||
9280 | An FST Grammar for Verb Chain Transfer in a Spanish-Basque MT System | OK | |
9282 | Finite-State Transducers Based on k-TSS Grammar for Speech Translation | OK | |
9285 | Unsupervised Morphology Induction Using Morfessor | OK | |
9286 | SProUT - a General-Purpose NLP Framework Integrating Finite-State and Unification-Based Grammar Formalisms | OK | YES |
9288 | Tool Demonstration: Functional Morphology | OK | YES |
9290 | From Xerox to ASpell: A First Prototype of a North Sámi Speller Based on TWOL Technology | OK | |
9292 | A Programming Language for Finite State Transducers | OK | YES |
9294 | FIRE Station |
These IDs are temporary and only to help us. When we deliver the material to Springer, each article will be given a new number in the range 0001 - 0350, reflecting the number of the first page of each article.
For purposes of packaging, make a directory whose name is PROCNNNN where NNNN = your paper ID code and move the source files of your article to this directory.
Send us the above directory as a tar or zip archive. When you compress the files, please do archiving so that the parent directory of the directory PROCNNNN is the root of the archive and the archive is named as PROCNNNN.tar.gz or PROCNNNN.zip, where NNNN is your paper ID code. We recommend the tar utility. With it you can create the archive with the commands
$ cd PROCNNNN/ $ cd .. $ tar cvfz PROCNNNN.tar.gz PROCNNNN/
For example, I would personally do it as follows
$ cd PROC9257/ $ cd .. $ tar cvfz PROC9257.tar.gz PROC9257/
The submission should include the following standard parts that should be present in the PROCNNNN/ directory:
PROCNNNN/Makefile | a Unix-style makefile (if you are using Windows, that's OK and then you do not have this) |
PROCNNNN/fsmnlp.tex | the main file of your paper contains just a few lines that load the style definitions and the two parts of the article (the abstract and the body) |
PROCNNNN/lib | the directory containg all the non-standard LaTeX style packages |
These files can be downloaded as a complete tar-archive from http://www.ling.helsinki.fi/events/FSMNLP2005/tex/fsmnlp-lib.tgz. Directory /tex/lib/ contains two versions of fsmnlp.tex that can be used if the tipe (phonetic IPA alphabet) or Hangul (Korean alphabet) are in use: fsmnlp-tipa.tex and fsmnlp-hangul.tex.
In addition to the standard parts, the submission should include the following parts that are unique to each contribution. These should be present in the PROCNNNN/ directory of the submitted package.
PROCNNNN/fsmnlp-abst.tex | the file containing the author and title definitions as well as the abstract |
PROCNNNN/fsmnlp-body.tex | the file containing the body of the article (the "body" starts after the abstract usually with the command \section{Introduction}) |
PROCNNNN/inc | the directory containing ALL the files that are not figures but are included in the body of the document |
PROCNNNN/fsmnlp-bibl.tex | the file containing either a \bibliography{...} or \thebibliography{99} command. We recommend using the first method (remember to include the .BIB-files). |
PROCNNNN/*.bib | the BIBTEX sources mentioned in fsmnlp-bibl.tex. |
PROCNNNN/fsmnlp-appx.tex | the file containing an optional appendix (will be printed after the bibliography) -- in most cases this file is empty. |
PROCNNNN/img | the directory containing ALL the EPS-figures as well as the source files of (gpic,) dot and Xfig figures, if any (this allows reproducing the corresponding EPS files if necessary to improve their quality) |
PROCNNNN/*.bib | the BIBTEX sources mentioned in fsmnlp-bibl.tex. |
PROCNNNN/fsmnlp.bbl | the references file generated by the bibtex program. |
PROCNNNN/fsmnlp.dvi | the DVI version of your contribution (PostScript figues need not be properly shown by xdvi — this is just for rebuilding the PostScript file, if needed.) |
PROCNNNN/fsmnlp.ps | the PS version of your contribution |
PROCNNNN/fsmnlp.pdf | the PDF version of your contribution |
PROCNNNN/fsmnlp.log | the log of the latest latex run. |
If you do not want to create the above directory structure from scratch or you do not understand what you should do, you can download a package that contains the required directory for your paper. This directory was created when we built your article for the pre-proceedings CD (for purposes of the official proceedings we changed the document style, though, so it can be used immediately as a basis for the revised version).
The secret download URL has been send to the authors by email.
Please download our version, extract it with "tar xfvz PRECNNNN.tgz" and try building the PDF file with make.
When you say simply make this assumes that bibliography is static and contained in fsmnlp-bibl.tex. If you can provide the BIBTEX sources, please insert \bibliography{...} command to fsmnlp-bibl.tex and say make all.
Figure captions should be under the figures and table captions above the tables!
There is no period at the end of the caption!
Headings (at least \title,\section,\subsection) should be Capitalized (i.e. Nouns, Verbs and all other Words except articles, prepositions and conjunctions). There is a special rule for words joinded with hyphen:
Non-deterministic, Self-determinations, Multi-flip,
but
Finite-State, User-Friendly, Context-Freeness, etc.
You must use \label and \ref or \pageref rather than absolute numbers for cross-references.
There should be no overfull boxes in the document. Please check the .log file created by LaTeX for error messages of the following type:
Do not use psfig or epsf for inclusion of EPS figures. Please use the epsfig package or/and the graphicx package instead. The epsfig is a uses the interface of psfig and techniques of epsf and makes them thus obsolete.
Let the Makefile to make the bibliography with bibtex each time when you compile the PDF file. This implies that
You do not write citation references directly into the text but you use the \cite command for references to the bibliography.
You can provide us with the Bibtex source of the bibliography in order to allow us to modify the bibtex style and to create hyperlinks to these places.
If necessary, you can insert a pagebreak before a figure, but normally please define figures as floating objects if possible. The reason is that the text may condense expand if the publisher uses different fonts than the authors.
In order to prevent hyphenation in the table of contents, use \protect\newline (instead of \\).
Do not try to expand the font capability of math mode with cmbright, sfmath or \DeclareMathAlphabet. The methods like these do not work properly as the maximum number of available math alphabets in TeX is amazingly only sixteen (this is a known LaTeX issue). Moreover, additional fonts have to be taken care of when PDF is generated, otherwise they are not safe in general. For example, the following is forbidden:
\DeclareMathAlphabet{\mymathsfsl}{T1}{lmss}{m}{sl}
\DeclareMathAlphabet{\mymathbfn}{T1}{lmr}{b}{n}
\newcommand{\myalpha}[1]{\ensuremath{\mymathsfsl{#1}}}
\newcommand{\myopname}[1]{\ensuremath{\mymathbfn{#1}}}
A convenient and very generic way to add more fonts to the math mode is to embed a \text{}-box inside the math mode. In the box you can use all non-math fonts. For example, this method is good for using slanted sans-serif in the math mode.
\newcommand{\mymathsfsl}[1]{\ensuremath{\text{\textsf{\textsl{#1}}}}}\newcommand{\myopname}[1]{\ensuremath{\text{\textbf{#1}}}}
The llncs document class uses automatically Computer Modern Roman (CM) fonts. Other fonts for Roman, Sans-Serif and Courier font families would deviate from the LNCS style and their use is therefore discouraged (see LNCS Authors' Instructions, page 7). We apologize the incorrect information in earlier phases when we asked to use the times package. Thus, the forbidden text font packages should include at least
courier, helvet, mathpazo, mathptmx, psfonts, pslatex, times.
Package inputenc Warning: No characters defined (inputenc) by input encoding change to `ascii' (inputenc) on input line 1.
If the default causes the compilation to fail at a certain file, this file contains a non-ASCII character and uses a different encoding. Determine then which encoding it is using. Then, change the default parameter value ascii in the above command into latin1 (papers 9145,9156,9191 ?), utf8 (papers 9003,9290 ?), cp850 (paper 9058) or ansinew (paper 9069) or something else according to the encoding you are using in each file.
However, if you make use of Hangul symbols in you text, then you should not set any input encoding. Hangulj.sty defines its own input methods and it assumes that the input encoding is unset.
If you use the tipa package to generate IPA fonts, please note that you need to copy lib/fsmnlp-tipa.tex onto fsmnlp.tex. Furthermore, you should observe that we reserve the following short commands for their original use and do not allow their redefinition by tipa.sty: \:, \;, \!, \*, \|. This is restriction is achieved with the \tipasafemode command after fsmnlp-tipa.tex have been declared.
Table of contents and the author index of the proceedings can be compiled automatically if we can compile all the contributions together as a single LaTeX document. We know how to do this, but your co-operation can reduce our work. However, Springer will compile each contribution separately and they do not require these things.
If you define commands etc. try to choose distinctive names for them (you can add, for example, you initials to the command names to make them more probably unique). Please refrain from defining LaTeX commands with very short names. The same applies to labels.
Note that you can prevent visibility of your definitions outside of a certain area by including the definitions of new commands inside a pair of curly brackets { ... }. Such brackets are, in fact, already employed in the fsmnlp.tex. In contrast to this, you cannot be completely sure that a certain command name has not been defined in a package that somebody else is using. Therefore, short names should be avoided.
In the current contributions, there are at least two used packages that define active characters (very short command names): gb4e.sty and hangulj.sty, but they are acceptable. Problematic packages such as these can be used in fsmnlp-body as follows:
Here the role of \makeatletter and \makeatother is to allow more complex command names, typical for style packages internals. Unfortunately, this method does not allow inclusion of many latex styles, as they have to be included in the preamble.{ \makeatletter % allows @ in commands \input{lib/hangulj.more} \input{lib/gb4e.sty} \makeatother % does not allow @ in commands \section{Introduction} The body of your contribution.... }
LaTeX internal commands must not be redefined, i.e.
All new LaTeX commands defined in the paper should be used somewhere. Please remove the definitions of all unused commands.
Check that the EPS figures are of high quality. It is true that screenshots and camera pictures are pixel graphics and have limited resolution. However, most of the illustrations in the articles are drawings. Drawing can be made with a tool that uses vector graphics, and we encourage you to ensure that the way you produce EPS does not reduce the quality by converting vector graphics into pixel graphics.
To check the quality of your drawings, open the final PDF file with AcroReader. Then, manify the page on the screen to some 3200% of the original size. If your drawings look perfect even at this resolution you can be faily sure that they make use of the vector graphics capability of the PostScript language. In some cases, you may have to re-draw the figures with a better tool, although the effort spent on producing perfect pictures should be kept within reasonable limits.
We recommend the following tools for the best drawing quality:
It is recommended that the author's include the sources of every drawing as well as the instructions (e.g. a Makefile target) into the submitted package of files. The sources should be stored to the PROCNNNN/img-src directory.
Remember that in the printed proceedings, the colors will not be implemented. Colors are, however, allowed in figures. Minimize shading as contrasts between various gray scale values may be difficult to reproduce.
Before you send the package to us, we ask you to check that you have verified the following items:
$ cd PROCNNNN/
$ make clean
$ make
$ acroread fsmnlp.pdf
Check that the PDF document looks what you want.
Check that the produced fonts in text and in mathematic formalas are of high resolution in the PDF. Sometimes your system does not support exactly the fonts you are using TeX substitutes them with something else. In such cases, the pagination and line breaks of your PDF may differ from the PDF procuded by the publisher.
Check that the page count is within the limit of 12 pages. You are not allowed to use direct formatting commands to squeeze the pages. Rather, if your article is too long, you should remove some material.
Read "Authors'Instructions for the preparation of camera-ready contributions to LNCS/LNAI Proceedings" in case that there are some relevant points that we have not mentioned here.
All the files in the directory PROCNNNN should be related to the official version of this paper. Please remove files related to older versions or to other articles, but keep the directory functional (test the cleaned directory with make).
Please do not confuse us by sending more than one PDF file at a time, unless you resend everything. We assume that the PDF file is built from the sources in the same directory.
A signed copyright form is required for each article. In the table of articles above, we have marked with "OK" all those papers for which we have already received a signed copyright form.
You can download (PDF, 10Kb) the copyright forms from Springer and print it. Note that Lecture Notes in Artificial Intelligence (LNAI) is a derivative of the concept of Lecture Notes in Computer Science (LNCS). Therefore the form is imprecisely entitled "Lecture Notes in Computer Science Copyright Form", but do not worry about this.
One of the authors signs for and accepts responsibility for releasing the material on behalf of any and all co-authors. Please make sure that the signed forms reach us by 10th December. You can use one of the following methods:
Dr. Anssi Yli-Jyrä
Department of General Linguistics
P.O. Box 9
FIN-00014 UNIVERSITY OF HELSINKI
FINLAND, EUROPE
Question: What we are supposed to fill inthe blanks related to
the "Conference/Book" and "Volume Editor" that are required in the
copyright form?
Answer: You can use the name of the conference "FSMNLP
2005" or alternatively the tentative book name "Proceedings of the
FSMNLP 2005". The volume editors include Lauri Karttunen, Juhani
Karhumäki, and Anssi Yli-Jyrä (the final order does not matter in the
Copyright form).
We are going to send a printed copy of the official proceedings in the Springer LNAI to all its authors who attended the conference.
Please inform us on the postal address of the receiver(s) of the printed proceedings. We realize that the OpenConf server provides only very incomplete information and the registeration of conference participants did not require this information.
If you want to provide supplementary material (e.g. presentation slides or demonstration), we will add links to them. Springer is also interested to add links to the supplementary material (you will still have the copyright).
There are two methods for providing the link:
You should also provide the material and a short description of it. The slides in PDF or PPT format might still exist on our conference laptop (denoted with "YES" in the "Slides Cached" column in the prior table of articles), if you want us to collect them for this purpose.
Question: By the way, are the pictures that you took during the conference
available on-line? Thank you.
Answer: They are on the way to web, but it will still take
time, unfortunately, to prepare a selection of those. But we will
tell all the participants when we are ready.