I was asked to help to prepare a paper with Springer's svmult class. The conference (held in Germany) organizers refer to http://www.springer.com/authors/ for formatting instructions but these instructions lack one important detail: how to format citations and the bibliography listing.

The editor of the conference the paper is to be submitted insists on using author(s) name(s)/publication year (Harvard) system but bibliography listing conformed to Harvard system can be formatted in several pretty different ways.

I have found the following recommendations concerning references in manuscript-guidelines.pdf:

One author: (Miller 1991) or Miller (1991)
Two authors: (Miller and Smith 1994) or Miller and Smith (1994)
Three authors or more: (Miller et al. 1995) or Miller et al. (1995)

It is not explained (among other things) however how to format multiple citations, ie:

(Miller and Smith 1994, Jones 2002, Jordan et al. 2004) or
(Miller and Smith 1994; Jones 2002; Jordan et al. 2004)

or in some other way?

To figure out the required bibliography format in desperation I decided to run bibtex with spbasic.bst (included in svmult package) and then I have discovered a pretty lengthy comment (inside spbasic.bst) on using svmult and Harvard system. Well documented package :@. Neue Deutsche Wirtschaft?.

The conference organizers were asked if the above is OK. The lack of detailed manuscript guidelines is suspicious (I suspect that person responsible for conference proceedings is not a LaTeX expert) but as they say `OK' the problem seems to be solved.

To save potential users of svmult reinventing the wheel in cases as the one described above:-) here is the template of the manuscript:

\documentclass[natbib]{svmult} %% option natbib for author/year bibliography
\usepackage{mathptmx}   % selects Times Roman as basic font
\usepackage{helvet} % selects Helvetica as sans-serif font
%%
%% cf. manuscript-guidelines-1.0.pdf
%% ---------------------------------
%% Cite references in the text with author name/s and year of 
%% publication in parentheses ("Harvard system"):
%% One author: (Miller 1991) or Miller (1991)
%% Two authors: (Miller and Smith 1994) or Miller and Smith (1994)
%% Three authors or more: (Miller et al. 1995) or Miller et al. (1995)
%%
%% Use the command \bibpunct with one optional and 6 mandatory arguments: 
%% arg 5 -- the punctuation that comes between the author names and the year 
%%   empty argument denotes space between author-name publication-year
\bibpunct{(}{)}{,}{a}{}{,}
%%
\begin{document}
%% 
%% In English texts all words of a heading (\title, \section, \subsection etc) 
%% have a leading capital letter except for articles (a, an, the), 
%% conjunctions and prepositions of up to four letters (e.g. on, of, at, to, etc...)
%% (cf cf. authinst.pdf)
%%
\title*{DC of Data Streaming with~a~GNG Network} %
\author{Janina Filipowicz and Tadeusz Owsianko}
\institute{Janina Filipowicz \at
Casimir Pulaski University of Technology and Humanities in Radom, 
Higher School of Engineering,
\email{filipj@ut-radom.edu.pl}
\and
Tadeusz Owsianko \at
Universitas Technologiae et Humanistica Radomensis Casimirus Pulaski, 
Faculty of~Materials Science
\email{owsianko@ut-radom.edu.pl}}

\maketitle

\abstract{%%
The paper presents the results of simulation research
concerning the possibility of applying self-learning GNG neural
networks in clustering data from the data streams.
%% Keywords are optional
\keywords{cluster analysis, data streams clustering, GNGN.}
}

\section{Introduction}

One of the problems connected with the analysis of data [...]

\section{Related Work}

[...] They may be divided into four basic groups. The
first group is made of the incremental or online classifiers. These
are such algorithms as the Very Fast Decision Tree algorithm 
(VFDT)~\citep{domingoshulten2000}.
The second group is made of the multi model algorithms such as Ensemble
Classifiers (EC)~\citep{wangfan2003,kranenassent2011}. The
third group of algorithms contains the low granularity Rule Based
Classifier proposed by~\cite{jirayusakulauwatanamongkol2007}. 

In cluster analysis
their applications were studied among others by~\cite{kaufmanrousseeuw1990},
as well as by~\cite{fritzke1994}.

In the study presented below, the data grouping algorithm based on the
self---learning of neural network of the Growing Neural Gas (GNG)
type will be presented. Compared to the classical Fritzke 
algorithm~\citeyearpar{fritzke1994} 
it has been substantially modified. 

\section{Algorithm of a GNG Network}

In the study presented the data grouping algorithm based on the
self-learning of neural network of the Growing Neural Gas (GNG)
type will be presented. Compared to the classical Fritzke 
algorithm~\citeyearpar{fritzke1994} 
it has been substantially modified.

\section{An Experimental Study}

Summing up the results of experiment, it can be said, that the
compatibility of grouping with the known model was very high 
(cf.~table~\ref{t:quality-speed-clusters}). 

\begin{table}[!tbh]
\caption{Quality and speed of learning 
 of~the GNG network\label{t:quality-speed-clusters}}
%%\tabcolsep=.7\tabcolsep
\begin{tabular}{l|rrrrrrrrrrrrrr} \hline
Clusters	        & 2	& 3	& 4	& 5	& 6	& 15   \\ \hline
Mean RAS	        & 0.75	& 0.92	& 0.97	& 0.98	& 0.99	& 0.83 \\
Mean SC	                & 0.50	& 0.74	& 0.86	& 0.91	& 0.93	& 0.76 \\
Mean Time/Iter ($s^{-4}$)& 4.08	& 4.20	& 4.29	& 4.31	& 4.35	& 4.51 \\ \hline
\end{tabular}
\end{table}

[...] The time of single learning iteration of the network depends on the
number of neurons in that network [...]

\section{Concluding Remarks}

It is difficult to determine the threshold value at which the network
should switch from the dynamic into the static phase. The above
mentioned problems will be subject of further research.

\begin{thebibliography}{}

%% Conference proceedings etc:
\bibitem[Domingos and Hulten(2000)]{domingoshulten2000} 
  Domingos P, Hulten G (2000): Mining High-Speed Data Streams. In:
  Proceedings of the sixth ACM SIGKDD International Conference on
  Knowledge Discovery and Data Mining, Boston, pp~71--80.

%% Book 1 author
\bibitem[Fritzke(1994)]{fritzke1994} 
  Fritzke B (1994): Growing Cell
  Structures -- a~Self-Organizing Network for Unsupervised and
  Supervised Learning. Neural Networks 7:1441--1460.

%% Journal article
\bibitem[Jirayusakul and Auwatanamongkol(2007)]{jirayusakulauwatanamongkol2007} 
  Jirayusakul A, Auwatanamongkol S (2007): A Supervised Growing Neural Gas
  Algorithm for Cluster Analysis, International Journal of Hybrid
  Intelligent Systems 4(2):129--141.
%% Book 2 authors
\bibitem[Kaufman and Rousseeuw(1990)]{kaufmanrousseeuw1990} 
  Kaufman L, Rousseeuw PJ (1990) Finding Groups in Data: a~Introduction
  to Cluster Analysis. Wiley, New York.

%% Journal article more than 2 authors:
\bibitem[Kranen et~al.(2011)]{kranenassent2011} 
  Kranen P, Assent I, Baldauf C,  Seidl T (2011) The ClusTree: Indexing
  Micro-Clusters for Anytime Stream Mining. Knowledge and Information
  Systems 29(2):249--272.

%% Conf proceeding more than 3 authors:
\bibitem[Wang et~al.(2003)]{wangfan2003} 
  Wang H, Fan W, Yu PS, HAN J (2003) 
  Mining Concept-Drifting Data Streams Using Ensemble
  Classifiers. In Proceedings of the ninth ACM SIGKDD International
  Conference on Knowledge Discovery and Data Mining, NY, pp~226--235.

\end{thebibliography}
\end{document}

BibTeX is not used. Optional argument of \bibitem command should contains text which is used by natbib to format citations. This text should looks like: author-list(year).