Dr. Bernoulli gets a job: Mathematics of the Job Search – Faculty Version

I recently found out that the Duke Computer Science department had 404 applicants for the open position in their department. I mentioned that to a CS professor from a different university, and he didn't seem surprised by that number. Moreover, when you think about how many "faculty candidate" lectures there usually are within a CS-like department each hiring season, and you consider that those interviewees are likely a small selection of the total applicants, then 404 starts sounding reasonable.

When there are 404 applicants who each have PhD degrees, publications, and possible post-doctoral or existing faculty appointments, let's also assume that the objective function that each department is maximizing is pretty flat. If you don't like that assumption, then assume we have no prior information, and so we will maximize entropy and assume that each applicant has a 1/404 chance of being picked for the job (in reality, this probability is itself conditioned on whether the state steps in and has a hiring freeze... so the real probability might be closer to 1/1000). So that is a very low number. Can we fight low probability with high volume of applications?

Assume we apply to N schools where the probability of getting an offer is

p = 1/404

at each of them. Then the probability of not getting an offer from each of them is

1 - p = 403/404,

and so the probability of not getting an offer from all of them is

(1 - p)^N = (403/404)^N.

So finally we arrive at the probability of getting an offer from at least one of them, which is

1 - (1 - p)^N = 1 - (403/404)^N.

Hypothetically speaking, let's say you apply to N = 50 such positions. Then you have a

1 - (403/404)^N = 1 - (403/404)⁵⁰ ≈ 11.65%

probability of getting the offer. Of course, if you were paying attention, you remember that p (1/404) is very small in this example. Consequently, the (1 - (1 - p)^N) curve looks linear for a wide region around the origin. So even though you remember your fourth-grade math teacher teaching you that you cannot additively accumulate probabilities (i.e., your probability of getting a job is not (N × p)), in this small-p case, it is a pretty decent approximation. In particular, even with our ostensibly large N, it is the case that

N × p = (50)(1/404) ≈ 12.38%,

which is pretty close to our slightly more dismal 11.65%.

In December, I ran into a woman who just got finished submitting all of her faculty positions. She said she applying to just 10 of them because she was exhausted and figured she was just practicing this round. Setting N = 10 reduces your chances to 2.45%. Having said that, the distribution across the applicant pool is certainly not flat. Her home institution, research, adviser, and other factors make her a very attractive candidate who will likely do well with such a low N... In fact, she was recently interviewed at a university near me (that, again, may have to deal with hiring freezes, etc., in the near future).

Now, in my case... Maybe I should burn my CV and dust off my résumé... I hope I'm not too old and outdated.

cleveref incompatibilities with Elsevier and Springer

elsart problems: If you try using cleveref with any of the elsart packages, you may get this error:

! TeX capacity exceeded, sorry [input stack size=5000].
\cl@part ->\cl@part 
                    \@elt {section}

This is due to a flaw in the elsart packages. They setup some part counters even though part isn't defined. This confuses cleveref. To fix this, add the following anywhere in your preamble:

\makeatletter \let\cl@part\relax \makeatother

Later, you can use cleveref's poorman option to strip all of the cleveref stuff from your document. After that, you can remove this line.

svjour problems: Similarly, if you try combining cleveref with Springer journal support files, you may get this error:

! TeX capacity exceeded, sorry [input stack size=5000].
\cl@chapter ->\cl@chapter
                          \@elt {theorem}

Again, this is due to a flaw in the Springer packages. They setup some chapter counters even though chapter isn't defined. To fix this, add the following anywhere in your preamble:

\makeatletter \let\cl@chapter\relax \makeatother

Again, this can be removed after cleveref's poorman procedure is applied.

Source and acknowledgments: The Springer solution was given by "Dan" in a comp.text.tex post. I used it to generate the elsart solution.

A reflection on (Springer) bibliographies (junior-part commas)...

In my previous post about Springer journal typesetting, I present the BiBTeX bibliography style files spmpsci.bst and spmpscinat.bst that were mostly generated by the docstrip driver files spmpsci.dbj and spmpscinat.dbj, respectively. I say mostly because I discovered that the original bibliography style files distributed by Springer were generated by docstrip but then manually hacked to remove the comma before the junior part (e.g., "Joe Public, Jr.") of the author and editor names.

You see, there is no way to use docstrip drivers based on merlin.mbs (from custom-bib) to have commas between author names and yet no commas before junior parts. Springer actually used the merlin.mbs option nm-rvv to get rid of commas and then hacked in between-author commas. I used the nm-rvvc option to include between-author commas and then hacked out the junior-part commas. The result is the same.

I have no problem listing author names without junior parts. As I mention in a Google groups thread, on page 3 of The Elements of Style by William Strunk Jr. and E.B. White, it is stated that the "Jr." after a last name should not have a comma before it (e.g., William Strunk Jr.). A number of journals agree with this, including The Journal of Mathematical Biology (i.e., a Springer journal using the Springer spmpscinat BiBTeX bibliography style).

At first I completely agreed with Springer and thought that merlin.mbs needed to be updated so that it had a junior part comma option. However, I gave it more thought and now I think that Springer should cease and desist. For one, this makes it difficult to regenerate the BST file from the DBJ driver options. The BST file distributed by Springer doesn't even mention that some manual editing has gone on, so anyone trying to generate their BST file will probably miss this hack. Second, while I think that authors who insert commas in their names need to stop this practice, if an author chooses to make his or her professional name comma-delimited, then journals should respect that author's wishes. Finally, if the BST junior-part comma is left in, commas can still be omitted by surrounding the junior (e.g., "Jr.") and the last name (e.g., "Smith") with curly braces (e.g., "{Smith Jr.}") in the BIB file. I haven't tested this, but I think that if this is done the Jr. becomes part of the BiBTeX last name (e.g., "{ll}") and thus there will be no comma before the junior in the resulting bibliography.

I'm sure Springer would tell me that they just want the bibliography entries to be consistent. However, names are messy things to mess with. Plus, modern-day BST files really should be specified entirely by a set of standard docstrip options (i.e., docstrip options and an MBS file). If non-standard options are really needed, a new MBS file should be distributed and the BST file should be spcified by options for that new MBS file. Would that be so hard to do for a publisher?

So that's my thought... Hey, publishers! Leave them BST files alone!

Preamble for Journal of Mathematical Biology

Update 5: The updated svglov2.clo and svglov3.clo are not needed if the preamble is used below (i.e., if the \RequirePackage{fix-cm} line is added before the documentclass line.

Update 4: See the section that redefines \autoref. Changed definition to support starred commands that are unlinked.

Update 3: See the section starting with the comment "%% Level 1 Subsections". It forces subsections to show up like sections in the PDF bookmarks; this is a (strange) JMB convention. In the comment I make some suggestions for improvements.

Update 2: See the small commented section "%%%% Table Support". If the LaTeX code is uncommented there, there is no need for the \noalign{\smallskip} stuff around \hline horizontal table rules. It uses the tabls package to do this.

Update 1: I've mirrored this content on my web page.

I'd like to combine some of the previous posts. I've been trying to fix and modernize some of the Springer LaTeX support files for the Journal of Mathematical Biology (JMB). I've come up with these...

• svglov2.clo - document class option for svjour2.cls. I used fix-cm to get rid of some of the warnings about not being able to scale the Computer Modern fonts. This is not needed if the preamble below is used (i.e., with the fix-cm line before the documentclass line).


• svglov3.clo - document class option for svjour3.cls. I used fix-cm to get rid of some of the warnings about not being able to scale the Computer Modern fonts. This is not needed if the preamble below is used (i.e., with the fix-cm line before the documentclass line).


• spmpsci.bst - BiBTeX bibliography style file for JMB. The original version of this did not reverse the first and last name of the editors as required; I fixed that.


• spmpscinat.bst - natbib compatible BiBTeX bibliography style file for JMB. Be sure to include natbib with numbers and sort&compress options.

I recommend using a preamble like this one...

%%%% Journal of Mathematical Biology (JMB) setup

% Allow Computer Modern fonts to be scaled (must be before svjour3)
\RequirePackage{fix-cm}

% Use running heads
\documentclass[runningheads]{svjour3}

% The journal's name
\journalname{Journal of Mathematical Biology}

% use Times fonts
\usepackage{mathptmx}

% flush right qed marks (Halmos square), 
% e.g., at end of proof
\smartqed

% Use natbib for \citet, \citep, etc.
% (JMB: numbered citations that are sorted 
%       and compressed)
\usepackage[numbers,sort&compress]{natbib}
%\bibliographystyle{spmpsci}
\bibliographystyle{spmpscinat}

%%%% Some useful packages

% Allow for smarter labeling of enumerations 
% and itemizations (consider using enumitem instead)
\usepackage{paralist}

% Provides \labelformat, which changes how \ref 
% references look
\usepackage{varioref}

% Mathematical symbols, etc.
\usepackage{amssymb,amsfonts,amsmath}

%%%% Table Support

% UNCOMMENT THE FOLLOWING LINES TO GET RID OF
% \noalign{\smallskip} STUFF AROUND \hline's IN
% TABLES
%
% % The following three lines remove the need for 
% % the \noalign{\smallskip} around the \hline's 
% % in the _Journal of Mathematical Biology_ LaTeX
% % template
% \usepackage{tabls}
% \addtolength\extrarulesep{\smallskipamount}
% \addtolength\extrarulesep{1pt}

%%%% Graphics and Figure Support

% Use subfig for subfigures
\usepackage[nearskip=-3pt,captionskip=4pt,
            listofformat=subsimple,
            labelformat=simple]{subfig}

% Make sure subfigures have parentheses around 
% them everywhere
\renewcommand\thesubfigure{(\alph{subfigure})}

% Use graphicx for including graphics
\usepackage{graphicx}

% When picture environments are used, use 
% pict2e for better resolution and flexibility
\usepackage{pict2e}

%%%% Hyperlink and Autoreference Support

%% Level 1 Subsections
%
% JMB does this weird thing where all 
% subsection-level PDF bookmarks are displayed as 
% section-level bookmarks. Strangely, they do 
% *NOT* turn on the ``bookmarksnumbered'' hyperref 
% option. I think their articles would be MUCH more 
% readable with all level-1 bookmarks if they were 
% prefixed by their section number. I'm sure they'd 
% argue that since the sections are linked within 
% the document, this is not needed.
%
% (note: another option is to set tocdepth=2 and turn 
%        on the ``bookmarksopen'' hyperref option)
%
% Anyway, the following redefines \addcontents line 
% to setup the bookmarks like JMB articles. Must do 
% this after hyperref because it redefines 
% \addcontentsline. Therefore, just use an 
% AtBeginDocument.
%
\usepackage{ifthen}
\AtBeginDocument{%
\let\orgaddcontentsline\addcontentsline
\renewcommand{\addcontentsline}[3]{
    \ifthenelse{\equal{#1}{toc}}
        {\ifthenelse{\equal{#2}{subsection}}
            {\orgaddcontentsline{#1}{section}{#3}}
            {\orgaddcontentsline{#1}{#2}{#3}}}
        {\orgaddcontentsline{#1}{#2}{#3}}}}

% Include hyperref for link support
% Include hyperref for link support
%\usepackage[dvipdfmx,   % If need dvipdfmx
%\usepackage[dvips,      % If need dvips
\usepackage[%           % Fine in most cases
            pdfpagelabels,hypertexnames=true,
            plainpages=false,
            naturalnames=false]{hyperref}

% Configure the hyperlink color
\usepackage{color}
\definecolor{darkblue}{rgb}{0,0.1,0.5}
\hypersetup{colorlinks,
            linkcolor=darkblue,
            anchorcolor=darkblue,
            citecolor=darkblue}

% In the future, it would be nice to use the package
%
%    cleveref
%
% Right now, it breaks svjour3 and svjour2, so we just 
% have to use autoref and labelformat :(.

% Make sure \autoref puts parentheses around 
% equations and enumeration items (similar to \eqref)
% NOTE: This allows us to use \ref instead of \eqref
\labelformat{equation}{\textup{(#1)}}
\labelformat{enumi}{\textup{(#1)}}

% Make sure equations are numbered by section
\numberwithin{equation}{section}

%% We need to redefine \autoref. We should use 
%% abbreviations inside the sentence and full names 
%% at the beginning of sentences. Additionally,
%% need to handle the plural cases.

% \Autoref is for the beginning of the sentence
\let\orgautoref\autoref
\providecommand{\Autoref}
        {\def\equationautorefname{Equation}%
         \def\figureautorefname{Figure}%
         \def\subfigureautorefname{Figure}%
         \def\sectionautorefname{Section}%
         \def\subsectionautorefname{Section}%
         \def\subsubsectionautorefname{Section}%
         \def\Itemautorefname{Item}%
         \def\tableautorefname{Table}%
         \orgautoref}

% \Autorefs is plural for the beginning of the sentence
\providecommand{\Autorefs}
        {\def\equationautorefname{Equations}%
         \def\figureautorefname{Figures}%
         \def\subfigureautorefname{Figures}%
         \def\sectionautorefname{Sections}%
         \def\subsectionautorefname{Sections}%
         \def\subsubsectionautorefname{Sections}%
         \def\Itemautorefname{Items}%
         \def\tableautorefname{Tables}%
         \orgautoref}

% \autoref is used inside a sentence 
% (this is a renew of the standard)
\renewcommand{\autoref}
        {\def\equationautorefname{Eq.}%
         \def\figureautorefname{Fig.}%
         \def\subfigureautorefname{Fig.}%
         \def\sectionautorefname{Sect.}%
         \def\subsectionautorefname{Sect.}%
         \def\subsubsectionautorefname{Sect.}%
         \def\Itemautorefname{item}%
         \def\tableautorefname{Table}%
         \orgautoref}

% \autorefs is plural for inside a sentence
\providecommand{\autorefs}
        {\def\equationautorefname{Eqs.}%
         \def\figureautorefname{Figs.}%
         \def\subfigureautorefname{Figs.}%
         \def\sectionautorefname{Sects.}%
         \def\subsectionautorefname{Sects.}%
         \def\subsubsectionautorefname{Sects.}%
         \def\Itemautorefname{items}%
         \def\tableautorefname{Tables}%
         \orgautoref}

This lets you do cool things like... (hyperlinks depicted in blue)

\ref{eq:1}		produces		linked		"(1)"
\ref*{eq:1}		produces		unlinked		"(1)"
\autoref{eq:1}		produces		linked		"Eq. (1)"
\autoref*{eq:1}		produces		unlinked		"Eq. (1)"
\Autoref{eq:1}		produces		linked		"Equation (1)"
\autoref{sec:2}		produces		linked		"Sect. 2.1"
\Autoref{sec:2}		produces		linked		"Section 2.1"
\Autorefs{sec:1} and \ref{sec:2}		produces		linked		"Sections 1 and 2.1"
\autorefs{sec:1} and \ref{sec:2}		produces		linked		"Sects. 1 and 2.1"
\autoref{item:1}		produces		linked		"item (i)"
\autorefs{item:1} and \ref{item:2}		produces		linked		"items (i) and (ii)"
\citep{SK86}		produces		linked		"[61]"
\citep[p.~41]{SK86}		produces		linked		"[61, p. 41]"
\citep{Cha76,SK86}		produces		linked		"[14, 61]"
\citep[e.g.,][]{SK86,Cha76}		produces		linked		"[e.g., 14, 61]"
\citeauthor{SK86}		produces		linked		"Stephens and Krebs"
\citet{SK86}		produces		linked		"Stephens and Krebs [61]"

spmpscinat.bst: A natbib style file for Journal of Mathematical Biology (e.g., Springer journals)

UPDATE: I learned how to use custom-bib's makebst. I've updated the spmpscinat.bst file in a more proper way by using makebst. I also regenerated the spmpsci.bst file so it properly reverses editor names to match the JMB format. I will submit these to JMB and hopefully they'll use them. FYI, these were mostly generated from spmpscinat.dbj and spmpsci.dbj, respectively.

I was frustrated to find out that the Journal of Mathematical Biology (JMB) has been using some outdated LaTeX support files. In particular, they distribute a BiBTeX style file that has no support for natbib use (or anything similar, like apacite). This appears to be less of a problem with JMB and more of a problem with Springer in general (i.e., most of its journals). They actually expect people to type the names of the authors that they reference directly (either that or use references (e.g., [2]) as nouns, which is bad style even though it saves space)! Can you believe that?

Civilized people use a package like natbib to do something like \citet{SK86} on every textual citation, which will then generate "Stephens and Krebs (1986)" (in Harvard mode) or "Stephens and Krebs [61]" (in numerical mode). There are lots of good reasons to do this. For one, it prevents me from accidentally turning "Krebs" into "Krbes".

To deal with this, I took the abbrvnat.bst style file packaged with natbib and hacked its support into the spmpsci.bst file from Springer. The result is spmpscinat.bst, which seems to work as desired for most bibliography entries I have tried.

I don't have a lot of experience with working with BiBTeX support files, so I have a feeling that it would have been better to use custom-bib to generate a BiBTeX style file in the JMB format that would be compatible with natbib. I may do that later.

So basically, I reserve the right to update or changed spmpscinat.bst at any time and without any notice, so look out for updates.

paralist, varioref, and hyperref: a happy combination

Consider the following...

\usepackage{varioref}
\labelformat{equation}{\textup{(#1)}}

This uses the clever varioref package to alter how equation references are displayed. The reference will be displayed with parentheses around the reference and the whole thing will be forced upright (so things don't look stupid in an italicized theorem environment). For example, for an equation labeled "eq:1" and a figure labeled "fig:1", I get these results:

\ref{eq:1} produces "(1)"
\ref{fig:1} produces "1"

This removes the need for \eqref from amsmath; just use \ref instead. Now, let's add hyperref:

\usepackage{hyperref}

So,

\ref{eq:1} produces a linked "(1)"
\ref*{eq:1} produces an unlinked "(1)"
\ref{fig:1} produces a linked "1"
\ref*{fig:1} produces an unlinked "1"
\autoref{eq:1} produces a linked "Equation (1)"
\autoref*{eq:1} produces an unlinked "Equation (1)"
\autoref{fig:1} produces a linked "Figure 1"
\autoref*{fig:1} produces an unlinked "Figure 1"

So this gives us the equivalent of an unlinked "\eqref*" through \ref*. Now, let's add paralist...

\usepackage{paralist}
\labelformat{equation}{\textup{(#1)}}

and an enumerate environment with automatic bracketed lowercase roman numeral numbering...

\begin{enumerate}[(i)]
    \item This is item one. \label{item:1}
    \item This is item two. \label{item:2}
\end{enumerate}

The \label macros can be associated with any enumerate item, even without paralist. The package paralist allows us to easily customize the displayed labels (enumitem is nice alternative). So, we get an enumerated output like...

 (i) This is item one.
(ii) This is item two.

Now we get...

\ref{item:1} produces a linked "(i)"
\ref*{item:1} produces an unlinked "(i)"
\autoref{item:1} produces a linked "item (i)"
\autoref*{item:1} produces an unlinked "item (i)"

Isn't life sweet? Notice the similarities between these things and stuff provided by prettyref and typedref. I think hyperref's autoref with the flexibility provided by varioref is a nice way to replace those old goodies.

Now can someone convince the Journal of Mathematical Biology to start using a bibliography style compatible with natbib? That would truly make me happy...