Welcome to the floatflt package!
Mats Dahlgren
(matsd@[Link])
1996/02/27
Abstract
The floatflt package is an extension of Thomas Kneser’s style option [Link] for LATEX 2.09. The
extension was done by Mats Dahlgren (matsd@[Link]).
This file and the package: Copyright c 1994-1996 by Mats Dahlgren. All rights reserved.
1 Introduction
The floatflt package defines two environments: floatingfigure and floatingtable. In principle the
two environments work the same way and do the same job, but there are some important differences, which are
explained below. The general idea is that a figure or table which is not very wide should be allowed to have regular
text typsetted beside itself. In Thomas Kneser’s file [Link] and in the documentation thereof, such
figures are called floating figures. In the floatflt package, the term floating is generally used for such figures and
tables, despite the slight different use of the term float in LATEX. Hence the term floating floats for such figures and
tables.
1.1 For [Link] users
There are five major differences in the use of floatflt as compared to floatfig:
The extension to the floattable environment.
There is no longer any need for initialization, thanks to the \AtBeginDocument command.
Optional arguments for the environments wihch allows the user to place the floating floats left, right or
according to the page number (as with [Link]); this is mainly due to the possibility to have
optional arguments in newly defined environments.
Optional argument for the package as a whole if the user wants to force all floating floats to be to the
right or left. The placement option at \begin{floatingfigure} (and \begin{floatingtable})
overrides this (and there is also the optional placement option to re-introduce the page-number dependent
placement for one figure or table).
Compatibility with the multicol package is now obtained.
2 Userguide
In the following, the usage of the package is explained in some detail. Users who are unfamiliar with the
[Link] package by Thomas Kneser may also find it useful to run LATEX on the file [Link].
2.1 How to call the package
The floatflt package is activated by the following line in your LATEX input file following your \documentclass
command:
\usepackage[option]{floatflt}
The option may be either one of the following: rflt, lflt, or vflt.
This document describes floatflt v.1.3 and was last updated 1996/02/27.
1
2.1.1 The package options
The three package options have the following functions:
rflt Forces the default for the floating figures and tables to be typset to the right in a paragraph.
lflt Forces the default for the floating figures and tables to be typset to the left in a paragraph.
vflt Forces the floating figures and tables to be typset to the right in a paragraph on odd-nubered pages, and to
the left on even-numbered pages; this is also the default.
The option vflt is actually not needed and does nothing but tells LATEX to use the default, it is also the internal
default. It is supported for symmetry reasons with the optional arguments for the environments temselves.
2.2 How to use the floatingfigure environment
In your LATEX document you invoke the floatingfigure environment by typing:
\begin{floatingfigure}[option]{width}
figure commands with or without \caption
\end{floatingfigure}
The option may be either one of the following: r, l, p, or v. The quantity width is the width you want your floating
figure to have (such as “75mm”).
2.2.1 The floatingfigure options
The options to the floatingfigure environment all overrule any present package option which may be in
effect. The options have the following functions:
r Forces the current floating figure to be typset to the right in a paragraph.
l Forces the current floating figure to be typset to the left in a paragraph.
p Forces the current floating figure to be typset to the right in a paragraph if the pagenumber is odd, and to the
left if even.
v Applies the package option to the current figure, and if no package option is specified, it forces the current
floating figure to be typset to the right in a paragraph if the pagenumber is odd, and to the left if even.
The p option is used when the default alternating right/left typesetting of floating figures is desired despite the use
of the rflt or lflt package option. The v option is the internal default, and does not have any effect other than
applying either the default or the option specified for the whole package.
2.3 How to use the floatingtable environment
In your LATEX document you invoke the floatingtable environment by typing:
\begin{floatingtable}[option]{
\begin{tabular}{tabular specifiers}
here you put your table entries
\end{tabular}}
\caption if desired
\end{floatingtable}
The option may be either one of the following: r, l, p, or v. No specification of the width is explicitly made,
instead the whole tabular environment is a mandatory argument. Thus, it is important to always have the “{”
before \begin{tabular} and the “extra” “}” after \end{tabular}. At the present, the tabbing envi-
ronment can not be used in the floatingtable environment. (Other commands and environments such as
the \parbox command may be used instead of the tabular environment within a floatingtable environ-
ment.)
2.3.1 The floatingtable options
The options for the floatingtable environment are exactly the same as for the floatingfigure environ-
ment, and they also have the same effects.
2
2.4 Compatibility
The present version of floatflt (v.1.3) has been tested with LATEX 2" of 1995/12/01 patch level 1 running TEX 3.1415
under Open VMS, using the article document class. Only minor changes of some of the internal commands have
been made since the previous verision (changes suggested by Werner Fink (werner@[Link])).
The regular \caption command of both the figure and tabular environments work fine. The
\hangcaption command of David M. Jones (dmjones@[Link]) does not work with float-
flt. The use of \hangcaption gives the same result as \caption. The listing commands \listoffigures
and \listoftables also work properly. However, since floating figures and tables normally are typsetted in a
paragraph, the order in the lists may be strange if also regular figure and table environments are used close
to the floating ones.
Unlike [Link], floatflt works with the multicol package, provided the floating floats are still less
wide than the columns. If not, or if the floating floats and the columns have almost the same width, text may dift
away and overlap text in adjacent column(s). However, this combination has several restrictions, and the placement
of the floating floats may not always be the correct according to the page numbering. Normally it also leads to
many ”Underfull \hbox” varnings. The use of this combination is discouraged.
2.5 Warnings and messages
In Thomas Kneser’s original work [Link], there are several warnings issued when problems occur.
These are kept in floatflt. Warnings will be issued if two floating figures, two floating tables, or if one floating table
is colliding with one floating figure. Also, message(s) will be displayed if a floating float is moved from its original
paragraph. Finally, each processed floating float will produce a message telling which page it has been typsetted
on.
3 Known problems and limitations
The following problems and limitations are known:
If a floating figure or table is called close to a sectioning command, the floating float may be lost or truncated.
This may lead to a “collision warning”, which normally has to be analyzed “by hand”.
If two consecutive paragraphs both contain a floating float, a “collision warning” is issued. This warning
may be ignored.
A floatingtable environment sometimes leads to a warning like:
Overfull \hbox (10.78334pt too wide) in paragraph at lines 287--289
where the line numbers refered to are those of the \end{tabular} statement. This could not be avoided
in all cases.
If you use floating floats of different kinds close to each other in your input, you are likely to have problems.
You may either lose one of the floating floats entirely, have it overwritten by the other, or have it truncated.
LATEX has problem in linebreaking short lines. This may give you bad linebreaks in the captions of floating
floats.
There are some problems when a floating float and a footnote appear on the same page, especially when the
floating float is moved to another page. Unfortunately, this may need some adjustments by hand. (Reported
by Jakob Schiotz (schiotz@[Link]).)
If a floating float is placed close to a list environment (such as enumerate or itemize), the text of the
list environment is likely to overwrite the floating float. See section 3.1 below for a semi-automatic solution
of this problem. (The bug was reported by Rafael Gallego (gallego@[Link]).)
The first two of these were pointed out by Thomas Kneser in the original documentation of [Link].
3.1 How to Avoid the List Problem
There is a semi-automatic way to avoid the “List Problem” mentioned above. The solution was introduced in
version 1.2. For items in a list which are colliding with a floating float, there is a special command \fltitem
avaliable, which takes one mandatory argument. This command is to be used in place of the ordinary \item, and
should have the text of the item as the argument. E. g. you would write
\fltitem{text}
3
instead of
\item text
to avoid the text of the \item command to be written in the floating float. \fltitem also takes a second,
optional, argument, which is an extra vertical space put in after the text of the \fltitem command. This is
meant to be used specially when there are two paragraphs in the same item text, and the latter is to use the full line
width (i. e. when the item text ends below the floating float). So, the full syntax of the \fltitem command is:
\fltitem[len]{text}
which inserts the extra vertical space len after the item text. If there is a second text which is to span the full
linewidth in the same item, one would write:
\fltitem[len]{text1} text2
which first typesets text1 narrowed according to the floating float, followed by text2 with full linewidth, and an
extra vertical space len between these two texts.
There is also a second, analogous command \fltditem which takes two mandatory and one optional argu-
ment. The optional argument is the same as the optional argument of \fltitem, the first mandatory argument is
the item label, e. g. the item to be described in a description environment. The second mandatory argument
of \fltditem is the item text. The syntax of the \fltditem command is:
\fltditem[len]{label}{text}
which produces an item with label label and text text and extra vertical space len after the item text. Additional text
spanning the full linewidth may follow, just like with \fltitem, but the indentation may turn out to be strange.
Both \fltitem and \fltditem can be used in an ordinary itemize, enumerate, or description
environment (however, the use of \fltditem in enumerate may produce strange results). For \fltitem
and \fltditem to work properly, the (otherwise) colliding floating float must be set flush right, that is have the
option r specified. The use of \fltitem or \fltditem is not demonstrated in the [Link] file.
3.2 Sending a bug report
Reports of new bugs in the package are most welcome. However, I do not consider this to be a “supported”
package. This means that there is no guarantee I (or anyone else) will put any effort into fixing the bug. But, on
the other hand, someone may, so filing a bug report is always a good thing to do! (If nothing else, your discoveries
may end up in future releases of this document.) Before filing a bug report, please take the following actions:
1. Ensure your problem is not due to your inputfile;
2. Ensure your problem is not due to your own package(s) or class(es);
3. Ensure your problem is not covered in the section “Known problems and limitations” above;
4. Try to locate the problem by writing a minimal LATEX 2" input file which reproduces the problem. Include
the command
\setcounter{errorcontextlines}{999}
in your input;
5. Run your file through LATEX 2" ;
6. Send a description of your problem, the input file and the log file via e-mail to:
matsd@[Link].
4 Conclusion
Without Thomas Kneser’s [Link] as a base, I could never have written the floatflt package. I hope that
some users will find the package useful and not too bugful. :–) Comments and suggestions for improvements are
always most welcome!
Enjoy your LATEX!
mats d.