Debian Codegraph @

texmf-dist/source/latex/paracol/probs.dtx

raw
1 % \section{Known and Unknown Problems}
2 % \label{sec:problem}
3 % \changes{v1.2-7}{2013/05/11}
4 % {Add the section ``Known and Unknown Problems'' to summarize a few
5 % typesetting issues and warn users of the possibility of bugs.}
6 % \changes{v1.3-1}{2013/09/17}
7 % {Remove the problem description of the placement of page-crossing
8 % spanning texts because it has been solved.}
9 % \changes{v1.3-3}{2013/09/17}
10 % {Remove the problem description of the lack of column-separating rule
11 % drawing because it has been implemented.}
12 %
13 % Here a few problems you could face in the use of \textsf{paracol} are
14 % summarized.
15 % 
16 % \begin{itemize}
17 % \item
18 % If your (e.g.,) left column goes ahead too much farther than the right
19 % column, \LaTeX{} could stop with the following error message.
20 % 
21 % \begin{quote}
22 % |! Package paracol Error: Too many unprocessed columns/floats.|
23 % \end{quote}
24 % 
25 % This usually means that the internal space to keep materials in the left
26 % column is exhausted. More specifically, suppose at some point in your
27 % |.tex| the left column is in the page $p$ while the right is in $q<p$.
28 % We need $(p-q)$ \emph{boxes} to keep the left column contents in the pages
29 % $q$, $q+1$, \ldots, $p-1$ because these pages cannot be \emph{printed} yet
30 % until the right column fills them. In addition, we also need two boxes
31 % for the left column in $p$ and the right column in $q$ so that you make
32 % \cswitch{} between them keeping unprinted contents in them. Therefore, at
33 % least we need to have $(p-q)+2$ boxes, while the number of them provided
34 % by \LaTeX{} is only 18\footnote{
35 % 
36 % Readers who are acquainted with \LaTeX{} implementation will understand
37 % that 18 is the cardinality of the set $\{\cs{bx@A},\ldots,\cs{bx@R}\}$ for
38 % floats acquired by \cs{newinsert}. Those who are more familiar with that
39 % might know that most \LaTeX{}, based on e-\TeX{} or others having similar
40 % extensions, now have 52 \cs{insert}s
41 % $\{\cs{bx@A},\ldots,\cs{bx@Z},\cs{bx@AA},\ldots,\cs{bx@ZZ}\}$ for floats
42 % and materials of \Paracol, since 2015}.
43 % 
44 % Therefore, \textsf{paracol} cannot continue its work if $(p-q)$ reaches
45 % 17. Furthermore, other stuff also consumes the boxes as follows.
46 % 
47 % \begin{itemize}
48 % \item
49 % If there are $n$ pages in $q$, $q+1$, \ldots, $p$ having \preenv{} or
50 % page-wise floats, $n$ boxes are consumed by them. Similarly, if $m$ pages
51 % in them have \Scfnote{}s, $m$ boxes are given to them.
52 %
53 % \item
54 % If the left (resp.\ right) column has \Mcfnote{}s in $p$ (resp.\ $q$), a box
55 % is used for them.
56 %
57 % \item
58 % If the left (resp.\ right) column has $k$ floats to be placed in $p$
59 % (resp.\ $q$) or to be deferred to $p+1$ (resp.\ $q+1$) or a succeeding page,
60 % $k$ boxes are reserved for them.
61 % \end{itemize}
62 % 
63 % Therefore, it should be safe to keep $(p-q)$ from exceeding 10 or so
64 % placing \!\switchcolumn! in both columns fairly frequently.
65 %
66 % \item
67 % As discussed in Section~\ref{sec:ref-switchcolumn}, setting a \sync{}ation
68 % point in a page brings the following side effects.
69 % 
70 % \begin{itemize}
71 % \item
72 % Stretch and shrink factors of all vertical skips in the page are
73 % nullified. The nullification of stretch factors could make a sparse
74 % column in the page have a vertical space at its bottom as if
75 % \!\raggedbottom! setting is in effect even with \!\flushbottom! one,
76 % rather than distributing the amount of the space to the skips so that the
77 % bottom line is aligned at the page bottom. As for the nullification of
78 % shrink factors, it makes the page have lines a little bit less than that
79 % it would have without \sync{}ation because lines above the (last)
80 % \sync{}ation point cannot be compressed. The other effect is a little bit
81 % subtle because the shrink factors below the last \sync{}ation point are
82 % taken care of by \TeX{}'s page builder when it examine the appropriateness
83 % of each breakable point, but they are nullified when the page is printed.
84 % That is, if \TeX{} finds a good break point which needs that the stuff
85 % between the \sync{}ation and break points is compressed a little bit, the
86 % stuff is printed without compression making its bottom edge a little
87 % bit below the page bottom.
88 %
89 % \item
90 % After a \sync{}ation point is set, columns in the page cannot have top
91 % floats any more even if a column has space above the \sync{}ation point
92 % and large enough to place the float. Therefore, if you like to exploit
93 % the space, you have to place the \env{figure} or \env{table} environment
94 % in question prior to the \cswitch{} command or environment for the
95 % \sync{}ation.
96 % \end{itemize}
97 %
98 % \item
99 % \changes{v1.3-6}{2013/09/17}
100 % {Add comments about usage of \cs{paragraph} etc.\ in spanning texts.}
101 % \begingroup\parskip\z@
102 % As the author did for Section\Tie1 to \ref{sec:float}, sometimes you will
103 % make a section header spanning all columns by giving a sectioning command
104 % such as \!\section!, \!\subsection! and \!\subsubsection! to the optional
105 % argument of \!\switchcolumn!|*| or \!\begin! of a \sync{}ing \csenv.
106 % These three commands work well and you will have what you intend to have,
107 % but you have to be careful with lower-level commands \!\paragraph! and
108 % \!\subparagraph!. Unlike higher-level relatives, these lower-level
109 % commands does \emph{not} put the header \emph{immediately} but keep it
110 % somewhere\footnote{
111 % 
112 % For people familiar to \TeX's \emph{dangerous bends}, the header is kept in
113 % \!\everypar!.}
114 % 
115 % so that when the paragraph following the command starts it is put as the
116 % leading part of the paragraph. Therefore if the \mctext{} has (e.g.)
117 % \!\paragraph! only, the header is not put as a \mctext{} but at the head of
118 % the first paragraph of the column to which you switch, leaving an
119 % empty \mctext{} with some large space as follows.
120 % 
121 % \par\leavevmode\Hrule
122 % \columnsep\z@
123 % \begin{paracol}{2}
124 % This left-column paragraph precedes a \sync{}ed \cswitch.
125 % \switchcolumn
126 % This right-column paragraph precedes a \sync{}ed \cswitch.
127 % \switchcolumn*[\paragraph{A Spanning Text Given by \cs{paragraph}}]
128 % This left-column paragraph follows the \sync{}ation but is led by
129 % \!\paragraph! given to the optional argument of \!\switchcolumn!|*| for
130 % \mctext.
131 % \switchcolumn
132 % This right-column paragraph follows the \sync{}ation with an empty \mctext.
133 % \end{paracol}
134 % \par\leavevmode\Hrule\par
135 % 
136 % Therefore, unless this is what you intend to do, you have to give some
137 % paragraph together with \!\paragraph! to the optional argument for
138 % \mctext. For example, \!\mbox!|{}| is a good candidate as the paragraph
139 % following \!\paragraph! because it produces (almost) nothing. By using
140 % this technique the example above becomes the followings.
141 % 
142 % \par\leavevmode\Hrule
143 % \begin{paracol}{2}
144 % This left-column paragraph precedes a \sync{}ed \cswitch.
145 % \switchcolumn
146 % This right-column paragraph precedes a \sync{}ed \cswitch.
147 % \switchcolumn*[\paragraph{A Spanning Text Given by \cs{paragraph} Followed
148 % by \cs{mbox}\texttt{\char`\{\char`\}}}\mbox{}]
149 % This left-column paragraph follows the \mctext{} above.
150 % \switchcolumn
151 % This right-column paragraph follows the \mctext{} above.
152 % \end{paracol}
153 % \par\leavevmode\Hrule
154 % \endgroup
155 %
156 % \item
157 % As shown in Section~\ref{sec:fnnp}, it is not easy to have good numbering
158 % and stacking order of \Scfnote{}s even with the supports from
159 % \!\footnote!|*| and its relatives. In addition, a footnote in a
160 % \env{paracol} environment cannot be broken into two (or more) pages.
161 %
162 % \item
163 % \changes{v1.3-2}{2013/09/17}
164 % {Add comments about the limitation of parallel-paging.}
165 % 
166 % As the author confessed in Section\Tie\ref{sec:ppts-paired}, right
167 % \parapag{}es cannot have \pwstuff{} but have blank spaces in the
168 % corresponding region for them. The author will try to remove this
169 % limitation from a future version of \Paracol, in the version 1.4 hopefully.
170 %
171 % \item
172 % \changes{v1.3-3}{2013/09/17}
173 % {Add comments about the imperfectness of extension of background
174 % painting regions.}
175 % \begin{Hfuzz}{1.5pt}
176 % As discussed in Section\Tie\ref{sec:bgpaint-me}, it is desirable that
177 % \bgpaint{} region definition in \!\backgroundcolor! has position dependent
178 % extensions. The author is fairly optimistic about the incorporation of
179 % this advanced feature in the version 1.4.
180 % \end{Hfuzz}
181 %
182 % \item
183 % \changes{v1.32-3}{2015/10/10}
184 % {Add comments about the out-of-order appearance of page-wise floats
185 % even with \protect\LaTeX-2015/01/10 or later.}
186 % In the release dated 2015/01/10, \LaTeX{} changed its mechanism of the
187 % placement of double-column floats (or in our terminology, \pwise{}
188 % floats) to avoid {\em out-of-order\/} appearance of them. That is, until
189 % the release on 2014/05/01 a double-column float (e.g., \env{figure*}) can
190 % be overtaken by a single-column float of the same category (e.g.,
191 % \env{figure}) when they cannot be put into the page in which texts around
192 % them are put. In order to cope with the problem, the new version merged
193 % two lists to keep {\em deferred\/} double- and single-column floats into
194 % one so that the appearance order of them is determined by their order in
195 % the single list. Though this change should have made people happy when
196 % they typeset {\em ordinary\/} two-column (or multiple-column) documents,
197 % the new feature might not be welcomed by \Paracol{} users because your
198 % parallel-columns have their own {\em streams\/} of floats to be put in the
199 % corresponding columns. Therefore, and for the sake of simplicity of
200 % \Paracol's implementation, the author decided to nullify this new feature
201 % in \env{paracol} environments. That is, even with new releases of
202 % \LaTeX{}, your \pwise{} floats given in a \env{paracol} environment can be
203 % overtaken by \cwise{} floats.
204 % \end{itemize}
205 %
206 % In addition to the problems above known to the author, there may be (or
207 % should be, honestly speaking) other unknown problems in \textsf{paracol}
208 % because it cannot be perfect though the author has made his best effort
209 % for testing and debugging it. Particularly, sometimes it is very tough,
210 % if not impossible, to make \textsf{paracol} compatible with other
211 % packages, especially with those having dark magic as \textsf{paracol} has
212 % in it\footnote{
213 % 
214 % For example, the author knows it is almost impossible to make
215 % \textsf{paracol} compatible with one of the author's own package available
216 % in CTAN.}.
217 % 
218 % Therefore, though reporting incompatibleness with a package you use is
219 % very welcome\footnote{
220 % 
221 % For example, \textsf{paracol} is now compatible with \textsf{color}
222 % package thanks to a report from a user.},
223 % 
224 % you should kindly understand the toughness of the compatibility issue.
225 %
226 % Furthermore, even without such problematic packages, \textsf{paracol}
227 % might produce weird results due to its bug. If your document has
228 % something to make unknown bugs visible, you might have one (or more) of
229 % the followings which the author encountered in his debugging work.
230 % 
231 % \begin{itemize}
232 % \item
233 % A page, a column, a footnote and/or a float disappears\footnote{
234 % 
235 % In fact, a bug fixed in version 1.2 caused page losing though it happens
236 % very very rarely but an unlucky user encountered it.}.
237 %
238 % \item
239 % A page, a column, a footnote and/or a float is duplicated.
240 %
241 % \item
242 % A message like ``{\tt Overfull |\vbox| (1.23456pt too high) has occurred
243 % while |\ouptut| is active}'' is shown.
244 %
245 % \item
246 % A message ``{\tt Underfull |\vbox| (badness 10000) has occurred while
247 % |\ouptut| is active}'' is shown. This message, however, does not always
248 % mean a bug but may just be a complaint that a column or a page is too
249 % sparse to meet your request to align the bottom of all columns and pages
250 % by \!\flushbottom! setting. Therefore, if you have this message and you
251 % cannot be sure whether it means a bug or not, try \!\raggedbottom! setting
252 % to see if you still have the message, before sending a bug report to the
253 % author.
254 % \end{itemize}
255 % 
256 % If you encounter anything like them (or whatever you cannot solve by
257 % yourself), don't hesitate to report it to the author with minimum source
258 % file to produce the problem\footnote{
259 % 
260 % And with patience because your problem might not be solved quickly.}.
261 %
262 % \endinput