# Indenting in enumerate (TeX/LaTeX)

I'm making a numbered list with the enumerate environment. Through the enumitem package I've changed the label to [label=Caso \arabic*)]. I have set itemindent to 10.5pt to align the label start with the start of the lines above. Trouble is, as in the picture below, the lines of an item besides the first one are not aligned with the first one but indented slightly to the left. How do I correct that? P.S. adding leftmargin=0em or any other length just moves the whole thing to the left, not solving the problem at all.

\ben[label=Caso \arabic*)]
\setlength{\itemindent}{10.5pt}
\item Supponiamo $F=]x,y]$ e $\dsum_{n=1}^\8\lg(F_n)<+\8$, perché se non è vero non ho niente
da dimostrare, perché devo dimostrare $\lg(F)\leq\dsum\lg(F_n)$; io so che $]x,y]$ è ricoperto
dagli $F_n$, il guaio è l'infinità, perché sto ricoprendo con intervalli non aperti un intervallo
non chiuso, ma poco male: ingrasso ogni intervallino in modo da renderlo aperto e chiudo $[x,y]$,
così ho un compatto ricoperto da aperti; per ogni $n$ dico che posso trovare $F_n^\1$ unione
finita di intervallini semiaperti tali che $F_n\sbs\pint F_n^\1$ e
$\lg(F_n^\1)\leq\lg(F_n)+\fr{\meg}{2^n}$; se:


NB preamble at https://gist.github.com/anonymous/9853413, it probably wouldn't fit due to its length.

-
Can you add the code you have now? However, I usually don't recommend splitting proofs in long lists. Just emphasizing “Caso 1” should be sufficient. –  egreg Mar 29 at 11:55
Not related directly to your question, but of some relevance for the overall impression of the typeset product: In running text, don't place the limits of summation above and below the \sum symbol; placing the limits next to the symbol, via \sum\nolimits_{n=1}^\infty, will probably give a better-looking paragraph. –  Mico Mar 29 at 12:04
@egreg I have added the code of the picture to the question with a gist of the long preamble behind it. In case you want to know, the code of the picture is in a file called 27.3 included in the main file with \include, and currently I'm working with \includeonly{27.3}. @Mico you may think that way, but I'm doing that on purpose (actually forcing \displaystyle - see definition of \dsum in gisted preamble) to avoid placing the limits on the side, which looks horrible in my opinion, so thanks, but no. One thing I'll try is adding \limits rather than forcing \displaystyle. –  MickG Mar 29 at 12:18
It shouldn't be difficult to make a preamble with just the code necessary for the example; the text is immaterial, so you can simply purge the personal commands. –  egreg Mar 29 at 12:34
@MickG the inline style for sum may look strange but no math journal would publish that paragraph with a display sum set inline like that. It distorts the baseline spacing so much, the problem with the text in the image is not mainly that the margin is out of line it is that the first line looks like a separate paragraph as it is set off with so much vertical space from the following lines, similarly the last line. –  David Carlisle Mar 29 at 12:53

Another alternative. The OP did not provide the commands used in math mode. This solution removed them to make a run.

Code

\documentclass{article}

\usepackage[utf8]{inputenc}
\usepackage{enumitem}
\usepackage{amsmath,amssymb}

\begin{document}

\noindent x\hrulefill x

\begin{enumerate}[
labelindent=*,
style=multiline,
leftmargin=*,
label=Caso \arabic*)
]

\item Supponiamo $F=]x,y]$ e $\sum_{n=1}^8, (F_n)<+$, perché se non è vero non ho niente
da dimostrare, perché devo dimostrare $\lg(F)\leq\sum(F_n)$; io so che $]x,y]$ è ricoperto
dagli $F_n$, il guaio è l'infinità, perché sto ricoprendo con intervalli non aperti un intervallo
non chiuso, ma poco male: ingrasso ogni intervallino in modo da renderlo aperto e chiudo $[x,y]$,
così ho un compatto ricoperto da aperti; per ogni $n$ dico che posso trovare $F_n^1$ unione
finita di intervallini semiaperti tali che $F_nF_n^1$ e
$(F_n^1)(F_n)+{2^n}$; se:

\item Supponiamo $F=]x,y]$ e $\sum_{n=1}^\infty\lambda(F_n)<+\infty$, perché se non è vero non ho niente da dimostrare\ldots
\end{enumerate}

\end{document}

-

## EDIT

Sorry for misunderstanding... To obtain what you want, you just have the use the option:

leftmargin=*


The following MWE (I've replaced your custom commands with known ones):

\documentclass{article}

\usepackage[utf8]{inputenc}
\usepackage{enumitem}

\begin{document}

\noindent x\hrulefill x

\begin{enumerate}[leftmargin=*,label=Caso \arabic*)]
\item Supponiamo $F=]x,y]$ e $\sum_{n=1}^\infty\lambda(F_n)<+\infty$, perché se non è vero non ho niente
da dimostrare\ldots
\end{enumerate}

\end{document}


produces the desired result:

If you want your list to behave like a normal paragraph with no indentation, you have to use the following options:

leftmargin=0pt,itemindent=*


The following MWE (I've replaced your custom commands with known ones):

\documentclass{article}

\usepackage[utf8]{inputenc}
\usepackage{enumitem}

\begin{document}

\noindent x\hrulefill x

\begin{enumerate}[leftmargin=0pt,itemindent=*,label=Caso \arabic*)]
\item Supponiamo $F=]x,y]$ e $\sum_{n=1}^\infty\lambda(F_n)<+\infty$, perché se non è vero non ho niente
da dimostrare\ldots
\end{enumerate}

\end{document}


produces the desired result:

-
The desired result is actually not that one, as "ho niente" is not aligned with "Supponiamo", as I wish it to be, but with the label. It is true that this way it's better than at the beginning. –  MickG Mar 29 at 13:30

Use correctly the parameters of enumitem.

\documentclass{article}

\usepackage[pass,showframe]{geometry} % just to show the page frame

\usepackage[T1]{fontenc}
\usepackage[utf8]{inputenc}
\usepackage[italian]{babel}
\usepackage{enumitem}
\usepackage{calc}

\newcommand{\pint}[1]{\mathring{#1}}
\newcommand{\eps}{\varepsilon}

\begin{document}
\begin{enumerate}[
label=Caso \arabic*),
labelwidth=\widthof{Caso 1)},
leftmargin=\widthof{Caso 1)\enspace},
labelsep=.5em,
]

\item Supponiamo $F=\mathopen{]}x,y]$ e $\sum_{n=1}^{\infty}\lambda(F_n)<+\infty$,
perché se non è vero non ho niente da dimostrare, perché devo dimostrare
$\lambda(F)\leq\sum\lambda(F_n)$; io so che $\mathopen{]}x,y]$ è ricoperto dagli $F_n$,
il guaio è l'infinità, perché sto ricoprendo con intervalli non aperti un intervallo non
chiuso, ma poco male: ingrasso ogni intervallino in modo da renderlo aperto e chiudo
$[x,y]$, così ho un compatto ricoperto da aperti; per ogni $n$ dico che posso trovare
$F_n'$ unione finita di intervallini semiaperti tali che $F_n\subset\pint{F}_n'$ e
$\lambda(F_n')\leq\lambda(F_n)+2^{-n}\eps$; se:

\end{enumerate}

\end{document}


I removed your personal commands. Some remarks:

1. using \displaystyle\sum in inline formulas is simply wrong, because it ruins the spacing between lines; the same hold for big fractions, I used 2^{-n}\eps instead of a fraction;

2. a name such as \meg for epsilon is a sure way for forgetting its meaning; also, abbreviating \frac into \fr is not recommendable;

3. you gain nothing with \lg instead of \lambda; the same with \8 instead of \infty;

4. typing X' in math mode is just the same as doing X^{\prime}, which is surely handier than X^\1;

5. the notation ]a,b] for half-open intervals is ugly; in any case you have to declare correctly ] when used as an opening item;

6. you gain nothing (and lose something) by using \ben instead of \begin{enumerate}.

Final word I would never use an enumerate environment for this. Just emphasizing “Caso” is sufficient.

Also the Italian is not very smooth, for example you have a pair of consecutive “perché”; “io so” should be “so”.

-
Thanks for the answer. On your final points: 1) We have divergent opinions on that, period; let's cut this argument; 2) Not true: I assure you I remember every single greek letter command, because there is iron logic behind them: g for greek at the end, a letter matching the greek letter in the middle, and an optional b for "big", e for "exponent scaling" (nothing is for normal scaling) and m for "\mathclap" at the start; as for \fr, it actually abbreviates \dfrac, and besides why would it not be recommendable; –  MickG Mar 29 at 14:54
3) \lg has a scale in it, and is far shorter than \lambda anyway, as \8 compared to \infty, I think; 4) Yes I know; I found that out when I was already too used to ^\1 to change my typing habit; 5) That's true, but I am adopting my teacher's notation in the notes; what do you mean, "declare correctly ] when used as an opening item"? 6) With \ben I gain in speed, which is appreciated when taking notes; what do I lose? 7) Yes, you already said you wouldn't do this; 8) The Italian is my professor's spoken Italian, which explains the non-smoothness. –  MickG Mar 29 at 14:57
@MickG Check the difference in spacing between $I=]a,b]$ and $I=\mathopen{]}a,b]$. Do you really open environments while taking notes? –  egreg Mar 29 at 15:02
Oh yes I do @egreg. And many of them. Indeed the main thing I do when running over notes is improve the spacing and add [name=, [label= to theorems, not much more. Actually, the [label=\arabic*)] was done while taking notes. As for the \mathopen, I agree the spacing is better with than without, but I never knew about that command. I'll keep that in mind. –  MickG Mar 29 at 15:20