@egreg Hello Prof. and good morning. I kindly ask you for an aesthetic opinion. I used this code to write this formula. Do you think there is a more elegant method to write it? I insert the fragment of my code and I apologize if I have disturbed you. Very cordial greetings.

\begin{equation}\label{TEL1}
\begin{split}
    \Delta f_y' &= f_y'(x+\Delta x, y+\Delta y, y'+\Delta y')-f_y'(x, y, y')= \\
    &=\bar{f}_{y'x}(x,y(x),y'(x))\Delta x+ \\
    &+\bar{f}_{y'y}(x,y(x),y'(x))\Delta y+\bar{f}_{y'y'}(x,y(x),y'(x))\Delta y'
  \end{split}
\end{equation}

@egreg This is the image i.sstatic.net/kOaSU.png

@Sebastiano I would use three lines, not two. Don’t repeat the + sign and remove the indent after the display.

@egreg Kindest Enrico thank you very much for your suggestion. ok for + sign. The code is:

\begin{equation}\label{TEL1}
\begin{split}
    \Delta f_y' &=\\
    &= f_y'(x+\Delta x, y+\Delta y, y'+\Delta y')-f_y'(x, y, y')= \\
    &=\bar{f}_{y'x}(x,y(x),y'(x))\Delta x+\bar{f}_{y'y}(x,y(x),y'(x))\Delta y+\bar{f}_{y'y'}(x,y(x),y'(x))\Delta y'
  \end{split}
\end{equation}

but the output it is ugly for the numered equation

@Sebastiano naturally if you have a line that long the equation number moves out of the way, if you don't like that split the equation over more lines

Debian release time.

Hi everyone

@Marijn very weird even by the standard of tikz questions, perhaps i should post a picture mode duck to add to the collection

@Marijn I was wondering the same thing. Probably the users were shocked

@Marijn because it is not a just-do-it-for-me question

@Marijn Ah no, I didn't read your comment, the code is not related with the image, it's an ordinary TikZ question, after all

If only Gutenberg had used Unicode. @barbarabeeton :)

@AlanMunn the answer being that those pesky Vikings took them to Iceland?

@DavidCarlisle Actually the pesky Vikings probably brought them to England. But then the pesky French took over.

@AlanMunn maybe we'll get them back after brexit

@AlanMunn @DavidCarlisle ^^

@DavidCarlisle Don't

@DavidCarlisle David, you're right, and I thank you for that. I had thought of moving the equation to the left with \mkern-20mu, but it goes too far to the left by overlapping on the green line of the edge of the theorem. My sincere greetings and thanks.

@JosephWright oh we broke the openmath standard:(

@DavidCarlisle ???

@JosephWright github.com/OpenMath/OMSTD/commit/…

@JosephWright it's not been run for a year or so but someone made a PR and travis errored, although actually the build failed on master.

@JosephWright who can you imagine suggested this catcode regime for the document?

@Sebastiano May I ask you a favor? Could you perhaps consider not to post your opinion in the comments so much? First of all, if several other users do that we will drown in comments. Also an explanation that you recently praised is not entirely correct.

\AtBeginDocument{
\let\cellsep&
\catcode`\&=12
\catcode`\_=12
\catcode`\^=12
\catcode`\#=12
\catcode`\$=12
\catcode`\£=1
\catcode`\ß=2
}

@DavidCarlisle Probably me?

@marmot Link?

@DavidCarlisle What change broke things?

@JosephWright no me, sometime around 1999 I guess,

@JosephWright Frank adding a % at ends of tables of contents, when % isn't a comment character...

@DavidCarlisle Not me then :)

@DavidCarlisle Ah, ooops,

@DavidCarlisle If it's FMi, that's Not Our Fault

@JosephWright I don't think the kernel should guard against that, if it can't assume \ { } % have standard catcodes then it can't do anything.

@DavidCarlisle Well the first three yes: ConTeXt in 'ASCII Mode' only assumes those

@JosephWright it was my bug report which triggered the change ;-(.

@DavidCarlisle hm, that still errors ;-).

@UlrikeFischer odd thing is it works for me, I suppose I should build the minimal tl2019 setup travis is using

@AlanMunn -- Lovely! (My husband, whom I met in a class on Old High German, will be visiting Iceland for New Year's. Now they had the sense to keep up the tradition.) Thanks!

@DavidCarlisle oh, I have a latex-dev in my path which is imho not quite up-to-date.

@DavidCarlisle when I remove the local stuff it still errors, but the error is different between xelatex and lualatex.

I wrote a generic package, just for fun, now if I put the following at the top of the file everything is fine:

\begingroup\expandafter\expandafter\expandafter\endgroup
\expandafter
\ifx\csname newdef\endcsname\relax
\else
  \errmessage{`\string\newdef' already defined. Aborting}
  \expandafter\endinput
\fi

But as soon as I put the following at the top, I get errors:

\begingroup\expandafter\expandafter\expandafter\endgroup
\expandafter
\ifx\csname eTeXversion\endcsname\relax
  \errmessage{newdef requires e-TeX's `\string\ifdefined'. Aborting}
  \expandafter\endinput
\fi

The error thrown is:

./newdef.tex:42: Extra \fi.
l.42     \fi

Lines 40 to 47 of the file:

\def\newdef@ifrelax#1\fi
  {%
    \fi
    \ifx\relax#1%
      \newdef@relax
    \fi
    \newdef@panic
  }

usual thing, yiou can't skip over \ifdefined but why are you using csname around eTeXversion ?

@DavidCarlisle I thought I could if it is in braces...

@Skillmon no \if...\fi matching doesn't see braces

@DavidCarlisle should I rather do \ifdefined\eTeXversion?

@DavidCarlisle interesting, didn't know that. That's a stupid behaviour, imho.

@Skillmon aren't you testing for if ifdefined is deifined? you can use \@backslashchar ifdefined in the message

@DavidCarlisle no, that wouldn't be generic anymore, but LaTeX kernel dependent :) But I have to do something like that...

\expandafter\string\csname ifdefined\endcsname (the one place you didn't use csname:-)

@DavidCarlisle yup :)

@DavidCarlisle no, I'm testing whether eTeX is available, because I need an expandable \ifdefined.

@DavidCarlisle Guess I could also test for \ifdefined, but that would require no stupid definition of \ifdefined is around.

@DavidCarlisle I think it is safer to test for eTeX with \eTeXversion.

@Skillmon you can not defend against such definitions. someone may have defined eTeXversion

@DavidCarlisle but that would be pretty useless. I think that is more unlikely than someone having defined \eTeXversion when not running eTeX. Though testing for eTeX nowadays is pretty useless anyway, as almost everybody is using it.

@marmot Ok there is no problem. Now I will remove last comments.

@DavidCarlisle thanks for your help!

@Skillmon tex is just about the only format in texlive without I think

@DavidCarlisle xmltex?

@JosephWright I'm in denial about that

@DavidCarlisle but isn't it completely bug free?

@Skillmon yes but also lacking any real use cases

@DavidCarlisle I thought there was some professional catalogue typesetting software using it.

@marmot Please, can you see if are there strange my comments now? Hi.

@Sebastiano Hi Sebastiano! Grazie! The comments are unfortunately still there. Please note that I do not want to convince you to do something you do not want to do.

@marmot not a single \begin{picture} in my last answer.

@marmot Don't worry, I won't be offended :-) I get angry when I see that they vote against my answers almost perfect and I do not understand the reason. Please enter the links. I've removed some of the links that don't have a positive comment.

@DavidCarlisle the tikz gold badge in action ;-)

@DavidCarlisle Nor in your last question.

@UlrikeFischer of course.

@DavidCarlisle Are you sure about \draw[fill] as opposed to \fill in that case?

@marmot does the first stroke and fill rather than just fill?

@Sebastiano It is impossible to understand voting behavior. The best strategy is to recall that the points are meaningless.

@DavidCarlisle Bingo! You really deserve the tikz gold badge. ;-)

@DavidCarlisle The thing is that the line width does not get transformed. So if you compare scale=1 and scale=2, the second one is only larger by a factor of 2 if you use \fill and not \draw[fill].

@marmot yes same as the old days with PostScript, fixed tex.stackexchange.com/posts/498928/revisions

@DavidCarlisle I saw it. But I upvoted before that because my crystal ball told me you would do it.

@egreg @DavidCarlisle always thank you for everything. Then I decided to write it like this. my warmest greetings.

@Sebastiano I'd put a \qquad before that + so it's not aligned with the =

@Sebastiano also do you have a blank line in the source before the equation ?

@DavidCarlisle David I add my code. Can you fix it with your comments?

\dimo Si osservi innanzitutto che $f_y'(x,y,y')$ è di classe $\mathcal{C}^1([a,b])$ in quanto, per ipotesi, $y=y(x)\in \mathcal{C}^1([a,b])$ soddisfa l'equazione di Eulero-Lagrange. Sia $\Delta x$ l'incremento della $x$, $\Delta y=y(x+\Delta x)-y(x)$ e $\Delta y'=y'(x+\Delta x)-y'(x)$. Fissato $x\in A$ essendo $f_y$ di classe $\mathcal{C}^1([a,b])$, essa è differenziaile nel punto $(x,y(x),y'(x))\in \mathcal{C}^2([a,b]\times\R^2)$, ciò significa che è possibile scrivere:


\begin{equation}\label{TEL1}

@StefanKottwitz Welcome into chat. Hi, did I take your advice well? :-)

I thought so, never put a blank line before display math otherwise you get the output you show, and as I say I'd use &\qquad + not &+

@Sebastiano You are still here, does it prove it? ;-)

@Sebastiano Thanks for the welcome!

@Sebastiano also I assume dove l'infinitesimo di ordin... is not the start of a paragraph so that should not have a blank line before it either

@StefanKottwitz They're still here. I took some magnesium citrate. I have SHBG above the norm.

@DavidCarlisle Wait till I send you all the complete code I have written. You are free to make the changes you want so that I can better understand what you would like to do.

@Sebastiano I don't need the full code it is obvious just from your image.

\documentclass[12pt,oneside]{book} % Default font size
\usepackage{structure-mod}
\begin{document}

\begin{theorem}
Sia $f(x,y,y')\in \mathcal{C}^2([a,b]\times\R^2)$ rispetto a tutte e tre le variabili; inoltre sia $y=y(x)\in \mathcal{C}^1([a,b])$ soluzione dell'equazione di Eulero associata ad $f$, sia cioè
\[f_y(x,y(x),y'(x))=\dfrac d{dx}f_{y'}(x,y(x),y'(x))\qquad\forall x\in[a,b].\]

Indicato poi con $A=\{x\in[a,b]\colon\,f_{y'y'}(x,y(x),y'(x))\ne0\}$, allora $y=y(x)$ è di classe $\mathcal{C}^2$ su $A$ e per ogni $x\in A$ si ha

@DavidCarlisle What exactly do I need to change, please? You could send me the correct answer. Thank you very much.

@Sebastiano it is always wrong to have a blank line before a display math equation. dove l'infinitesimo and dove l'infinitesimo d` doesn't start with a capital so I assume it is not the start of a sentence or start of a paragraph, and as I have said twice now I would indent the +

@Sebastiano There should never be a blank line before a display. Never in the sense of mai e poi mai per qualsivoglia ragione. A blank line can follow it if the text starts a new paragraph. It could be the case for “Indicato poi”, but probably it isn't.

@Sebastiano remove all the blank lines I do not know what else I can say to make that easier to understand

@egreg Into the structure.tex there is \parskipindent (any similar) =0.

@Sebastiano but just look at the image you posted it was obvious the markup was wrong just from the bad spacing.

@Sebastiano your image shows that your paragraph indent is not zero, dove is indented.

@Sebastiano That means nothing. Read again what I wrote and repeat with me: no blank line for whatsoever reason before a math display (unless you know precisely why you are having one, in very special situations, which is not your case).

@DavidCarlisle here the code:

%----------------------------------------------------------------------------------------
%	PACKAGES AND OTHER DOCUMENT CONFIGURATIONS
%----------------------------------------------------------------------------------------

\documentclass[12pt,oneside]{book} % Default font size
\usepackage{structure-mod}
\begin{document}

\begin{theorem}
Sia $f(x,y,y')\in \mathcal{C}^2([a,b]\times\R^2)$ rispetto a tutte e tre le variabili; inoltre sia $y=y(x)\in \mathcal{C}^1([a,b])$ soluzione dell'equazione di Eulero associata ad $f$, sia cioè

@egreg I have removed the blank lines before a math display. Done!

@Sebastiano you still haven't indented the +

@Sebastiano Why aren't you using the standard \begin{proof}...\end{proof}? No blank line before \end{proof}, of course.

I should do that in my textbook, too. I have an old 32-bit W7 pc and very slow. So I often work in pieces. I have to remember to delete the empty lines in relation to your warning and David's.

@Sebastiano you don't have to remember just look at the output, if it pokes you in the eye then the markup is wrong...

@DavidCarlisle now there is: \qquad+

@Sebastiano shouldn't there be a full stop at the and? (there would be in English I think)

@DavidCarlisle Of course, but I still have to complete the theorem. :-)

@egreg You're absolutely right. When I started writing the book in April 2017 I used legrand orange book and I added some parts, if I remember correctly, of a famous colleague of yours from the University of Padua who had been retired for 3 years. Wait, let's see if I find the part of how I write a theorem.

@Sebastiano Giuseppe DM?

@egreg Siiiiiiiiiiiiiiiiiiii :)

%----------------------------------------------------------------------------------
%	DeclareMathOperator
%----------------------------------------------------------------------------------
\def\a{\alpha}
\def\be{\beta}
\def\men{\smallsetminus}
\def\ve{\varepsilon}

\def\xy{(x,y)}
\newcommand\xyz{(x,y,z)}
\def\xyzt{(x,y,z,t)}
\def\axyzt{(x',y',z',t')}

\def\vf{\varphi}
\def\z{\zeta}
\def\vt{\vartheta}
\def\la{\lambda}
\def\s{\sigma}
\def\de{\delta}
\def\ga{\gamma}
\def\om{\omega}

\newcommand{\field}[1]{\mathbb{#1}}

@Sebastiano He was my teacher of general topology. Great teacher and good friend.

he's a great person like you. We've known each other, not personally, but by email and phone since 2002.

@Sebastiano Convert every \def into \newcommand as soon as possible and change name to the already defined macros.

@egreg Here there is your structure:

\documentclass{amsart}
\usepackage{amsmath, amsthm}
\usepackage{amscd}
\usepackage{verbatim}
\usepackage{amssymb}
\usepackage[italian]{babel}
\usepackage[utf8]{inputenc}
\usepackage{textcomp}
\usepackage{pstricks,pst-math,pstricks-add, subfigure}
%\usepackage{boxedeps}
\usepackage{epsfig}
\usepackage{epsf}
\graphicspath{ {img/} }
\input xy
\xyoption{all}
%\input LabelFig.tex
%\renewcommand{\figurename}{Figura}
%\SetTexturesEPSFSpecial\HideFigureFrames
%\usepackage[pdftex]{graphicx}
%\usepackage[dvips]{xcolor}

@Sebastiano If you use babel, you don't need to \renewcommand{\proofname}. Just use \begin{proof} and \end{proof}. For “solutions”: \newenvironment{solution}{\begin{proof}[Risoluzione]}{\end{proof}}

Here there is the part of my structure.tex

%----------------------------------------------------------------------------------
%	THEOREM STYLES
%----------------------------------------------------------------------------------

\usepackage{amsmath,amsfonts,amssymb,amsthm} % For math equations, theorems, symbols, etc

\newcommand{\intoo}[2]{\mathopen{]}#1\,;#2\mathclose{[}}
\newcommand{\ud}{\mathop{\mathrm{{}d}}\mathopen{}}
\newcommand{\intff}[2]{\mathopen{[}#1\,;#2\mathclose{]}}
\newtheorem{notation}{Notazione}[chapter]

% Boxed/framed environments

@egreg I must delete this?

with \newenvironment{solution}{\begin{proof}[Risoluzione]}{\end{proof}}?

@Sebastiano You gain nothing by not using the proper \begin and \end tags.

@Sebastiano don't use \def unless you know what you are doing, \def\a{\alpha} breaks the standard \a command

@Sebastiano son't use either of these \usepackage{epsfig}\usepackage{epsf}

@egreg @DavidCarlisle I have changhed \def with \newcommand now

%----------------------------------------------------------------------------------
%	DeclareMathOperator
%----------------------------------------------------------------------------------
\newcommand{\a}{\alpha}
\newcommand{\be}{\beta}
\newcommand{\men}{\smallsetminus}
\newcommand{\ve}{\varepsilon}

\newcommand{\xy}{(x,y)}
\newcommand{\xyz}{(x,y,z)}
\newcommand{\xyzt}{(x,y,z,t)}
\newcommand{\axyzt}{(x',y',z',t')}
\newcommand{\vf}{\varphi}
\newcommand{\z}{\zeta}
\newcommand{\vt}{\vartheta}
\newcommand{\la}{\lambda}

@Sebastiano but now you will get errors on \a etc, that is the point of newcommand, to warn you

@Sebastiano \def\SS{\field{S}}\def\I{\field{I}}\def\H{\field{H}} breaks German

@Sebastiano do't use \bf, \it, \sc

@DavidCarlisle \usepackage{epsfig} and \usepackage{epsf} are into a old structure of a friend of egreg and mine. They are not into my structure. I have changed \def - see the previous code.

@Sebastiano you said the file that you showed was the file you were using. In which case it should not have epsf or epsfig

@DavidCarlisle The code:

\documentclass{amsart}
\usepackage{amsmath, amsthm}
\usepackage{amscd}
\usepackage{verbatim}
\usepackage{amssymb}
\usepackage[italian]{babel}
\usepackage[utf8]{inputenc}
\usepackage{textcomp}
\usepackage{pstricks,pst-math,pstricks-add, subfigure}
%\usepackage{boxedeps}

is of the our friend :-). I have not use it.

@Sebastiano I can not understand you at all, why are you postingf fragments of tex if they are not from your document?

@Sebastiano this will also give an error (have you tried the file at all??) \newcommand{\r}{\sqrt{x^2+y^2}}

@DavidCarlisle Because I have responded to egreg because there is still definition and not newcommand. I wrote that I had taken only a part of this friend of ours and not all his whole code.

@DavidCarlisle No :( still. I have not very quick.

@DavidCarlisle :-(

! LaTeX Error: Command \a already defined.
               Or name \end... illegal, see p.192 of the manual.

See the LaTeX manual or LaTeX Companion for explanation.
Type  H <return>  for immediate help.
 ...

l.324 \newcommand{\a}{\alpha}

?

@Sebastiano most one-letter commands like \r are pre-defined, so if you use \def you change them with no warning but latex can break in unxepected ways, if you simply change def to newcommand then you get errors at the definition. \r is ring accent so \def\r{\sqrt{x^2+y^2}}` means you can not have an Angstrom symbil

@Sebastiano yes of course that is the main function of \newcommand to stop you accidentally redefining \a and breaking latex.

@DavidCarlisle @egreg Thus I have not "Dimostrazione" (\dimo) and at the end the square of (\findim).

@Sebastiano the amsthm begin{proof}...end{proof} adds the final square by default

@DavidCarlisle I have this:

%----------------------------------------------------------------------------------
%	DeclareMathOperator
%----------------------------------------------------------------------------------
\def\a{\alpha}
\newcommand{\be}{\beta}
\newcommand{\men}{\smallsetminus}
\newcommand{\ve}{\varepsilon}

\newcommand{\xy}{(x,y)}
\newcommand{\xyz}{(x,y,z)}
\newcommand{\xyzt}{(x,y,z,t)}
\newcommand{\axyzt}{(x',y',z',t')}
\newcommand{\vf}{\varphi}
\newcommand{\z}{\zeta}
\newcommand{\vt}{\vartheta}
\newcommand{\la}{\lambda}

and this:

\def\a, \def\SS \def\r` are all wrong

%----------------------------------------------------------------------------------
%	THEOREM STYLES
%----------------------------------------------------------------------------------

\usepackage{amsmath,amsfonts,amssymb,amsthm} % For math equations, theorems, symbols, etc

\newcommand{\intoo}[2]{\mathopen{]}#1\,;#2\mathclose{[}}
\newcommand{\ud}{\mathop{\mathrm{{}d}}\mathopen{}}
\newcommand{\intff}[2]{\mathopen{[}#1\,;#2\mathclose{]}}
\newtheorem{notation}{Notazione}[chapter]

% Boxed/framed environments

@DavidCarlisle What am I to write then? Excuse me :-(

@Sebastiano use some other name (any other name) try this document

\documentclass{article}
\def\r{\sqrt{x^2+y^2}}
\def\SS{\field{S}}

\begin{document}

hmm Ångstrom

hmmm \MakeUppercase{straße}

\end{document}

@DavidCarlisle language skills ;-)

@DavidCarlisle You have right: Missing chars Å,ß. Into my book there are many \r and no \SS

Ulrike hi

@Sebastiano not missing chars!! don't look at the output, look at the log

! Missing $ inserted.
<inserted text>
                $
l.7 hmm Å
          ngstrom
?
! Missing $ inserted.
<inserted text>
                $
l.8

?
! Undefined control sequence.
\ss ->\field
             {S}
l.9 hmmm \MakeUppercase{straße}

I have understood

@UlrikeFischer Ich spreche fließend Deutsch

@DavidCarlisle und schwedisch.

@Sebastiano if you use \def then you can over-write any part of latex and you will get no warning until later when you try to use the broken part

@DavidCarlisle Ok. I'm trying to do my best, I can assure you. I'm going to go to bed in a bit. Tomorrow morning I will wake up early to go to sea. It is expected to be 40 degrees and there will be a mass of people already around 9 am Italian time.

I always get the value of this goddamn SHBG back up. I should rest mentally and physically :-( I send you and I send you my most sincere greetings. Thanks for everything.