Re: Accessibility questions that are less technical

David Carlisle <[email protected]> Wed, 28 Jan 2026 12:15:37 +0000
Newsgroups gmane.comp.tex.texhax
Message-ID <CAEW6iOi5BmzTK+AwAy66LesAzp8ksKGybcqndyGkAOareOvhUw@mail.gmail.com>
--0000000000005760b5064971b34c
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

not sure this is the best list for discussing this but that file is
somewhat broken  show-pdf-tags can't open it, verapdf complains, and pac
similarly gives a fatal error "EMC has no closing BMC or BMD, pdf document
will be closed"

I could try the tex source later, but you may do better with lualatex

certainly for mathematics if you specify (as this file does) that it is
UA-2 then you must supply MathML for every math expression
LaTex can only automatically do that  with luatex. So if using pdftex you
need to supply the mathml, or target UA-1/PDF 1.7



On Wed, 28 Jan 2026 at 11:49, Jim Hefferon <[email protected]> wrote:

> Thanks again for the replies.  There was some interest when I said this:
>
> 5) There is sometimes a tradeoff between a presentation that is useful
> for visually-impaired users and others.  To again name a concrete
> example from ltx-talk, I sometimes show a sequence of graphics as
> slides inside of a frame to illustrate changes (say, developing arc
> length as line segments become shorter).  Only the graphic changes,
> not the surrounding text.  I understand screen readers to re-read the
> entire slide; certainly Okular does.  That's pretty tedious after five
> slides.  Does anyone have advice about working through that?
>
> Folks told me that the answer is that the reader I was using, Okular
> on Linux, is not sufficient.  But I got some questions so just to make
> the example concrete, I uploaded a file to my website
> https://hefferon.net/review_integration.pdf that shows the behavior.
> FWIW, it uses ltx-talk.cls and the source is
>
> https://gitlab.com/jim.hefferon/calc-ii-talks/-/blob/master/review_integr=
ation/review_integration.tex?ref_type=3Dheads
> .
>   Under Okular, if I ask it to read the whole document then it reads
> the frame headed "Prototype" a new time for each slide, which is a lot
> of times.
>
> Best,
> Jim
>
> On Tue, Jan 27, 2026 at 3:01=E2=80=AFPM Jim Hefferon <jim.hefferon@gmail.=
com>
> wrote:
> >
> > Thank you Ulrike and David.  Very helpful replies.  For sure the video
> > acts exactly like I want the doc to act, which is not what I was
> > getting from Okular.  I'll look into all of the suggestions.  (My
> > Firefox is #147; I couldn't immediately get orca to work with it but
> > I'll look more.)
> >
> > I should try the blind-mathematicians list again; for myself, I know
> > I've lost emails before now.
> >
> > Perhaps there are other people interested in discussing and sharing
> > accessibility more as users than as developers?  If there is
> > significant interest then we could figure out a convenient way.
> >
> > Regards,
> > Jim
> >
> > On Tue, Jan 27, 2026 at 9:44=E2=80=AFAM David Carlisle <d.p.carlisle@gm=
ail.com>
> wrote:
> > >
> > >
> > >
> > > On Tue, 27 Jan 2026 at 13:36, Jim Hefferon <[email protected]>
> wrote:
> > >>
> > >> Hello.  I have a number of questions that are perhaps less technical
> > >> than most posts here.  I apologize; they seem too soft SE or latex3'=
s
> > >> GitHub list, the tug.org accessibility list does not have substantiv=
e
> > >> traffic, and the blind-mathematicians mailing list did not accept my
> > >> email. I hope folks here don't mind.
> > >>
> > >> I intend to re-make a book of mine to be accessible and I'm struggli=
ng
> > >> with some things that arose out of some preliminary, practice, work
> > >> that I did.
> > >>
> > >> 1) At the bottom is an MWE document (I have today's TeX Live and I'm
> > >> using LuaLaTeX).  I do not currently have access to Windows so the
> > >> NVDA screen reader is not an option.  On Linux, Okular does not read
> > >> the math correctly
> > >
> > >
> > > You might do better to target pdf 1.7 and UA-1 rather than UA-2
> > > or at least if using UA-2 add an Alt to the Formula (which is the
> default
> > > with UA-1) by adding
> > >
> > > tagging-setup=3D{ math/alt/use,math/setup=3Dmathml-SE}
> > >
> > > alt/use specifies that LaTeX should add an Alt property to every
> Formula (with the TeX source of that expression)
> > >
> > > this is not as good as tagging with mathml, but better than nothing.
> > >
> > > Note however that what matters most is the screen readers used by you=
r
> (human) not the one you have access to, so that may be nvda or jaws on
> windows.
> > >
> > > You could try orca with  firefox 148 or higher (currently in beta)
> which has support for tagged pdf, although
> > > I haven't had great success with orca so far.
> > >
> > >
> > >
> > >>
> > >> and does not read the alt text or "Black bird." at
> > >> all.  Are there better screen reader options?  I'd sure like to do t=
he
> > >> highest PDF standard.  At the moment I am developing my docs without
> > >> being able to know what is the outcome and it is awkward.
> > >
> > >
> > >
> > > I can recommend using ngpdf  https://ngpdf.com/loadFile to test your
> file, that will show the tag structure
> > > and also the derived HTML file, which is a reasonable representation
> of the accessible view of the document.
> > >
> > >
> > >>
> > >>  (I have
> > >> not tried converting to HTML and using an HTML reader.  I hope to po=
st
> > >> the PDF for free download.)
> > >>
> > >> 2) Can alt text contain math?  If not, I must write "x times y" and
> > >> "the sine function"?
> > >
> > >
> > >
> > > As noted above latex will add the tex source as alt text if you
> specify pdf 1.7 and UA-1 or if you explicitly set
> > > math/alt/use key to true.
> > >
> > >
> > >>
> > >>
> > >> 3) Is there any way to influence screen readers outside of textual?
> > >> To give a concrete example, I would like to get a pause after
> > >> execution of ltx-talk's \frametitle{..} command.
> > >
> > >
> > > if the title is marked up as a title you would hope the screen reader
> announces it as such and adds a suitable pause.
> > >
> > >
> > >>
> > >>   I see that the
> > >> package's documentation recommends periods in the title but in the t=
he
> > >> typographic tradition that I am used to, headlines don't have period=
s.
> > >> Is this just a matter of hoping that screen readers improve over tim=
e?
> > >
> > >
> > > Check the tagging of the title, if it is tagged as a heading it ought
> to be automatically set off from the following text.
> > >
> > >>
> > >>
> > >> 4) Is there some best practices document that folks could recommend?
> > >> For example, googling for guidelines on writing alt text gives many
> > >> results but even the most high-profile ones seem to me to be both
> > >> vague and contradictory.
> > >
> > >
> > > vague and contradictory unfortunately describes many of the documents
> in this area
> > > both the standards themselves and tutorials from various places.
> Things are getting better though.
> > >
> > >> For instance, some advocate a character
> > >> limit while others recommend extensive description. I confess that I
> > >> find it impossible to briefly describe a mathematical graph of the
> > >> area under f(x)=3Dsin(x) between x=3Da and x=3Db broken into ten Rei=
mann
> > >> boxes, particularly without mathematical symbols.
> > >
> > >
> > > well exactly. You can use unicode math symbols in an alt text but not
> math layout
> > > but mostly what you lose is navigation. You can't really follow a ful=
l
> math display
> > > in a "one take" reading of an alt text whatever text you place in the
> attribute.
> > > Using mathml allows the reader to navigate into the expression,
> > > reading subterms at their pace and in their order. But of course if
> you are using a screen reader that doesn't understand the tagging that's
> just a theoretical benefit, and using Alt is all you have.
> > >
> > >>
> > >> In general, the
> > >> advice I have seen online seems more directed to business documents
> > >> than to documents for math or science.
> > >
> > >
> > > that and often assuming that the tagging is being done by hand by
> remediating the document after production rather than being added as part
> of the typesetting as in LaTeX
> > >>
> > >>
> > >> 5) There is sometimes a tradeoff between what is useful for
> > >> visually-impaired users and others.  To again name a concrete exampl=
e
> > >> from ltx-talk, I sometimes show a sequence of graphics as slides
> > >> inside of a frame to illustrate changes (say, developing the arc
> > >> length as a limit by showing line segments become shorter).  Only th=
e
> > >> graphic changes, not the surrounding text.  I understand screen
> > >> readers to re-read the entire slide; certainly Okular does.
> > >
> > >
> > > That sounds wrong, the default behaviour is just to read the last
> slide (which is what you want for gradual reveals)
> > > If you have more complicated interactions with text coming and going
> it has some  possibility to control which text
> > > gets marked to be read and which is marked as Artifact.
> > >
> > >> That's
> > >> pretty tedious after five slides.  Does anyone have advice about
> > >> working through that?
> > >
> > >
> > > It's not supposed to happen.
> > >
> > >>
> > >>
> > >> Thank you,
> > >> Jim Hefferon
> > >>
> > >> ................ Example document .............
> > >
> > >
> > > A video of your document unchanged being read by nvda (from foxit) is
> at
> > >
> > >
> https://www.dropbox.com/scl/fi/wbyumxcv16wk8vozlox9f/jh1.mp4?rlkey=3Dqrmm=
rykcdv964zyvrpakvsy3m&st=3D0igpgs0v&dl=3D0
> > >
> > >
> > >>
> > >>
> > >> \DocumentMetadata{
> > >>   lang        =3D en,
> > >>   pdfstandard =3D ua-2,
> > >>   pdfstandard =3D a-4f, %or a-4
> > >>   tagging=3Don,
> > >>   tagging-setup=3D{math/setup=3Dmathml-SE}
> > >> }
> > >> \documentclass{article}
> > >> \usepackage{unicode-math}
> > >> \usepackage{graphicx}
> > >> \usepackage{mathtools}
> > >>
> > >> \begin{document}
> > >> Red dog.
> > >> \begin{equation*}
> > >>   \int_{x=3D0}^1 x^2\,dx=3D\pi
> > >> \end{equation*}
> > >> Blue fish.
> > >> \begin{center}
> > >>   % \includegraphics[alt=3D{Show the integral $\int x^2 dx$ as a
> > >> graph.}]{example-image-a}
> > >>   \includegraphics[alt=3D{Show the integral as a
> graph.}]{example-image-a}
> > >> \end{center}
> > >> Black bird.
> > >> \end{document}
> > >>
> > >> --
> > >> "Does not the Captain seek your advice sir?"
> > >> "Not always," said Stephen.
> >
> >
> >
> > --
> > "Does not the Captain seek your advice sir?"
> > "Not always," said Stephen.
>
>
>
> --
> "Does not the Captain seek your advice sir?"
> "Not always," said Stephen.
>
>

--0000000000005760b5064971b34c
Content-Type: text/html; charset="UTF-8"
Content-Transfer-Encoding: quoted-printable

<div dir=3D"ltr"><div>not sure this is the best list for discussing this bu=
t that file is somewhat broken=C2=A0 show-pdf-tags can&#39;t open it, verap=
df complains, and pac similarly gives a fatal error &quot;EMC has no closin=
g BMC or BMD, pdf document will be closed&quot;</div><div><br></div><div>I =
could try the tex source later, but you may do better with lualatex</div><d=
iv><br></div><div>certainly for mathematics if you specify (as this file do=
es) that it is UA-2 then you must supply MathML for every math expression</=
div><div>LaTex can only automatically do that=C2=A0 with luatex. So if usin=
g pdftex you need to supply the mathml, or target UA-1/PDF 1.7</div><div><b=
r></div><div><br></div></div><br><div class=3D"gmail_quote gmail_quote_cont=
ainer"><div dir=3D"ltr" class=3D"gmail_attr">On Wed, 28 Jan 2026 at 11:49, =
Jim Hefferon &lt;<a href=3D"mailto:[email protected]">jim.hefferon@gma=
il.com</a>&gt; wrote:<br></div><blockquote class=3D"gmail_quote" style=3D"m=
argin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left=
:1ex">Thanks again for the replies.=C2=A0 There was some interest when I sa=
id this:<br>
<br>
5) There is sometimes a tradeoff between a presentation that is useful<br>
for visually-impaired users and others.=C2=A0 To again name a concrete<br>
example from ltx-talk, I sometimes show a sequence of graphics as<br>
slides inside of a frame to illustrate changes (say, developing arc<br>
length as line segments become shorter).=C2=A0 Only the graphic changes,<br=
>
not the surrounding text.=C2=A0 I understand screen readers to re-read the<=
br>
entire slide; certainly Okular does.=C2=A0 That&#39;s pretty tedious after =
five<br>
slides.=C2=A0 Does anyone have advice about working through that?<br>
<br>
Folks told me that the answer is that the reader I was using, Okular<br>
on Linux, is not sufficient.=C2=A0 But I got some questions so just to make=
<br>
the example concrete, I uploaded a file to my website<br>
<a href=3D"https://hefferon.net/review_integration.pdf" rel=3D"noreferrer" =
target=3D"_blank">https://hefferon.net/review_integration.pdf</a> that show=
s the behavior.<br>
FWIW, it uses ltx-talk.cls and the source is<br>
<a href=3D"https://gitlab.com/jim.hefferon/calc-ii-talks/-/blob/master/revi=
ew_integration/review_integration.tex?ref_type=3Dheads" rel=3D"noreferrer" =
target=3D"_blank">https://gitlab.com/jim.hefferon/calc-ii-talks/-/blob/mast=
er/review_integration/review_integration.tex?ref_type=3Dheads</a>.<br>
=C2=A0 Under Okular, if I ask it to read the whole document then it reads<b=
r>
the frame headed &quot;Prototype&quot; a new time for each slide, which is =
a lot<br>
of times.<br>
<br>
Best,<br>
Jim<br>
<br>
On Tue, Jan 27, 2026 at 3:01=E2=80=AFPM Jim Hefferon &lt;<a href=3D"mailto:=
[email protected]" target=3D"_blank">[email protected]</a>&gt; wr=
ote:<br>
&gt;<br>
&gt; Thank you Ulrike and David.=C2=A0 Very helpful replies.=C2=A0 For sure=
 the video<br>
&gt; acts exactly like I want the doc to act, which is not what I was<br>
&gt; getting from Okular.=C2=A0 I&#39;ll look into all of the suggestions.=
=C2=A0 (My<br>
&gt; Firefox is #147; I couldn&#39;t immediately get orca to work with it b=
ut<br>
&gt; I&#39;ll look more.)<br>
&gt;<br>
&gt; I should try the blind-mathematicians list again; for myself, I know<b=
r>
&gt; I&#39;ve lost emails before now.<br>
&gt;<br>
&gt; Perhaps there are other people interested in discussing and sharing<br=
>
&gt; accessibility more as users than as developers?=C2=A0 If there is<br>
&gt; significant interest then we could figure out a convenient way.<br>
&gt;<br>
&gt; Regards,<br>
&gt; Jim<br>
&gt;<br>
&gt; On Tue, Jan 27, 2026 at 9:44=E2=80=AFAM David Carlisle &lt;<a href=3D"=
mailto:[email protected]" target=3D"_blank">[email protected]</a>=
&gt; wrote:<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; On Tue, 27 Jan 2026 at 13:36, Jim Hefferon &lt;<a href=3D"mailto:=
[email protected]" target=3D"_blank">[email protected]</a>&gt; wr=
ote:<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; Hello.=C2=A0 I have a number of questions that are perhaps le=
ss technical<br>
&gt; &gt;&gt; than most posts here.=C2=A0 I apologize; they seem too soft S=
E or latex3&#39;s<br>
&gt; &gt;&gt; GitHub list, the <a href=3D"http://tug.org" rel=3D"noreferrer=
" target=3D"_blank">tug.org</a> accessibility list does not have substantiv=
e<br>
&gt; &gt;&gt; traffic, and the blind-mathematicians mailing list did not ac=
cept my<br>
&gt; &gt;&gt; email. I hope folks here don&#39;t mind.<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; I intend to re-make a book of mine to be accessible and I&#39=
;m struggling<br>
&gt; &gt;&gt; with some things that arose out of some preliminary, practice=
, work<br>
&gt; &gt;&gt; that I did.<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; 1) At the bottom is an MWE document (I have today&#39;s TeX L=
ive and I&#39;m<br>
&gt; &gt;&gt; using LuaLaTeX).=C2=A0 I do not currently have access to Wind=
ows so the<br>
&gt; &gt;&gt; NVDA screen reader is not an option.=C2=A0 On Linux, Okular d=
oes not read<br>
&gt; &gt;&gt; the math correctly<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; You might do better to target pdf 1.7 and UA-1 rather than UA-2<b=
r>
&gt; &gt; or at least if using UA-2 add an Alt to the Formula (which is the=
 default<br>
&gt; &gt; with UA-1) by adding<br>
&gt; &gt;<br>
&gt; &gt; tagging-setup=3D{ math/alt/use,math/setup=3Dmathml-SE}<br>
&gt; &gt;<br>
&gt; &gt; alt/use specifies that LaTeX should add an Alt property to every =
Formula (with the TeX source of that expression)<br>
&gt; &gt;<br>
&gt; &gt; this is not as good as tagging with mathml, but better than nothi=
ng.<br>
&gt; &gt;<br>
&gt; &gt; Note however that what matters most is the screen readers used by=
 your (human) not the one you have access to, so that may be nvda or jaws o=
n windows.<br>
&gt; &gt;<br>
&gt; &gt; You could try orca with=C2=A0 firefox 148 or higher (currently in=
 beta) which has support for tagged pdf, although<br>
&gt; &gt; I haven&#39;t had great success with orca so far.<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; and does not read the alt text or &quot;Black bird.&quot; at<=
br>
&gt; &gt;&gt; all.=C2=A0 Are there better screen reader options?=C2=A0 I&#3=
9;d sure like to do the<br>
&gt; &gt;&gt; highest PDF standard.=C2=A0 At the moment I am developing my =
docs without<br>
&gt; &gt;&gt; being able to know what is the outcome and it is awkward.<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; I can recommend using ngpdf=C2=A0 <a href=3D"https://ngpdf.com/lo=
adFile" rel=3D"noreferrer" target=3D"_blank">https://ngpdf.com/loadFile</a>=
 to test your file, that will show the tag structure<br>
&gt; &gt; and also the derived HTML file, which is a reasonable representat=
ion of the accessible view of the document.<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt;=C2=A0 (I have<br>
&gt; &gt;&gt; not tried converting to HTML and using an HTML reader.=C2=A0 =
I hope to post<br>
&gt; &gt;&gt; the PDF for free download.)<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; 2) Can alt text contain math?=C2=A0 If not, I must write &quo=
t;x times y&quot; and<br>
&gt; &gt;&gt; &quot;the sine function&quot;?<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; As noted above latex will add the tex source as alt text if you s=
pecify pdf 1.7 and UA-1 or if you explicitly set<br>
&gt; &gt; math/alt/use key to true.<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; 3) Is there any way to influence screen readers outside of te=
xtual?<br>
&gt; &gt;&gt; To give a concrete example, I would like to get a pause after=
<br>
&gt; &gt;&gt; execution of ltx-talk&#39;s \frametitle{..} command.<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; if the title is marked up as a title you would hope the screen re=
ader announces it as such and adds a suitable pause.<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt;=C2=A0 =C2=A0I see that the<br>
&gt; &gt;&gt; package&#39;s documentation recommends periods in the title b=
ut in the the<br>
&gt; &gt;&gt; typographic tradition that I am used to, headlines don&#39;t =
have periods.<br>
&gt; &gt;&gt; Is this just a matter of hoping that screen readers improve o=
ver time?<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; Check the tagging of the title, if it is tagged as a heading it o=
ught to be automatically set off from the following text.<br>
&gt; &gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; 4) Is there some best practices document that folks could rec=
ommend?<br>
&gt; &gt;&gt; For example, googling for guidelines on writing alt text give=
s many<br>
&gt; &gt;&gt; results but even the most high-profile ones seem to me to be =
both<br>
&gt; &gt;&gt; vague and contradictory.<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; vague and contradictory unfortunately describes many of the docum=
ents in this area<br>
&gt; &gt; both the standards themselves and tutorials from various places. =
Things are getting better though.<br>
&gt; &gt;<br>
&gt; &gt;&gt; For instance, some advocate a character<br>
&gt; &gt;&gt; limit while others recommend extensive description. I confess=
 that I<br>
&gt; &gt;&gt; find it impossible to briefly describe a mathematical graph o=
f the<br>
&gt; &gt;&gt; area under f(x)=3Dsin(x) between x=3Da and x=3Db broken into =
ten Reimann<br>
&gt; &gt;&gt; boxes, particularly without mathematical symbols.<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; well exactly. You can use unicode math symbols in an alt text but=
 not math layout<br>
&gt; &gt; but mostly what you lose is navigation. You can&#39;t really foll=
ow a full math display<br>
&gt; &gt; in a &quot;one take&quot; reading of an alt text whatever text yo=
u place in the attribute.<br>
&gt; &gt; Using mathml allows the reader to navigate into the expression,<b=
r>
&gt; &gt; reading subterms at their pace and in their order. But of course =
if you are using a screen reader that doesn&#39;t understand the tagging th=
at&#39;s just a theoretical benefit, and using Alt is all you have.<br>
&gt; &gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; In general, the<br>
&gt; &gt;&gt; advice I have seen online seems more directed to business doc=
uments<br>
&gt; &gt;&gt; than to documents for math or science.<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; that and often assuming that the tagging is being done by hand by=
 remediating the document after production rather than being added as part =
of the typesetting as in LaTeX<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; 5) There is sometimes a tradeoff between what is useful for<b=
r>
&gt; &gt;&gt; visually-impaired users and others.=C2=A0 To again name a con=
crete example<br>
&gt; &gt;&gt; from ltx-talk, I sometimes show a sequence of graphics as sli=
des<br>
&gt; &gt;&gt; inside of a frame to illustrate changes (say, developing the =
arc<br>
&gt; &gt;&gt; length as a limit by showing line segments become shorter).=
=C2=A0 Only the<br>
&gt; &gt;&gt; graphic changes, not the surrounding text.=C2=A0 I understand=
 screen<br>
&gt; &gt;&gt; readers to re-read the entire slide; certainly Okular does.<b=
r>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; That sounds wrong, the default behaviour is just to read the last=
 slide (which is what you want for gradual reveals)<br>
&gt; &gt; If you have more complicated interactions with text coming and go=
ing it has some=C2=A0 possibility to control which text<br>
&gt; &gt; gets marked to be read and which is marked as Artifact.<br>
&gt; &gt;<br>
&gt; &gt;&gt; That&#39;s<br>
&gt; &gt;&gt; pretty tedious after five slides.=C2=A0 Does anyone have advi=
ce about<br>
&gt; &gt;&gt; working through that?<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; It&#39;s not supposed to happen.<br>
&gt; &gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; Thank you,<br>
&gt; &gt;&gt; Jim Hefferon<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; ................ Example document .............<br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt; A video of your document unchanged being read by nvda (from foxit=
) is at<br>
&gt; &gt;<br>
&gt; &gt; <a href=3D"https://www.dropbox.com/scl/fi/wbyumxcv16wk8vozlox9f/j=
h1.mp4?rlkey=3Dqrmmrykcdv964zyvrpakvsy3m&amp;st=3D0igpgs0v&amp;dl=3D0" rel=
=3D"noreferrer" target=3D"_blank">https://www.dropbox.com/scl/fi/wbyumxcv16=
wk8vozlox9f/jh1.mp4?rlkey=3Dqrmmrykcdv964zyvrpakvsy3m&amp;st=3D0igpgs0v&amp=
;dl=3D0</a><br>
&gt; &gt;<br>
&gt; &gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; \DocumentMetadata{<br>
&gt; &gt;&gt;=C2=A0 =C2=A0lang=C2=A0 =C2=A0 =C2=A0 =C2=A0 =3D en,<br>
&gt; &gt;&gt;=C2=A0 =C2=A0pdfstandard =3D ua-2,<br>
&gt; &gt;&gt;=C2=A0 =C2=A0pdfstandard =3D a-4f, %or a-4<br>
&gt; &gt;&gt;=C2=A0 =C2=A0tagging=3Don,<br>
&gt; &gt;&gt;=C2=A0 =C2=A0tagging-setup=3D{math/setup=3Dmathml-SE}<br>
&gt; &gt;&gt; }<br>
&gt; &gt;&gt; \documentclass{article}<br>
&gt; &gt;&gt; \usepackage{unicode-math}<br>
&gt; &gt;&gt; \usepackage{graphicx}<br>
&gt; &gt;&gt; \usepackage{mathtools}<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; \begin{document}<br>
&gt; &gt;&gt; Red dog.<br>
&gt; &gt;&gt; \begin{equation*}<br>
&gt; &gt;&gt;=C2=A0 =C2=A0\int_{x=3D0}^1 x^2\,dx=3D\pi<br>
&gt; &gt;&gt; \end{equation*}<br>
&gt; &gt;&gt; Blue fish.<br>
&gt; &gt;&gt; \begin{center}<br>
&gt; &gt;&gt;=C2=A0 =C2=A0% \includegraphics[alt=3D{Show the integral $\int=
 x^2 dx$ as a<br>
&gt; &gt;&gt; graph.}]{example-image-a}<br>
&gt; &gt;&gt;=C2=A0 =C2=A0\includegraphics[alt=3D{Show the integral as a gr=
aph.}]{example-image-a}<br>
&gt; &gt;&gt; \end{center}<br>
&gt; &gt;&gt; Black bird.<br>
&gt; &gt;&gt; \end{document}<br>
&gt; &gt;&gt;<br>
&gt; &gt;&gt; --<br>
&gt; &gt;&gt; &quot;Does not the Captain seek your advice sir?&quot;<br>
&gt; &gt;&gt; &quot;Not always,&quot; said Stephen.<br>
&gt;<br>
&gt;<br>
&gt;<br>
&gt; --<br>
&gt; &quot;Does not the Captain seek your advice sir?&quot;<br>
&gt; &quot;Not always,&quot; said Stephen.<br>
<br>
<br>
<br>
-- <br>
&quot;Does not the Captain seek your advice sir?&quot;<br>
&quot;Not always,&quot; said Stephen.<br>
<br>
</blockquote></div>

--0000000000005760b5064971b34c--