Ok that's my try:
\documentclass{beamer}
\usepackage{tikz, etoolbox}
\usepackage{etoolbox}
\newcount\mycount
\def\mytempcoord{}
\def\mytempopa{}
%
\newcommand\doatpos[3]{%
\bgroup
\renewcommand*{\do}[1]{\ifnum\mycount<2
\edef\mytempcoord{\mytempcoord ##1%
\ifnum\mycount<1
,
\fi
}
\else
\edef\mytempopa{##1}
\fi
\advance\mycount by 1
}
\docsvlist{#2}\global\advance\mycount by -3
\begin{tikzpicture}[remember picture, overlay]
\node[anchor=#1, opacity=\mytempopa] at (\mytempcoord) {#3};
\end{tikzpicture}%
%
%
\egroup
%
}
\begin{document}
\frame{
\doatpos{west}{0cm, 3cm,.5}{This is text \rule{1cm}{1cm}}
\doatpos{west}{1cm, -3 cm,.7}{This is text a\rule{1cm}{1cm}}
}
\end{document}
I basically wrote a parser which divides this three item comma list in such a way that it can be used by an ordinary sequence like my TikZ overlay thing here.
You might find that your third argument, the pseudo z-variable, needs no dimension. Well yes. Involving a certain key handler here could be useful.
However, my opinion is different. I rather would introduce a new argument (in braces). That would make things much easier. (My very personal view is that mimicking the behavior of completely different systems is rather pointless.)

Same output, but no etoolbox
:
\documentclass{beamer}
\usepackage{tikz}
\def\mytemparray{}
%
\newcommand\doatpos[3]{%
\bgroup
\def\mytemparray{{ #2 }}
\pgfmathparse{\mytemparray[0]} \edef\mya{\pgfmathresult}
\pgfmathparse{\mytemparray[1]} \edef\myb{\pgfmathresult}
\pgfmathparse{\mytemparray[2]} \edef\myc{\pgfmathresult} %also possible \pgfmathsetmacro
\begin{tikzpicture}[remember picture, overlay]
\node[anchor=#1, opacity=\myc] at ( \mya pt, \myb pt) {#3};
\end{tikzpicture}%
%
%
\egroup
%
}
\begin{document}
\frame{
\doatpos{west}{0, 3 cm, .5}{This is text \rule{1cm}{1cm}}
\doatpos{west}{1cm, -3 cm,.7}{This is text a\rule{1cm}{1cm}}
}
\end{document}
This alternative uses pgfmath
for list-processing. Since pgfmathparse
is very expansion unfriendly i decided to hard-code and put those three elements in distinguished macros. This solution is again very specific and inflexible, but i think it's more obvious how to process more items in the list.
z
? Are you able to project 3-dimensional PDFs?z
coordinate should define order of visibility of objects. That's why there's no dimension in my MWE (minimal wanted example)pst-abspos
ortextpos
and/or TikZ using the specialcurrent page
node.\tikz[remember picture,overlay]\node[anchor=center] at (current page.center) {\includegraphics{foo}};
would do. Thez
part can be achieved by defining appropiate layers.