langsci / 259

Müller, Stefan et al. (eds): Head-Driven Phrase Structure Grammar: The handbook
Creative Commons Attribution 4.0 International
22 stars 7 forks source link

AVMs in type hierarchies are not at the top formal background.tex #85

Open stefan11 opened 3 years ago

stefan11 commented 3 years ago

grafik

Using anchor=north does not help, since then individual type are put as much up as possible and case would be higher than other types: grafik

stefan11 commented 3 years ago

This is related to #93.

stefan11 commented 3 years ago

Hi @sasozivanovic and @kopeckyf, I added Saso to this project since I finally understood that he was right. We have to find a way to align avms at the top. I always thought that using align=north to get an alignment of the AVMs in trees since this worked for the cases I looked at. But now I have a case with mixes of simple words and AVMs and if some of the words have high letters like "l" in "boolean" and others do not ("case") then the alignment breaks down. So would there be a way to shift the AVMs to the top? Maybe some option to \avm? I think Saso hat some proposal.

Thanks!

sasozivanovic commented 3 years ago

Hi all!

Stefan Müller je 26. 02. 21 ob 10:39 napisal:

Hi @sasozivanovic https://github.com/sasozivanovic and @kopeckyf https://github.com/kopeckyf, I added Saso to this project since I finally understood that he was right. We have to find a way to align avms at the top. I always thought that using align=north to get an alignment of the AVMs in trees since this worked for the cases I looked at. But now I have a case with mixes of simple words and AVMs and if some of the words have high letters like "l" in "boolean" and others do not ("case") then the alignment breaks down. So would there be a way to shift the AVMs to the top? Maybe some option to |\avm|? I think Saso hat some proposal. Yes, it should be in some pull request. Best, Sašo

Thanks!

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/langsci/259/issues/85#issuecomment-786530690, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAZJ2JEGRBRTXJ6OBQREDRLTA5T57ANCNFSM4XNUHBCQ.

kopeckyf commented 3 years ago

Hi,

I can't replicate the behaviour. For me, everything is fine:

grafik

\documentclass[tikz]{standalone}
\usepackage[linguistics]{forest}
\usepackage{langsci-avm}
\begin{document}
    \begin{forest} for tree = {font=\itshape, anchor=north}
        [object
            [\avm{[\type*{substantive}
                   prd & boolean]}
               [\avm{[\type*{verb}
                      vform & vform\\
                      prd   & plus]}
               ]   
               [\avm{[\type*{noun}
                      case & case]}
               ]
            ]
            [case]
            [vform]
            [boolean [plus] [minus] ]
        ]
    \end{forest}
\end{document}
kopeckyf commented 3 years ago

@sasozivanovic 's proposal for top-alignment of AVMs had implications for runtime, which I think are too big a drawback.

sasozivanovic commented 3 years ago

Felix Kopecky je 26. 02. 21 ob 12:14 napisal:

@sasozivanovic https://github.com/sasozivanovic 's proposal for top-alignment of AVMs had implications for runtime, which I think are too big a drawback. I know what you mean: for top-alignment, the AVM is typeset twice. But I don't see how this is a drawback, as the double-typesetting does not apply to center alignment. In other words, it only affects the situation which is currently not addressed.

Of course, I would prefer a single-typesetting solution for top-alignment as well, but I couldn't think of a way to do it, back then. Maybe there's a way, I don't know.

Best, Sašo

— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/langsci/259/issues/85#issuecomment-786583215, or unsubscribe https://github.com/notifications/unsubscribe-auth/AAZJ2JF4P2XKAEETUV2QT7LTA57AHANCNFSM4XNUHBCQ.

kopeckyf commented 3 years ago

As I currently understand the situation, anchor=north solves the issue reported here (see my example above). Besides not needing to construct the box twice, it also seems easier to just write for tree = {anchor=north} than any user input I could imagine that uses \avmsetup{}.