tweaking the placement of sidenotes
Lynn Newton
We have a couple of cases where the text of \footnote drops below
the footer level. I thought I was being clever when I used negative
vskip, until I realized that this does not move the footnote number
in the margin as well, e.g., if I write:
\footnote{\vskip-10pt Footnotetext}
the text "Footnotetext" gets raised but the footnote number stays
put. Grrr.
How can I accomplish what I need to do here?
--
Lynn David Newton
Interhack Corporation
web.interhack.com
Kevin Godby
On Thu, Jan 15, 2009 at 7:57 AM, Lynn Newton <lynn....@gmail.com> wrote:
>
> I know I can't bethe first person to ask about this ...
>
> We have a couple of cases where the text of \footnote drops below
> the footer level. I thought I was being clever when I used negative
> vskip, until I realized that this does not move the footnote number
> in the margin as well, e.g., if I write:
>
> \footnote{\vskip-10pt Footnotetext}
>
> the text "Footnotetext" gets raised but the footnote number stays
> put. Grrr.
>
> How can I accomplish what I need to do here?
I've run into that problem as well. Let me work on it a bit today and
I'll see what I can do for you. The solution I attempt will probably
be something like:
\sidenote[-3\baselineskip][]{My sidenote text.}
where the -3\baselineskip is the amount of vertical space to move
(negative is up, positive is down). The empty [] brackets would be
for the sidenote number (since the actual definition of \footnote is
\footnote[footnotenum]{footnote text} where footnotenum is optional).
I've been slowly tweaking the stuff from footmisc in our own style
anyway, so this may be the tipping point to just rip out the footmisc
stuff what we use and modify it for our own purposes. This would
allow us to shed the footmisc dependency (but would put the onus on me
for bug fixes).
I'll work on this today and let you know what I come up with.
--Kevin
Kevin Godby
I just dropped some code into the repository that allows you to modify
the vertical position of the sidenotes. The syntax is:
\sidenote[num][dimen]{Sidenote text.}
where ``num'' is the sidenote number and ``dimen'' is the amount of
space you wish to move the sidenote down the page.
You can leave ``num'' blank (empty brackets []) and it will use the
normal sidenote number. This parameter exists for
backward-compatibility with LaTeX's \footnote command.
The ``dimen'' parameter should be a TeX dimension such as 0pt or 1in
or 5pc or 3\baselineskip. If left empty (or completely left out), it
will default to 0pt. If you use a negative value (e.g.,
-3\baselineskip), it will move the sidenote up the page.
If you want to set the ``dimen'' parameter, you *must* also set the
``num'' parameter (or leave empty brackets as a placeholder):
\sidenote[][-1.0cm]{This sidenote is 1 centimeter higher than it
normally would be and uses its original sidenote number.}
If you only want to change the sidenote number (it's hard for me to
imagine why one would wish to do this, but I'm trying to keep the
commands compatible with LaTeX's \footnote command), you can leave the
``dimen'' parameter entirely:
\sidenote[12345]{This sidenote is numbered ``12345.''}
Note that setting the sidenote number manually will not adjust the
sidenote (footnote) counter, so the number will be used for the
current sidenote and the following sidenote will be numbered
sequentially using the original numbering. (So the sidenote following
our last example above would likely not be numbered 12346.)
In the writing this code, I've dropped our dependency on the
``footmisc'' package, but added a dependency on the ``optparams''
package. (The ``optparams'' package allows us to use multiple
optional arguments with a command.)
Feel free to give the new code a try and let me know if you find any
bugs with it.
--Kevin Godby
Lynn Newton
I'm a bit behind keeping up on changes myself. Real production
work and all, you know. (And we are fully using our new
classes that inherit from the tufte-latex package for
everything now, so this stuff is important to us.)
> I just dropped some code into the repository that allows you to
> modify the vertical position of the sidenotes. The syntax is:
> \sidenote[num][dimen]{Sidenote text.} ...
Etc. Cool, thank you. I'll have to wait until I'm at home to get
an update because we are not able to run svn to the source base
even with proxy service enabled in my servers file. But I have a
couple of preliminary questions.
Are the changes you describe carried over to bibliographic
citations? I'm working on a piece in handout class that our
Practice Leader presented at a local college early this week,
where the material in the main body column is largely short
paragraphs, but in the course of three pages, there are 34
bibliographic citations, most of them within a span of two pages.
Naturally, these run right off the page, and present me with the
onerous "Float(s) lost" error.
--
Lynn Newton
Interhack Corporation
web.interhack.com
Kevin Godby
On Thu, Jan 29, 2009 at 12:36 PM, Lynn Newton <lynn....@gmail.com> wrote:
^ snip *
> Etc. Cool, thank you. I'll have to wait until I'm at home to get
> an update because we are not able to run svn to the source base
> even with proxy service enabled in my servers file. But I have a
> couple of preliminary questions.
>
> Are the changes you describe carried over to bibliographic
> citations? I'm working on a piece in handout class that our
> Practice Leader presented at a local college early this week,
> where the material in the main body column is largely short
> paragraphs, but in the course of three pages, there are 34
> bibliographic citations, most of them within a span of two pages.
> Naturally, these run right off the page, and present me with the
> onerous "Float(s) lost" error.
The current \cite command doesn't accept a vertical offset, but I'll
work on adding that this weekend. Should be easy enough, since it
effectively calls \sidenote.
Wow.. that's a lot of citations in a short space. Is the sidenotes
margin completely filled, or do you have some whitespace that would
allow you to slide the sidenotes up the page a bit?
You can also force a page break to help split up the paragraphs (and
their citations) across multiple pages.
Another option may be to find a more concise bibliography style. I
haven't explored too many of them, so I don't know if there's one
that's useful or not. You could always create your own, too, I
suppose.
Without seeing the pages, it's hard to make sensible recommendations.
Let me know if there's anything the document classes could do to help.
--Kevin Godby
Lynn Newton
>> home to get an update because we are not able to run
>> svn to the source base
>> Are the changes you describe carried over to bibliographic
> The current \cite command doesn't accept a vertical
> offset, but I'll work on adding that this weekend.
> Should be easy enough, since it effectively calls
> \sidenote.
use of that.
> Wow.. that's a lot of citations in a short space. Is
> the sidenotes margin completely filled, or do you
> have some whitespace that would allow you to slide
> the sidenotes up the page a bit?
little bit with trying to add page breaks.
The document is a summary and background style of
article designed to accompany a presentation our
practice leader gave at Franklin College on Monday
about how to remain a viable worker in technology over
the next twenty-five years. One of his main points was
to learn the foundations -- the science and the
history, rather than just trying to keep up with the
latest new thing and never being able to do so.
Therefore, this summary included a lot of references to
seminal books and articles from a broad cross-section
of technologies, from operating systems, to networking
to GUIs and other things. For such a production, having
the full bibliographical quote available in line with
the terse text was a real advantage.
> Another option may be to find a more concise
We use the plainnat bibliography style pretty
consistency.
I'll take another pass at trying to repaginate this
beast today, but because \cite is just a variation of
\sidenote, I know that ability to move them up and down
is going to be an ongoing problem for us, so we will
look anxiously to you for a project sanctified
solution.
Thanks much.
Kevin Godby
>> The current \cite command doesn't accept a vertical
>> offset, but I'll work on adding that this weekend.
>> Should be easy enough, since it effectively calls
>> \sidenote.
>
> That would be very nice, and we would definitely make
> use of that.
Okay, I just committed code to svn for \cite[offset]{bibkey}. I also
modified \cite{bibkey1,bibkey2,...} so that all the bibliography
entries are printed in a single sidenote paragraph (instead of
separate sidenotes as before). This is more in line with how Tufte
prints his citations and should save some vertical space in the
sidenote column.
I've only given it some cursory tests, so please let me know if you
run into any problems with it.
--Kevin
Bil Kleb
> On Thu, Jan 29, 2009 at 12:36 PM, Lynn Newton <lynn....@gmail.com> wrote:
>
>> Are the changes you describe carried over to bibliographic
>> citations? I'm working on a piece in handout class that our
>> Practice Leader presented at a local college early this week,
>> where the material in the main body column is largely short
>> paragraphs, but in the course of three pages, there are 34
>> bibliographic citations, most of them within a span of two pages.
>> Naturally, these run right off the page, and present me with the
>> onerous "Float(s) lost" error.
>
> The current \cite command doesn't accept a vertical offset, but I'll
> work on adding that this weekend. Should be easy enough, since it
> effectively calls \sidenote.
I feel offering this would stray from the Tuftian spirit of sidenotes.
> Wow.. that's a lot of citations in a short space. Is the sidenotes
> margin completely filled, or do you have some whitespace that would
> allow you to slide the sidenotes up the page a bit?
This gets to the philosophy of sidenotes vis-a-vis footnotes.
I found that sidenote citations forced me to focus my thoughts
w.r.t. to citations precisely because it created a constraint on
how many you could use. IIRC, I've also seen Tufte make similar
remarks.
BibTeX makes it so easy to use citations that I found I had to
think carefully about which ones *really* were essential. In the end,
I feel my papers are more readable, which, after all, is the point?
FWIW, I often heard Clay Jenkinson of the http://jeffersonhour.org
radio show (and now also podcast) remark that Jefferson disliked the
elaborate Ciceronian rhetoric style and especially that which John Adams
produced because it was heavily laden with footnotes. I look to
Jefferson as a model for extremely clear, succinct writing.
Regards,
--
Bil Kleb
http://fun3d.larc.nasa.gov
Bil Kleb
>
> I feel offering this would stray from the Tuftian spirit of sidenotes.
Nevermind; since \cite is now like \sidenote it should
have similar capabilities.
Regards,
--
Bil Kleb
http://twitter.com/bil_kleb
Lynn Newton
this diff in tufte-book.tex:
6c6
< \author[The Tufte-LaTeX Developers]{The Tufte-\LaTeX\ Developers}
---
> \author[The Tufte-LaTeX Developers]{The Tufte-LaTeX\ Developers}
Was that intentional?
Regarding Bil's comment - point well taken --
unquestionably clarity should be the objective.
It turns out we do have the need for this in
one one particular anomalous situation, and in
some cases it's better to have an option
available than not to have it. Just as it's
sometimes desirable to be able to change
page dimensions temporarily, it's sometimes
desirable to be able to make minor adjustments
of other typesetting matters, and with the
power of TeX driving it, there's no reason
why it shouldn't be possible.
Also, I would suggest that while Dr. Tufte
is the master and his books the inspiration for
this package, few authors or typesetters will
want to duplicate another's exactly. That's
why, for instance, at the company where I work
we developed our own set of classes, inheriting
first from the tufte-latex classes (which
themselves inherit from standard LaTeX), adding
our own things such as letter, proposal, outline,
presentation, and table of contents. In the case
of the latter, our documents are the type that
tend to be referred to, particularly those that
deal with legal and forensic matters, and therefore
need to retain section numbers. So we designed
our own that keep those, and that also list the
sections and subsections in the TOC, while
retaining other features of the Tufte look.
Thanks Kevin for the quick turnaround on this!
--
Lynn Newton
Interhack Corp.
web.interhack.com
Kevin Godby
On Sat, Jan 31, 2009 at 5:08 AM, Bil Kleb <bil....@gmail.com> wrote:
>
> Kevin Godby wrote:
>> On Thu, Jan 29, 2009 at 12:36 PM, Lynn Newton <lynn....@gmail.com> wrote:
>> The current \cite command doesn't accept a vertical offset, but I'll
>> work on adding that this weekend. Should be easy enough, since it
>> effectively calls \sidenote.
>
> I feel offering this would stray from the Tuftian spirit of sidenotes.
In /Beautiful Evidence/, Tufte often moves the sidenotes around to
accommodate graphics, page boundaries, and other elements. In fact,
I've been doing some more measurements on this book lately so I can
establish some grid spacing in our document classes, and I've
discovered that he breaks the grid more often than not! Sometimes his
sidenotes are 12 picas wide, sometimes 14 picas wide. Graphics and
sidenotes along will break out of their normal bounding boxes and hang
into margins.
A few notes from Tufte himself on his style can be found here:
http://www.edwardtufte.com/bboard/q-and-a-fetch-msg?msg_id=0000ld&topic_id=1
``I also seek to optimize each double-page layout, sacrificing global
consistency (or a global theory about sidenote use) across all the
layouts; this is sometimes called `breaking the grid.' '' ---Edward
R. Tufte, 4 September 2006
>> Wow.. that's a lot of citations in a short space. Is the sidenotes
>> margin completely filled, or do you have some whitespace that would
>> allow you to slide the sidenotes up the page a bit?
>
> This gets to the philosophy of sidenotes vis-a-vis footnotes.
>
> I found that sidenote citations forced me to focus my thoughts
> w.r.t. to citations precisely because it created a constraint on
> how many you could use. IIRC, I've also seen Tufte make similar
> remarks.
In support of your claim (from the same link above):
``But footnotes should not be overdone or pretentious, as is often the
case in PhD dissertations. Also don't build in little all-too-knowing
commentaries into the references, a graduate student disease, as in
`the seminal article by G. P. K. Sitzplotz, [reference].' '' ---Edward
R. Tufte, 4 September 2006
``Of course, the big-column/little-column format should not be done
mechanically; there needs to be worthwhile material that naturally
belongs in the margin. If the marginal materials are simply
references, then the standard footnotes (at the bottom of each page,
not ganged together at the end of the document) are fine. Maybe the
empty space in the margin can be left for marginal notes written by
the readers, if any.'' ---Edward R. Tufte, 13 September 2003
> BibTeX makes it so easy to use citations that I found I had to
> think carefully about which ones *really* were essential. In the end,
> I feel my papers are more readable, which, after all, is the point?
I agree completely. When I first started using LaTeX, I went a little
overboard with footnotes and citations. Why? Because LaTeX made is
so *easy* to use them that I almost felt obligated to take full
advantage of the capability.
After trying to read a few of my documents and seeing how slipshod
they appeared, I started to moderate my use of footnotes and
integrated the citations into the text to make it more readable.
(More on the citations below.)
Just because one *can* doesn't mean one *should*.
> FWIW, I often heard Clay Jenkinson of the http://jeffersonhour.org
> radio show (and now also podcast) remark that Jefferson disliked the
> elaborate Ciceronian rhetoric style and especially that which John Adams
> produced because it was heavily laden with footnotes. I look to
> Jefferson as a model for extremely clear, succinct writing.
My brother has a collection of Jefferson's (and Madison's writings) on
his bookshelf. I'll have to take a look at them sometime. Thanks for
the intriguing reference!
On the subject of citations, I prefer to integrate them into the text
more. (This can be difficult if your field's citation style demands
otherwise.) Compare:
``In /Beautiful Evidence/, Tufte suggests...'' (and then providing the
full citation in the margin if there is room, or letting the reader
find the appropriate entry in the bibliography at the end of your
document)
with:
``Tufte suggests[1] ...'' (with a full citation in the sidenote,
footnote, or bibliography).
I find the first much easier to read. It doesn't demand that I
immediately stop and find the reference, and if I do with to find the
full citation, it's easier to remember the author's name and the book
title than a citation number (or something more hideous like [Tuf06]).
If you'd like to include an entry in the bibliography but not print it
in a sidenote, use the \nocite{bibkey} command.
--Kevin Godby
Kevin Godby
>
> Bil Kleb wrote:
>>
>> I feel offering this would stray from the Tuftian spirit of sidenotes.
>
> Nevermind; since \cite is now like \sidenote it should
> have similar capabilities.
Since LaTeX has a number of problems with layout out sidenotes (e.g.,
it can't detect when they're being set atop some other element such as
a full-width graphic that runs into the margin), I figured I'd opt for
consistency and provide a vertical offset for any of the elements we
place in the sidenotes area (\sidenote, \cite, \marginnote,
\marginfigure, \margintable).
Of course, providing the offsets should be done only in the final
stage of document preparation as adding or removing material in the
main body text will cause the sidenotes-like material to move around
as well. So the vertical offsets should be used as a last resort.
--Kevin Godby
Kevin Godby
On Sat, Jan 31, 2009 at 7:01 AM, Lynn Newton <lynn....@gmail.com> wrote:
>
> Kevin, when I pulled the new source, I noted
> this diff in tufte-book.tex:
>
> 6c6
> < \author[The Tufte-LaTeX Developers]{The Tufte-\LaTeX\ Developers}
> ---
>> \author[The Tufte-LaTeX Developers]{The Tufte-LaTeX\ Developers}
>
> Was that intentional?
Yes. I got tired of the sample-book document complaining about
typesetting \LaTeX in the sans serif font on the title page. (It has
something to do with the sans serif font lacking some math symbols, I
think, but I haven't taken the time to dive into it too deeply yet.)
* snip *
> Also, I would suggest that while Dr. Tufte
> is the master and his books the inspiration for
> this package, few authors or typesetters will
> want to duplicate another's exactly. That's
> why, for instance, at the company where I work
> we developed our own set of classes, inheriting
> first from the tufte-latex classes (which
> themselves inherit from standard LaTeX), adding
> our own things such as letter, proposal, outline,
> presentation, and table of contents. In the case
> of the latter, our documents are the type that
> tend to be referred to, particularly those that
> deal with legal and forensic matters, and therefore
> need to retain section numbers. So we designed
> our own that keep those, and that also list the
> sections and subsections in the TOC, while
> retaining other features of the Tufte look.
My general philosophy on this topic is that the Tufte-LaTeX document
classes should be relatively opinionated, so to speak. if we didn't
take a stand on many of the design issues that crop up, then we'd be
writing a meta-class like memoir. When possible, however, I try not
to prevent someone from modifying or tweaking things to better fit
their situation (as you've done above). (However, I would turn down a
request to be able to disable the sidenotes, for instance, because at
that point you should probably find a more appropriate base class.)
--Kevin Godby
Lynn Newton
whether footnotes or citations, if I have two or more
of them right in a row. Say, for instance, I have this:
Blah blah blah blah blah blah blah blah
blah\cite{first}\cite{second}\cite{third} and still
more blah blah blah blah blah blah blah.
There seems to be quite a bit of white space
surrounding these items.
BTW, in most caces I would prefer to use the style
shown in the example rather
than\cite{first,second,third} In practice, there are
always at least a couple of words between citations,
and are completely separate items.
Thoughts?
I was able to use the new dimension option to \cite to
tune up the document that started this discussion.
Kevin Godby
On Mon, Feb 2, 2009 at 10:13 AM, Lynn Newton <lynn....@gmail.com> wrote:
>
> What is the default vertical spacing between items,
> whether footnotes or citations, if I have two or more
> of them right in a row. Say, for instance, I have this:
>
> Blah blah blah blah blah blah blah blah
> blah\cite{first}\cite{second}\cite{third} and still
> more blah blah blah blah blah blah blah.
>
> There seems to be quite a bit of white space
> surrounding these items.
The smallest amount of space between two
sidenotes/citations/marginnotes defaults to the value of
\marginparpush which, in our case, is set to \baselineskip.
> BTW, in most caces I would prefer to use the style
> shown in the example rather
> than\cite{first,second,third} In practice, there are
> always at least a couple of words between citations,
> and are completely separate items.
>
> Thoughts?
I've been working on an update that sets some of the line-spacing and
font sizes to be more in line with Tufte's books. I haven't compared
them directly to the values we're currently using (LaTeX's defaults),
so I don't know if they will help you or hurt you in this situation.
Interestingly, Tufte uses different line spacing for body text, block
quotations, and sidenotes. Just to make life interesting (and
non-gridlike), i suppose.
> I was able to use the new dimension option to \cite to
> tune up the document that started this discussion.
Great! I'm glad it worked out for you.
--Kevin Godby
Lynn Newton
> The smallest amount of space between two
> sidenotes/citations/marginnotes defaults to the value
> of \marginparpush which, in our case, is set to
> \baselineskip.
Ok. It looks like more. But maybe that's my eyeballs.
> I've been working on an update that sets some of the
> line-spacing and font sizes to be more in line with
> Tufte's books. I haven't compared them directly to
> the values we're currently using (LaTeX's defaults),
> so I don't know if they will help you or hurt you in
> this situation.
I noticed your comments about measuring things in the
Tufte books. This is a good idea.
In the document wherein I've been using this feature,
almost all cases of \cite and \footnote have been hard
coded to some value. This makes it pretty fragile in
the case of any future revisions, but it's nearing
completion.
I guess I'm wondering this: if I specify, for instance,
\cite[-12pt]{bibtag}, that will move it 12 point up
from where it would normally print. But I don't know if
changes to the amount of default spacing between margin
items will change where that would print, requiring me
to change hard coded dimensions. (I'm guessing it
will.)
Also, I don't know exactly what the technique is for
determining precisely where in the margin a given item
will land when there is nothing to interfere with it,
and what factors cause that and other items that get
added to change, which direction, and by how much.
You're dealing with some pretty hairy stuff!
> Interestingly, Tufte uses different line spacing for
> body text, block quotations, and sidenotes. Just to
> make life interesting (and non-gridlike), i suppose.
I guess. Who needs grids? :-) Authors just want to be
able to put specks of ink on the page exactly where
they want them to go.