YABQ (yet another bib question...)

YABQ (yet another bib question...)

[Thomas A. Schmitz]

I'm pleased as punch with the new bib module and have been writing my  
new lists with it all day. There is one thing that I could not make  
work, and I wonder if it is possible at all. I have 2 entries:

[1] Some collected volume

[2] Some article, in: a journal [reprinted in [1]].

Is it possible to refer back to item [1] automatically, by sticking  
the "reprinted in" in a "note" field in the database? I tried a  
number of combinations of \crossref or so, but couldn't get this to  
work.

Suggestions welcome!

Thomas

Re: YABQ (yet another bib question...)

[Sanjoy Mahajan]

> Is it possible to refer back to item [1] automatically, by sticking
> the "reprinted in" in a "note" field in the database? I tried a
> number of combinations of \crossref or so, but couldn't get this to
> work.

I need to repeat my experiments, but I think it worked to use simply
\cite in the bibtex entry.  Ah, here is one example.  It uses
xampl.bib, which comes with old distributions of ConTeXt (e.g. teTeX
3.0's distribution).

[Taco: The reference to Knuth below comes out as (Knuth, 1981c) even
though there are no other 1981 entries for Knuth cited in the text.
Is that hard to fix given ConTeXt's way of dealing with bibtex, in
that it just incorporates the entire database, in which there are
several 1981 entries from Knuth?]

===================== cut here ========
\usemodule[bib]
\usemodule[bibltx]
\setupbibtex[database=xampl]
\starttext

Here is a citation to a book that contains a cross reference:
\cite[book-crossref].  The bibliography entry (next page) for it
should contain a bib reference to another item in the bibliography.

\completepublications
\stoptext
===================== cut here ========

Here are the relevant entries from xampl.bib, which indicates that
it's the crossref= line that you need (not \cite as I said above)v:

@BOOK{book-crossref,
   crossref = "whole-set",
   title = "Seminumerical Algorithms",
   volume = 2,
   series = "The Art of Computer Programming",
   edition = "Second",
   year = "{\noopsort{1973c}}1981",
   note = "This is a cross-referencing BOOK entry",
}

@BOOK{whole-set,
   author = "Donald E. Knuth",
   publisher = "Addison-Wesley",
   title = "The Art of Computer Programming",
   series = "Four volumes",
   year = "{\noopsort{1973a}}{\switchargs{--90}{1968}}",
   note = "Seven volumes planned (this is a cross-referenced set of
   BOOKs)",
}

-Sanjoy

`Never underestimate the evil of which men of power are capable.'
         --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.

Re: YABQ (yet another bib question...)

[Taco Hoekwater]

Sanjoy Mahajan wrote:
> 
> I need to repeat my experiments, but I think it worked to use simply
> \cite in the bibtex entry.  Ah, here is one example.  It uses
> xampl.bib, which comes with old distributions of ConTeXt (e.g. teTeX
> 3.0's distribution).

Yeah, that's how I expected it to be done.

> [Taco: The reference to Knuth below comes out as (Knuth, 1981c) even
> though there are no other 1981 entries for Knuth cited in the text.
> Is that hard to fix given ConTeXt's way of dealing with bibtex, in
> that it just incorporates the entire database, in which there are
> several 1981 entries from Knuth?]

Problematic. There is a small bit of influence possible: the `c' is
inside macro (\maybeyear) that could be redefined (nullified), but
that is a document-global solution, and it may not even work too well
at that. It's been a long time since I looked at this problem. It is
not totally unsolvable, but definately not simple either.

Best,
Taco

Re: YABQ (yet another bib question...)

[Thomas A. Schmitz]

On Sep 10, 2006, at 3:30 PM, Taco Hoekwater wrote:

>> [Taco: The reference to Knuth below comes out as (Knuth, 1981c) even
>> though there are no other 1981 entries for Knuth cited in the text.
>> Is that hard to fix given ConTeXt's way of dealing with bibtex, in
>> that it just incorporates the entire database, in which there are
>> several 1981 entries from Knuth?]
>
> Problematic. There is a small bit of influence possible: the `c' is
> inside macro (\maybeyear) that could be redefined (nullified), but
> that is a document-global solution, and it may not even work too well
> at that. It's been a long time since I looked at this problem. It is
> not totally unsolvable, but definately not simple either.
>
Taco,

I just ran into the same question. When using a citation style that  
doesn't quote the year (like refcommand=num), it seems more logical  
to drop the maybeyear letter. It seems to work setting

\def\maybeyear{\gobbleoneargument}

I haven't seen any side effects yet, but maybe I'm missing something.  
Would it be possible to make this behavior a global option/value?

Best

Thomas

Re: YABQ (yet another bib question...)

[Taco Hoekwater]

Thomas A. Schmitz wrote:
> 
> I just ran into the same question. When using a citation style that  
> doesn't quote the year (like refcommand=num), it seems more logical  
> to drop the maybeyear letter. It seems to work setting
> 
> \def\maybeyear{\gobbleoneargument}
> 
> I haven't seen any side effects yet, but maybe I'm missing something.  
> Would it be possible to make this behavior a global option/value?

The problem with redefining \maybeyear is that it affects all
citations equally: when you have three Knuth records with the
same year and two Tuftes with the same year in the database, then
if you use only one of Knuths but both Tuftes, you can no longer
see which one of the Tuftes you were citing.

I could add an interface setting, so that you would do not have to
resort to using \def, but it would be much better if I could fix
the problem internally. Needs some thinking, though

Taco

Re: YABQ (yet another bib question...)

[Thomas A. Schmitz]

On Sep 15, 2006, at 9:57 AM, Taco Hoekwater wrote:

> The problem with redefining \maybeyear is that it affects all
> citations equally: when you have three Knuth records with the
> same year and two Tuftes with the same year in the database, then
> if you use only one of Knuths but both Tuftes, you can no longer
> see which one of the Tuftes you were citing.
>
> I could add an interface setting, so that you would do not have to
> resort to using \def, but it would be much better if I could fix
> the problem internally. Needs some thinking, though
>
> Taco

Yes, I see the difficulty. I don't know enough about bibtex styles  
and habits in general, but I'm just wondering if \maybeyear should be  
used in a different way - in my discipline, I think it would.  
Consider this example:

Hoekwater, First article, journalA, vol. 5 (2006)

Hoekwater, Second article, journalB, vol. 10 (2006)

IMHO, bibtex should now generate keys such as Hoek2006a and Hoek2006b  
or authoryear references like Hoekwater (2006a) and Hoekwater  
(2006b), but it should not (!) append the "a" and "b" to the years in  
the bibliographic list itself. But I may be wrong here - what do you  
and other users say? Should the (2006) in the example above come out  
as "(2006a)" and "(2006b)"? And as I was suggesting: if you use  
numered references, I think it would be best to just switch maybeyear  
off completely, so my suggestion would be to have a switch "maybeyear  
= on/off" for  the \setuppublicationlist.

Best

Thomas

Re: YABQ (yet another bib question...)

[Taco Hoekwater]

Thomas A. Schmitz wrote:
> IMHO, bibtex should now generate keys such as Hoek2006a and Hoek2006b  
> or authoryear references like Hoekwater (2006a) and Hoekwater  
> (2006b), but it should not (!) append the "a" and "b" to the years in  
> the bibliographic list itself. 

Actually, sometimes it should, namely in authoryear styles with
the list in database ordering as opposed to cite ordering.

> But I may be wrong here - what do you  
> and other users say? Should the (2006) in the example above come out  
> as "(2006a)" and "(2006b)"? And as I was suggesting: if you use  
> numered references, I think it would be best to just switch maybeyear  
> off completely, so my suggestion would be to have a switch "maybeyear  
> = on/off" for  the \setuppublicationlist.

Sure, that is simple. Will upload in a few minutes

Taco

Re: YABQ (yet another bib question...)

[Thomas A. Schmitz]

On Sep 15, 2006, at 11:47 AM, Taco Hoekwater wrote:

>
>
> Thomas A. Schmitz wrote:
>> IMHO, bibtex should now generate keys such as Hoek2006a and Hoek2006b
>> or authoryear references like Hoekwater (2006a) and Hoekwater
>> (2006b), but it should not (!) append the "a" and "b" to the years in
>> the bibliographic list itself.
>
> Actually, sometimes it should, namely in authoryear styles with
> the list in database ordering as opposed to cite ordering.

Oh dear, once you go into the gory details of bibliographies, there's  
no end to it...

>
>> But I may be wrong here - what do you
>> and other users say? Should the (2006) in the example above come out
>> as "(2006a)" and "(2006b)"? And as I was suggesting: if you use
>> numered references, I think it would be best to just switch maybeyear
>> off completely, so my suggestion would be to have a switch "maybeyear
>> = on/off" for  the \setuppublicationlist.
>
> Sure, that is simple. Will upload in a few minutes
>
> Taco

Thanks Taco, that's sweet. Maybe other users can chime in. I'm still  
thinking whether it would be useful to add an option that would  
append maybeyyear only to the reference key, not to the list... Such  
as a bibliographic style that would produce this sort of list:

Hoekwater 2006a = Hoekwater, "Article 1,"
                   JournalA 2 (2006) 20--30
Hoekwater 2006b = Hoekwater, "Article 2,"
                   JournalB 2 (2006) 30--40

But disregard this if you think it's silly or if it's too difficult  
to implement.

Re: YABQ (yet another bib question...)

[Sanjoy Mahajan]

> IMHO, bibtex should now generate keys such as Hoek2006a and
> Hoek2006b or authoryear references like Hoekwater (2006a) and
> Hoekwater (2006b), but it should not (!) append the "a" and "b" to
> the years in the bibliographic list itself. But I may be wrong here
> - what do you and other users say? Should the (2006) in the example
> above come out as "(2006a)" and "(2006b)"? 

At first I was sure that (2006a) and (2006b) are the right answer for
the list.  Othewise how else could the user know which entry to look
up when they see, say, Hoekwater (2006b) in the text?  But I just
figured out the answer to that question: Users count 'a', 'b',
... starting with the first 2006 entry.  However, I still don't think
it's a good idea to make them do that.  Let's not ask users to do what
computers do very well (counting)!

> And as I was suggesting: if you use numered references, I think it
> would be best to just switch maybeyear off completely, so my
> suggestion would be to have a switch "maybeyear = on/off" for the
> \setuppublicationlist.

Let me know whether I'm understanding it correctly.  If you have a
numbered reference list, then the year can still end up with a letter
tag, e.g.

1. Taco Hoekwater.  JournalA.  2006a
2. Taco Hoekwater.  JournalB.  2006b

Ah, I hadn't thought of that problem.  You're right, there shouldn't
be a maybeyear in this case since the list number disambiguates the
reference completely.

-Sanjoy

`Never underestimate the evil of which men of power are capable.'
         --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.

Re: YABQ (yet another bib question...)

[Thomas A. Schmitz]

On Sep 15, 2006, at 2:34 PM, Sanjoy Mahajan wrote:

> At first I was sure that (2006a) and (2006b) are the right answer for
> the list.  Othewise how else could the user know which entry to look
> up when they see, say, Hoekwater (2006b) in the text?  But I just
> figured out the answer to that question: Users count 'a', 'b',
> ... starting with the first 2006 entry.  However, I still don't think
> it's a good idea to make them do that.  Let's not ask users to do what
> computers do very well (counting)!
>
> Let me know whether I'm understanding it correctly.  If you have a
> numbered reference list, then the year can still end up with a letter
> tag, e.g.
>
> 1. Taco Hoekwater.  JournalA.  2006a
> 2. Taco Hoekwater.  JournalB.  2006b
>
> Ah, I hadn't thought of that problem.  You're right, there shouldn't
> be a maybeyear in this case since the list number disambiguates the
> reference completely.
>
Sanjoy,

yes, I agree completely: let the computer do the counting and  
bookkeeping! And you've hit the nail on the head: what I meant was,  
in cases where the form of  the list makes the reference completely  
unambiguous (because it is numbered or because keys/authoryear tags  
are prefixed), adding another number in the bibliographic entry is  
superfluous and somewhat ugly.

So question to Taco: maybe we need three options for maybeyear?

1. off [always]

2. on [always]

3. on for tags and authoryear etc., off for the date entry in the  
list itself.

Am I making sense? Are we working you to the bones? ;-)

Best

Thomas

Re: YABQ (yet another bib question...)

[Taco Hoekwater]

Thomas A. Schmitz wrote:
> So question to Taco: maybe we need three options for maybeyear?
> 
> 1. off [always]

This is the one that is missing, yes? Because the

   \setuppublicationlist[maybeyear=off]

is actually option 3.

  > 2. on [always]
> 
> 3. on for tags and authoryear etc., off for the date entry in the  
> list itself.

Re: YABQ (yet another bib question...)

[Thomas A. Schmitz]

Sanjoy,

thanks for your rely. Of course you're quite right, this is the wau  
to go, and it works. But my question was imprecise because I did not  
mention what was causing the problem: I'm working with split  
bibliographies and want to refer in chapter 2 to an item already  
included in the biblio of chapter 2. If I just put a \cite command  
somewhere, this item will be included a second time in the list of  
chapter 2. So my question is: this item does already have a counter  
attached to it. Is it possible to refer to this raw counter across  
chapters? Like so:

Chapter 1

[1] main reference

Chapter 2

[2] another thing

[3] and yet another [reprinted in [1]]

without having [1] main reference repeated. This my very well be  
impossible, I'm just wondering.

Thanks, and all best

Thomas

On Sep 9, 2006, at 12:40 AM, Sanjoy Mahajan wrote:

>> Is it possible to refer back to item [1] automatically, by sticking
>> the "reprinted in" in a "note" field in the database? I tried a
>> number of combinations of \crossref or so, but couldn't get this to
>> work.
>
> I need to repeat my experiments, but I think it worked to use simply
> \cite in the bibtex entry.  Ah, here is one example.  It uses
> xampl.bib, which comes with old distributions of ConTeXt (e.g. teTeX
> 3.0's distribution).
>
> [Taco: The reference to Knuth below comes out as (Knuth, 1981c) even
> though there are no other 1981 entries for Knuth cited in the text.
> Is that hard to fix given ConTeXt's way of dealing with bibtex, in
> that it just incorporates the entire database, in which there are
> several 1981 entries from Knuth?]
>
> ===================== cut here ========
> \usemodule[bib]
> \usemodule[bibltx]
> \setupbibtex[database=xampl]
> \starttext
>
> Here is a citation to a book that contains a cross reference:
> \cite[book-crossref].  The bibliography entry (next page) for it
> should contain a bib reference to another item in the bibliography.
>
> \completepublications
> \stoptext
> ===================== cut here ========
>
> Here are the relevant entries from xampl.bib, which indicates that
> it's the crossref= line that you need (not \cite as I said above)v:
>
> @BOOK{book-crossref,
>    crossref = "whole-set",
>    title = "Seminumerical Algorithms",
>    volume = 2,
>    series = "The Art of Computer Programming",
>    edition = "Second",
>    year = "{\noopsort{1973c}}1981",
>    note = "This is a cross-referencing BOOK entry",
> }
>
> @BOOK{whole-set,
>    author = "Donald E. Knuth",
>    publisher = "Addison-Wesley",
>    title = "The Art of Computer Programming",
>    series = "Four volumes",
>    year = "{\noopsort{1973a}}{\switchargs{--90}{1968}}",
>    note = "Seven volumes planned (this is a cross-referenced set of
>    BOOKs)",
> }
>
> -Sanjoy
>
> `Never underestimate the evil of which men of power are capable.'
>          --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.
> _______________________________________________
> ntg-context mailing list
> ntg-context@ntg.nl
> http://www.ntg.nl/mailman/listinfo/ntg-context

Re: YABQ (yet another bib question...)

[Taco Hoekwater]

Thomas A. Schmitz wrote:
> Sanjoy,
> 
> thanks for your rely. Of course you're quite right, this is the wau  
> to go, and it works. But my question was imprecise because I did not  
> mention what was causing the problem: I'm working with split  
> bibliographies and want to refer in chapter 2 to an item already  
> included in the biblio of chapter 2. If I just put a \cite command  
> somewhere, this item will be included a second time in the list of  
> chapter 2. So my question is: this item does already have a counter  
> attached to it. Is it possible to refer to this raw counter across  
> chapters? Like so:
> 
> Chapter 1
> 
> [1] main reference
> 
> Chapter 2
> 
> [2] another thing
> 
> [3] and yet another [reprinted in [1]]
> 
> without having [1] main reference repeated. This my very well be  
> impossible, I'm just wondering.

You can try this:

   \def\silentcite[#1]%
     {\bgroup
      \let\addthisref\gobbleoneargument
      \cite[#1]%
      \egroup}

(untested). If that does not work, it is not possible.

Taco

Re: YABQ (yet another bib question...)

[Thomas A. Schmitz]

On Sep 11, 2006, at 7:11 PM, Taco Hoekwater wrote:

> You can try this:
>
>    \def\silentcite[#1]%
>      {\bgroup
>       \let\addthisref\gobbleoneargument
>       \cite[#1]%
>       \egroup}
>
> (untested). If that does not work, it is not possible.
>
> Taco

Taco,

thanks for the suggestion. It doesn't work (still  produces an entry  
in the bibliography), but it's not a big deal, these are very special  
cases, and I can edit the bbl manually in order to get them right.

On a related note: all lovers of the bib module (and I know there are  
many of them out there) may be interested to know of two nice things.  
I found them out while I was switching my bibliographies to the new  
system. So if you use the one and only editor, emacs, you may want to  
try this:

1. reftex works with context! It cannot automatically determine the  
database it is supposed to use (I have to look into the customization  
again to see whether it can be taught to do so), but you can tell it  
explicitly about your .bib file, and then, all the usual keybindings  
work, you get a nice buffer where you can choose your references etc.  
There was a discussion a while ago on the mailing list, and my  
impression was it didn't work, so I was thrilled to see it did!

2. I discovered another nice little app that works within emacs. It's  
called ebib. It is not very good as an editor for bibtex databases,  
but it has a great tab-completion feature for inserting references  
into your files. I tweaked it a bit so it would insert the proper  
\cite[] and \nocite[] commands for ConTeXt, and it is the fastest and  
most convenient thing I've ever seen. Maybe I should say something  
about on the wiki?

Anyway, Taco, thanks for your help, and all best

Thomas

Re: YABQ (yet another bib question...)

[Sanjoy Mahajan]

Thomas,

> Is it possible to refer to this raw counter across chapters?

Not sure -- it is beyond my ConTeXt bib module competence, but I hope
one of the resident experts has an idea...

-Sanjoy

`Never underestimate the evil of which men of power are capable.'
         --Bertrand Russell, _War Crimes in Vietnam_, chapter 1.