terms for letters and interviews

Hi–

Trying again:

I need to add two terms,“letter” and “interview.” They are needed in
at least the following cases:

Subsequent citation for interviews in Chicago note style, such as:
Hedges, interview

Bibliographic reference for letters in Chicago note w/ Bibliography
style, for example:
Creel, George. Letter to Colonel House. 25 September 1918. Edward M.
House Papers, Yale University Library.

I have SVN access to Zotero so I added the terms there, but I don’t
have SVN access to xbiblio so I can’t add them to CSL schema.

I’m happy to discuss this but really there is no way to properly
format these item types without these terms–the alternatives are
text values which would not be localizable.

I would think citing these item types properly should be just as
important as citing editions of books properly.

Thanks,
Elena

Elena Razlogova wrote:

I need to add two terms,“letter” and “interview.” They are needed in
at least the following cases:

I’m happy to discuss this but really there is no way to properly
format these item types without these terms–the alternatives are
text values which would not be localizable.

I would think citing these item types properly should be just as
important as citing editions of books properly.

Sure, but I’d been under the vague impression that you were handling
this on the Zotero end? Isn’t the CSL variable you’re using to format
these “genre” (which I’m again wondering if we should change to
"type"?), and in that case isn’t Zotero providing the appropriate string?

Just asking. The thing is, if we add these two, I worry we then have to
add terms for any imaginable type, and formatting won’t work without
them. I can see arguments on either side really.

Simon, what do you think?

Bruce

PS - If you want SVN access, send me your SF user name off-list.

Sure, but I’d been under the vague impression that you were handling
this on the Zotero end? Isn’t the CSL variable you’re using to format
these “genre” (which I’m again wondering if we should change to
"type"?), and in that case isn’t Zotero providing the appropriate
string?

“Type” would certainly be more descriptive–I’m only worried that it
would be confusing to have both “type” for “item type” and “type” as
a variable.

My message was about citing “e-mail” and “instant message” item types
(and I still haven’t heard from Simon on whether what I was proposing
was even possible). Unfortunately, with letters we cannot populate
"genre" automatically because for letters the word “letter” is
omitted in a citation but included in a bibliographic reference. So
even if we could set “genre” as “letter” in Zotero, we would end up
with the following:

George Creel to Colonel House, letter, 25 September 1918, Edward M.
House Papers, Yale University Library.

when it should be:

George Creel to Colonel House, 25 September 1918, Edward M. House
Papers, Yale University Library.

In other words, in citations type of letter is only included if it’s
something other than a letter–a telegram, memo, etc.

The interview item type doesn’t have a “genre” field at all, it has
"medium." I suppose Simon could still program Zotero to set “genre"
as “interview” (“medium” would be taken up by strings like
"transcript” or “tape”), but then wouldn’t CSL formatting for
"interview" be useless in all cases except within Zotero? It seems
more like a hack to me than a rule.

Just asking. The thing is, if we add these two, I worry we then
have to
add terms for any imaginable type, and formatting won’t work without
them. I can see arguments on either side really.

In theory, would it be possible to use something like <text
variable=“type” to display the name of the item type (i.e. letter,
interview, patent, etc.? Then we wouldn’t need terms either for
interview or letter, or any other terms that are the same as item
type names.

Elena

I need to add two terms,“letter” and “interview.” They are needed in
at least the following cases:

I added these this morning- together with a few of my own.

Is there a way to validate an rnc file like there is for xml files?

Julian.

I need to add two terms,“letter” and “interview.” They are needed in
at least the following cases:

I added these this morning- together with a few of my own.

Hang on, though. We’ve not settled whether this is within scope of CSL
to handle.

Is there a way to validate an rnc file like there is for xml files?

In oXygen, yes; you validate the same way.

With nxml, you can just load the schema; nxml will complain if there’s an error.

Bruce

My message was about citing “e-mail” and “instant message” item types

Ah yes; I’m just saying this is the same problem.

(and I still haven’t heard from Simon on whether what I was proposing
was even possible). Unfortunately, with letters we cannot populate
"genre" automatically because for letters the word “letter” is
omitted in a citation but included in a bibliographic reference.

How is that (citation vs. bibliography) relevant?

So even if we could set “genre” as “letter” in Zotero, we would end up
with the following:

George Creel to Colonel House, letter, 25 September 1918, Edward M.
House Papers, Yale University Library.

when it should be:

George Creel to Colonel House, 25 September 1918, Edward M. House
Papers, Yale University Library.

In other words, in citations type of letter is only included if it’s
something other than a letter–a telegram, memo, etc.

OK, but I just want to emphasize this is one of those corner case
rules I’m really leary of bending over backwards to support. No matter
what CMS says, no sane copy editor is going to call your first example
"wrong" and force you to change it.

So my criteria here is not what CMS says, but whether we can support
it in a) a general way , and b) without too much pain.

The interview item type doesn’t have a “genre” field at all, it has
"medium."

Well, does one ever cite different kinds of interviews? My thought is
no, so the current support is fine (notwithstanding that it’s flat,
and so fails in a lot of cases altogether!).

Just asking. The thing is, if we add these two, I worry we then
have to
add terms for any imaginable type, and formatting won’t work without
them. I can see arguments on either side really.

In theory, would it be possible to use something like <text
variable=“type” to display the name of the item type (i.e. letter,
interview, patent, etc.? Then we wouldn’t need terms either for
interview or letter, or any other terms that are the same as item
type names.

That’s what I’m suggesting. It’s a simple and general solution that
will work in most cases.

But it doesn’t solve your example above. The other downside is we
can’t localize it (which we define as “not our problem”).

Bruce

So even if we could set “genre” as “letter” in Zotero, we would
end up
with the following:

George Creel to Colonel House, letter, 25 September 1918, Edward M.
House Papers, Yale University Library.

when it should be:

George Creel to Colonel House, 25 September 1918, Edward M. House
Papers, Yale University Library.

In other words, in citations type of letter is only included if it’s
something other than a letter–a telegram, memo, etc.

OK, but I just want to emphasize this is one of those corner case
rules I’m really leary of bending over backwards to support. No matter
what CMS says, no sane copy editor is going to call your first example
"wrong" and force you to change it.

It’s looks terribly wrong to me, actually. To include “letter” as a
"genre" adds nothing to the citation and goes against the main
principle of footnotes–to use as little space and ink as possible
while providing accurate citations. Think how much extra space would
this needlessly take in a book that cites hundreds of letters (such
as mine).

In theory, would it be possible to use something like <text
variable=“type” to display the name of the item type (i.e. letter,
interview, patent, etc.? Then we wouldn’t need terms either for
interview or letter, or any other terms that are the same as item
type names.

That’s what I’m suggesting. It’s a simple and general solution that
will work in most cases.

But it doesn’t solve your example above.

Well, there are two options here:

  1. Rename “personal_communication” into “letter” and add <text
    variable=“type” capability.

  2. Add term “letter” and add <text variable=“type” capability.

I would vote for Option 1, because:

a) True, “personal_communication” is a more inclusive label than
"letter", but “letter” I think is more precise for this item type. It
describes all kinds of missives likely to be cited under this
category (telegrams, email, im) and does not describe “verbal
communication”–but wouldn’t verbal communication be cited as an
"interview" anyway?

b) “Personal_communication” is an abstract label whereas “letter” is
concrete in the sense that it can be used in CSL formatting.

c) This is also the solution that requires no coding, just global
search/replace.

The other downside is we
can’t localize it (which we define as “not our problem”).

Why would item types not be localizable? (Zotero item types are
already localized, I believe).

Best,
Elena> Bruce

Elena Razlogova wrote:

OK, but I just want to emphasize this is one of those corner case
rules I’m really leary of bending over backwards to support. No matter
what CMS says, no sane copy editor is going to call your first example
"wrong" and force you to change it.

It’s looks terribly wrong to me, actually. To include “letter” as a
"genre" adds nothing to the citation and goes against the main
principle of footnotes–to use as little space and ink as possible
while providing accurate citations. Think how much extra space would
this needlessly take in a book that cites hundreds of letters (such
as mine).

So for the short-run (e.g. until we resolve this and it gets
incorporated into Zotero), I’d suggest you just not print the type in
your citations at all.

In theory, would it be possible to use something like <text
variable=“type” to display the name of the item type (i.e. letter,
interview, patent, etc.? Then we wouldn’t need terms either for
interview or letter, or any other terms that are the same as item
type names.
That’s what I’m suggesting. It’s a simple and general solution that
will work in most cases.

But it doesn’t solve your example above.

Well, there are two options here:

  1. Rename “personal_communication” into “letter” and add <text
    variable=“type” capability.

  2. Add term “letter” and add <text variable=“type” capability.

I would vote for Option 1, because:

a) True, “personal_communication” is a more inclusive label than
"letter", but “letter” I think is more precise for this item type. It
describes all kinds of missives likely to be cited under this
category (telegrams, email, im) and does not describe “verbal
communication”–but wouldn’t verbal communication be cited as an
"interview" anyway?

Not necessarily in my view. I’ve had informal telephone conversations
that I have (or could) cite that I wouldn’t necessarily call so formal
as an “interview.”

But there’s admittedly some ambiguity here.

b) “Personal_communication” is an abstract label whereas “letter” is
concrete in the sense that it can be used in CSL formatting.

c) This is also the solution that requires no coding, just global
search/replace.

I’d probably opt for your option 2, which would allow markup like:

That does what you’re looking for (though is obviously rather complex).

We’d have to introduce some fallback expectations for the generic type,
though. Otherwise, we’d probably have to do something like:

See what I mean about complex? I can’t think of a totally easy and
generic solution to this problem.

Any other ideas?

The other downside is we
can’t localize it (which we define as “not our problem”).

Why would item types not be localizable? (Zotero item types are
already localized, I believe).

Sure, but I mean not within CSL. We would not be responsible for
maintaining the (long) list of localizations needed.

Bruce

I’d probably opt for your option 2, which would allow markup like:

i don’t think this would work–how can the same item be both
type=“personal_communication” and type=“letter”? Are we talking about
CSL types, Zotero item types, or “genre” variables?

That does what you’re looking for (though is obviously rather
complex).

We’d have to introduce some fallback expectations for the generic
type,
though. Otherwise, we’d probably have to do something like:

See what I mean about complex? I can’t think of a totally easy and
generic solution to this problem.

Any other ideas?

How about adding the term “letter”? That is the easiest solution and
requires one simple <if statement that is already in Chicago CSL.

  <else-if type="personal_communication">
    <choose>
      <if variable="genre">
	<text variable="genre" text-case="capitalize-first"/>
      </if>
      <else>
	<text term="letter" text-case="capitalize-first"/>
      </else>
    </choose>
    <text macro="recipient" prefix=" "/>
    <text macro="issued" prefix=". "/>
  </else-if>

Elena>>>

Elena Razlogova wrote:

I’d probably opt for your option 2, which would allow markup like:

i don’t think this would work–how can the same item be both
type=“personal_communication” and type=“letter”? Are we talking about
CSL types, Zotero item types, or “genre” variables?

I’d say we’re talking about the user-defined description first, and if
not present, the type label (as determined by whatever implementation,
including Zotero).

That does what you’re looking for (though is obviously rather
complex).

We’d have to introduce some fallback expectations for the generic
type,
though. Otherwise, we’d probably have to do something like:

See what I mean about complex? I can’t think of a totally easy and
generic solution to this problem.

Any other ideas?

How about adding the term “letter”? That is the easiest solution and
requires one simple <if statement that is already in Chicago CSL.

It is “easy” but it is not a general solution. I want to avoid having to
define a term for every possible type, medium, etc. we might need.

Bruce

Okay, that wasn’t originally my point. I want to change this:

    <choose>
      <if type="interview">
	<text term="interview" text-case="lowercase"/>
      </if>
    </choose>

to this:

    <choose>
      <if type="interview">
	<text variable="type" text-case="lowercase"/>
      </if>
    </choose>

to display the name of the item type from the
statement. Currently it’s not possible. Can this be made possible in
CSL (without any Zotero programming)? Why not?

Thanks,
Elena

Okay, that wasn’t originally my point. I want to change this:

        <choose>
          <if type="interview">
            <text term="interview" text-case="lowercase"/>
          </if>
        </choose>

to this:

        <choose>
          <if type="interview">
            <text variable="type" text-case="lowercase"/>
          </if>
        </choose>

I’m not sure this would work exactly without tightening up the definition.
type is actually sort of multivalued. As you know, for a given object,
type=“interview” and type=“article” might be simultaneously true. In which
case would variable=“type” give “interview” or “article” or indeed
"radio_interview", “conversation” or something else in the hierarchy.

could give all manner of things for instance.

Julian.

Okay, that wasn’t originally my point. I want to change this:

        <choose>
          <if type="interview">
            <text term="interview" text-case="lowercase"/>
          </if>
        </choose>

to this:

        <choose>
          <if type="interview">
            <text variable="type" text-case="lowercase"/>
          </if>
        </choose>

I’m not sure this would work exactly without tightening up the
definition. type is actually sort of multivalued. As you know, for
a given object, type=“interview” and type=“article” might be
simultaneously true. In which case would variable=“type” give
"interview" or “article” or indeed “radio_interview”,
“conversation” or something else in the hierarchy.

could give all manner of things for instance.

Ah, thanks! Now I remember. Well, we’re back to having terms for both
"interview" and “letter” until something can be done within Zotero
for this.

Elena

I’d probably opt for your option 2, which would allow markup like:

That does what you’re looking for (though is obviously rather
complex).

Now that I understand what you’re talking about, would it be possible
to add a variable “original_type” to store the type coming from
Zotero or any other implementation? It would solve most of my
problems, I think, without any Zotero programming, and would give
flexibility to any other biblio project that would want to use CSL.

We’d have to introduce some fallback expectations for the generic
type,
though. Otherwise, we’d probably have to do something like:

Well, I have conditions like that now for other things–I don’t think
this kind of complexity can be avoided.

Best,
Elena> See what I mean about complex? I can’t think of a totally easy and

Elena Razlogova wrote:

Now that I understand what you’re talking about, would it be possible
to add a variable “original_type” to store the type coming from
Zotero or any other implementation?

Explain again what “original_type” would mean?

Bruce

Well, if I understand Julian’s explaination correctly, “type” may
contain several values at once for a given item, for example for this
item:

George Creel to Colonel House, September 25, 1918, Edward House
Papers, Yale University Library.

original type = “letter” (user-defined, in this case Zotero)
main type = “personal_communication” (CSL)
fallback type = “article” (CSL)

Zotero maps its item types to CSL item types, but in addition could
we assign the user-defined item type to its own variable
"original_type" (or “user_type,” whatever)? Then the relevant CSL for
a footnote citation would look like this:

    <else-if type="personal_communication">
      <group prefix=", " delimiter=", ">
	<choose>
	<if variable = "genre">
          <text variable="genre"/>
	</if>
	<elseif original_type="letter">
	</elseif>
	<else>
          <text variable="original_type"/>		
	</else>
	</choose>
	<text macro="issued"/>
      </group>
    </else-if>

This would work for Zotero “e-mail” and “instant message” types.
Moreover, if another (hypothetical) team creates their own
bibliographic program and decides to add a bunch of other item types
for personal communication, such as “telegram”, or “secret code
message”, or whatever, mapping these to CSL "personal_communication"
would automatically result in a correct citation.

But I’m still not sure I understand correctly how “type” variable
works–let me know if I am wrong about this.

Best,
Elena

Well, if I understand Julian’s explaination correctly, “type” may
contain several values at once for a given item, for example for this
item:

George Creel to Colonel House, September 25, 1918, Edward House
Papers, Yale University Library.

original type = “letter” (user-defined, in this case Zotero)
main type = “personal_communication” (CSL)
fallback type = “article” (CSL)

I think Julian’s point is even more basic: even an implementation like
Zotero could allow more than one “original type.” An interview can
also be an article.

This is a good point, and might present a problem.

Zotero maps its item types to CSL item types, but in addition could
we assign the user-defined item type to its own variable
"original_type" (or “user_type,” whatever)? Then the relevant CSL for
a footnote citation would look like this:

        <else-if type="personal_communication">
          <group prefix=", " delimiter=", ">
            <choose>
            <if variable = "genre">
              <text variable="genre"/>
            </if>
            <elseif original_type="letter">
            </elseif>
            <else>
              <text variable="original_type"/>
            </else>
            </choose>
            <text macro="issued"/>
          </group>
        </else-if>

This would work for Zotero “e-mail” and “instant message” types.
Moreover, if another (hypothetical) team creates their own
bibliographic program and decides to add a bunch of other item types
for personal communication, such as “telegram”, or “secret code
message”, or whatever, mapping these to CSL "personal_communication"
would automatically result in a correct citation.

But I’m still not sure I understand correctly how “type” variable
works–let me know if I am wrong about this.

I think you’re on the same track as I. But we need to think about this
a little more in lieu of the point up top. I’ll try to do that, but
it’s finals week.then I have a manuscript deadline.

Bruce

OK, on types and such:

I think the solution is going to be a little verbose, but for
simplicity’s sake should be something like what Elena was originally
thinking:

I don’t believe that requires any schema changes, aside from adding
needed terms.

Will that work? Shall we go with it and see if that works OK?

Bruce

PS - Might be good to integrate the “note” variable into the macro; not
sure.

OK, on types and such:

I think the solution is going to be a little verbose, but for
simplicity’s sake should be something like what Elena was originally
thinking:

I don’t believe that requires any schema changes, aside from adding
needed terms.

Will that work? Shall we go with it and see if that works OK?

yes, this is pretty much how it works now–julian already added the
terms. thanks for resolving this.

Bruce

PS - Might be good to integrate the “note” variable into the macro;
not
sure.

FYI–note is currently mapped to “extra” in Zotero, so it doesn’t
really fit here at present.
Elena