Please add a general purpose element for theorems,
definitions, and similar displayed text blocks.
This element, say <theorem>, should have the following
content model:
(head?, p+)
Its attributes include besides global ones:
* "type" (e.g., "theorem" or "definition")
* "typeN" (the theorem number, as in "Definition 3")
The "n" attribute of <theorem>, however, should be
reserved for specifying running numbers of displayed
blocks in general.
Consider the following example:
<theorem n="12" type="definition" typeN="3">
<head>Multiplication</head>
<p>...</p>
</theorem>
which could be rendered as:
(12) Definition 3 (Multiplication)
...
Lou Burnard
2004-08-20
Logged In: YES
user_id=1021146
What is the defining characteristic of a "theorem"? If it's
simply that it has to be formatted in a particular way, maybe
it would be better to specify "theorem" as a value for the
rend attribute on some more generic element such as <q> or
<eg> or <formula> ? (OK, I know, <eg> isn't in P4 yet...)
Alternatively, if <theorem> is actually intended to have some
semantics, could you spell them out a bit for us? Not the
same as the existing <formula>, I assume.
Andreas Nolda
2004-08-20
Logged In: YES
user_id=950793
Theorem-like entities are well-established structural units in
scientific (in particular, logical and mathematical) texts.
(There are several LaTeX implementations providing
environments for them.)
Semantically, theorem-like entities are theoretical sentences
(definitions, axioms, theorems, proofs, etc.). They are
neither quotations (so <q> is not appropriate) nor examples
(which could be tagged as <eg>). Unlike <formula>e,
'theorems' are often formulated in natural language.
In general, 'theorems' are formatted as a block.
Normally, a 'theorem' has a label such as "Definition" or "D"
specifying its type. In many cases, canonical references are
constructed from the label and a proper counter ("Definition
3" or "D3").
Sometimes, 'theorems' are numbered instead by the general
counter used for all numbered displayed block
elements--including <formula>e--(e.g. "(12) Definition").
There are also cases, where both numbers are given ("(12)
Definition 3").
Finally, 'theorems' can have an optional header, supplying
some characterization or 'nick-name' for it, for example:
"Definition 3 (Multiplication)" or "Theorem 5 (Gdel's
Incompleteness Theorem)".
Lou Burnard
2004-08-22
Logged In: YES
user_id=1021146
Thank you for the clarification. There are analogous
structures in e.g. legal and linguistic texts, so I think I agree
that it might be helpful to add this concept to the TEI's world
view. I'm not sure where to put it though.
Syd Bauman
2004-09-20
Logged In: YES
user_id=686243
I agree that a theorem needs someplace to go in the TEI
scheme. For the time being (i.e., for encoders using P4),
probably an <ab type="theorem"> is the way to go. I think
it's worth trying to collect the set of things that tend to
be treated in the same way (not part of the running prose,
but part of the main text unlike a <note>; generally occur
at a specific spot in the text, although may float and be
referred to from the main text; may have a heading, and may
be referred to be a number or other label; often rendered by
being indented on both sides) and seeing if we can extract
any commonalities. (Andreas has already done this for us, to
some extent, with "theorem" and "definition". :-) This might
permit a single element to handle them all (perhaps with a
type= attribute).
I don't see the need for the typeN= attribute nor
restricting the semantics of the global n= attribute for
<theorem>.
Lou Burnard
2004-09-22
Logged In: YES
user_id=1021146
I wonder if this <theorem> thing is a possible child of the
redefined <figure>, along with <eg>?
Lou Burnard
2004-09-22
Andreas Nolda
2004-09-27
Logged In: YES
user_id=950793
Regarding Syd Bauman's comment on the proposed "typeN"
attribute: As my original example shows, theorems can have
up to two numbers: a theorem-type related number ("3" in by
example above) and a common running number for displayed
elements like formulae, examples, and theorems ("(12)").
That may be bad style, but can nevertheless be found in
real-world papers.
So, if the encoder (or the author, for that matter) wishes
to specify both numbers of the theorem explicitly, he needs
two attributes for that purpose.
Andreas Nolda
2004-09-27
Lou Burnard
2004-09-27
Logged In: YES
user_id=1021146
But surely this kind of numbering scheme (where individual
items get a special number, as well as the general sequence
numbering) is not specific to theorems? A picture might, for
example, be Figure 123 with the title "(3) A moonlit scene"
Would a <label> within the <head> not be a more generic
approach to solving this kind of problem?
Andreas Nolda
2004-09-29
Logged In: YES
user_id=950793
The problem is that, according to my view, <head> should be
optional:
<theorem n="12" type="definition" typeN="3">
<p>...</p>
</theorem>
corresponding to:
(12) Definition 3
...
The 'bare' label "Definition 3" can be generated from the
"type" and "typeN" values.
Syd Bauman
2006-09-25
Lou Burnard
2006-09-25
Lou Burnard
2006-09-25
Logged In: YES
user_id=1021146
We agree that <theorem> is a useful new element, even if it
is only as syntactic sugar for <ab type="theorem">. The
question is, where should it be documented? The most likely
candidate module seems to be FT, unless we add it to the
core. I think the numbering issues are less important, and
are not restricted to theorems anyway.
SourceForge Robot
2006-10-10
Logged In: YES
user_id=1312539
This Tracker item was closed automatically by the system. It was
previously set to a Pending status, and the original submitter
did not respond within 14 days (the time period specified by
the administrator of this Tracker).
SourceForge Robot
2006-10-10
Lou Burnard
2006-10-10
Lou Burnard
2006-10-10
Logged In: YES
user_id=1021146
Reopened pending further discussion
Lou Burnard
2007-02-01
Lou Burnard
2007-03-16
Lou Burnard
2007-03-16
Logged In: YES
user_id=1021146
Originator: NO
After further discussion on the TEI Council list, there is general agreement that this proposal requires more substantial work before it can be considered for inclusion in TEI P5. The idea is sound, but the details need to be worked on further to make clearer how it fits into the existing structure. The proposal is to use <ab> for the moment, and to work more on this for releases subsequent to P5 1.0