How to do an anstype
# How to add a new anstype (called here `_name`)
Each anstype contains two files, `_name.input` is the input form element, and
`_name` is the answer processing file.
Specific anstypes need not be put into this directory.
One can create a subdirectory "anstype" in the module, then put the files there.
Any name will do, but we recommend that module-specific anstypes use a specific
prefix for its names, like "mynumeric", "myfunction", ..., in order to avoid
possible confusions.
In an oef exercise, the anstype is called by the line
`\answer{xxx}{yyy}{type=_name}{option=zzz}`
see later for the intern variables corresponding to the values.
The input file should contain the following two definitions:
1. define `anstype=yes`. Otherwise the OEF manager will not recognize the
existence of the anstype.
2. define the form style using the variable `anstyle`. Here are the possible
words to declare in this variable:
* `mc`: multiple choice style.
* `symtext`: allows symtext processing.
* `numeric`: the data is numeric or with an input field
* `dprompt`: correct answer prompt given under double-underlined variable (internal `reply__$i`);
in this mode, the correct answer is not present by default.
* `nogood`: never present correct answer (because there cannot be).
* `noanswer`: never present the answer analysis (because there is no correct or wrong answer)
These scripts accept the following input variables:
* `i`: The number of the reply field.
* `replyname$i`: The prompt of the form element.
* `reply$i`: The answer given by the student.
* `replygood$i`: The correct answer given by the author.
so the value of the second field of `\answer{}{}`.
* `replyoption$i`: Option words. Anstypes are more or less free to
define their own recognized option words.
In oef exercises, it correspond to the value `zzz` in `{option=zzz}`
* `inputsize`: the second item of the field \embed{} (the first one is
* `r$i` ou `reply$i`). In general, it contains the size of the form element.
Moreover, if no comparison is wanted, the answer processing file can have a
calling parameter "nocompare" (not useful in oef exercices).
Output variables of the answer processing file:
* `Test`: Put `bad $i` if the correct answer given by the author
is not understandable.
* `test`: Put `NaN` followed possibly by an error indicator,
if the answer given by the student is not understandable.
some available error indicators:
* `badsize` (following by a number)
* `badform`
* `badform noreduced`
* `nocompute`
* `nocomputecomplex`
* `notunit`
* `bad_variable`
* `unknownword` followed by some words
* `chemdraw_empty_data`
* `chemclick_empty_data`
* `diareply$i`: Diagnostics result. Possible values: `good`, `bad`.
* `precreply$i`: If the answer misses the correct result just by a
problem of precision, put `yes` to this variable and put
`bad` to `diareply$i`. In the computation of the score, a coefficient
depending on the level is introduced. In thise case, one must
not advance `freegot` at all.
* `partialgood$i`: If the answer should be considered as partially correct,
put `yes` to this variable and put `good` to `diareply$i`.
It has no consequence on the calculation of the real score (use `freegot` for that).
* `freegot`: This is a numerical variable. Its content is incremented
by 1 if the answer is 100% OK, none if it is false,
or anything between 0 and 1 for partially correct answer.
It is not recommended to decrement this value, nor to
increment it more than 1. The real score is computed using it.
So,
`partialgood$i=yes + diareply$i=good + freegot` advances of a number strictly between 0 and 1
or
`!advance precgood + diareply$i=bad + fregot=0 + precreply$i=yes`
There are also some optional output variables (they can be left empty).
* `replyGood$i`: This variable can be defined for the correct answer
shown to the students after replying to the exercise,
It should be different than the real `replygood$i`.
It is not shown in the analysis of the answer for false student
answer if dprompt is declared in anstyle.
* `m_reply$i`: This variable can be defined for the answer shown back
in the oef variable `reply$i`, if it should be different than the
real answer typed in.
* `reply_$i`: The answer of the student shown in the html page in place
of `\embed{}` or in the analysis zone (for oef exercises).
* `reply__$i`: A special prompt in the analysis of the answer
to show when `dprompt` is declared in anstyle. If it is empty, its
value is the value of `reply_$i`. If the variable `noshow$i` is yes,
the student reply is not shown in the answer analysis.
* `m_sc_reply$i`: by default 0, 0.5 or 1 according to `diareply$i`, `partialgood$i` and
`freegot`. Can be set to special values. Give the possibility
to the developer of exercises to give feedbacks without testing again.
There are also some variables which can be used (new/2009)
* `oef_formnosubmit`: if `yes`, no "Send answer" button appear.
* `oef_js_submit`: can be used to put javascript in the answer button,
the line `onclick="javascript:$oef_js_submit;"` will be present
in this button (in test).
* `oef_anstype_css`: style css (what is inside ``)
will be include at the end of the web page.
If the following variables are defined in the file `_name.input`, they can be reused
in the file `_name`:
`oef_applet_option`,`oef_applet_command`,`oef_fill_option`,`oef_answer_option1`...,
(for reuse applet options or embed options for example, if one wants to use the anstype several
times in the same html page, it is recommanded to use `oef_answer_option$i` indexed by the
answer's number).
If you use an external software, use the variable `ans_require` (best if there is a version number).
Add the help in `scripts/help/anstype` and `scripts/help/anstype.$lang` and in the list
`help/$lang/reply.phtml`.
See the lists of the anstype in the OEF documentation even if they can be used
in any pedagogical modules (with adaptation of the syntax).
Language
