## Question types: Correct writing

qtype_correctwriting
Maintained by Oleg Sychev
Correct Writing question type are focused on the learning grammar of the new language. It can found various mistakes in student's response and give mistake messages and hints.
92 sites
13 fans

Correct Writing question type are focused on the learning grammar of the new language. It can found various mistakes in student's response and give mistake messages and hints.

It can find token sequence mistakes: absent tokens, extraneous tokens and misplaced tokens. Token is a minimal language unit like word, punctuation mark etc. For now two language supported: English language and C programming language.

Also it can find typos, including missing and extraneous spaces between words.

You can download this question type along with all subplugins, necessary for it, using our BitBucket site

### Sets

This plugin is part of set Correct Writing Question Type.

### Contributors

Oleg Sychev (Lead maintainer): Idea, string analysis method, general architecture, hints implementation
Dmitry Mamontov: Question type code, sequence analysis, picture generation

• Wed, 13 Mar 2013, 9:32 AM
Oleg - Many thanks for sharing the correct writing question type. I was able to install without any issues. One concern I have is about the dependencies. This question type is listed as being for Moodle 2.4 and many of the dependencies do not show having a version for 2.4. We may want to be a little careful to ensure that everything seems to match up with respect to the versions (i.e. a 2.4 version of each plugin being available). I know you have a link to have them all together but if there begin to be differences between the versions that system may fall apart without carefully managing the branches. I presume that you want this plugin added to the PERL set. I will go ahead and approve this plugin and add it to the set. Please do not hesitate to let me know if there is anything I can do to be supportive of your efforts. Peace - Anthony

p.s. - Any objections to changing the URL for the source code to https://code.google.com/p/oasychev-moodle-plugins/source/browse/#hg%2Fquestion%2Ftype%2Fcorrectwriting which is more specific (rather than the more generic one that points to the root of the repository)?
• Tue, 6 Aug 2013, 5:05 PM
'It can found various mistakes in student's response' - very clever!
• Sat, 8 Nov 2014, 3:04 AM
There is some bug in string_pair.php. Look at the first line. It should be <?php not <?. It makes problems in strickt mode.
• Wed, 26 Nov 2014, 3:28 AM
In the manual picture there are 7 penalties to configure. I've installed current version and can find only 4 of them. Where are the missing penalties?
-missing token
-extra token
-absent token mistake hints
In this configuration a forgotten ")" in student's answer of for ( int i = 4; i <= 10; i++)
leads to 33% minus. This ist too strict.
• Wed, 26 Nov 2014, 4:22 AM
Sepp, penalties with green asterisk are "advanced" and hidden by default.
did you see "Show advanced" or similar button on the form? Press it and you will see all settings.
• Wed, 26 Nov 2014, 6:20 AM
Sorry, no such button. Some of green-asterisk-entries are visible, e.g. :
-Penalty for misplaced token
-Maximum percent of mistakes
• Sat, 6 Dec 2014, 12:47 AM
Thanks, Sepp. Two penalties was really disappeared from the question editing screen during internal refactoring. A placed a new version there (2.6.1.) that should make them visible, please try it.
• Tue, 9 Dec 2014, 7:57 AM
Thank you. Now it works! I can set the penalty for missing token.

But, ...
not every token is as important as each of the others. Is it possible to add an indiviual missing penalty in the description token window? eg. separated by ~
for: for command~10%
( : bracket~5%
int : data type~20%
and so on

Sepp
• Fri, 9 Jan 2015, 5:37 AM
@Khac: Edit first line of string_pair.php (and other php files) from "<?" to "<?php"
• Fri, 2 Oct 2015, 5:59 PM
Is there a possibility to use another language istead of english (I do not mean C/C++ etc.) - german for example?
• Fri, 2 Oct 2015, 6:25 PM
It's depends on the language (or words used in answer to particular question). CorrectWriting use language to tokenize input - i.e. break answer down to the tokens - words, punctuation marks etc. That may not be always intuitive (cf. single apostrophe in English at the end of the word which can be a quote or a part of word). If the rules of tokenizing German (for you answers) are identical with English you can just use English in the question. If not, that will require some additional coding either from you or from us. I would expect problems with German letters that absent in English, I do not know about other details of German.

We can be contacted about developing support for other languages, thought considering other tasks we can't guarantee to do it fast. Thought if you are researcher, there are ways to motivate us to work faster under support of some language.
• Mon, 17 Oct 2016, 9:16 PM
These type of tool are really so needed while writing to ensure that your content are error free especially if you’re the type that can hardly spot out errors on your writing.

I am using this plugin as a sentence checker and improved my grammar as a writers to AceWriters.org writing service.

All Best!
• Thu, 29 Dec 2016, 8:29 PM
Very clever plugin! Helps me a lot especially when I'm i a hurry and letters appear in a jumble.
But the real brain is still the best grammar checker and unfortunately even it makes mistakes. And writing tons of papers and essays for students in customwritings review center this plugin is a great help.
If there is a German language checker it will be priceless. Hope you'll add additional coding from your side.
• Sat, 14 Oct 2017, 9:02 PM
Thanks for this very helpful plugin. We're going to use it for English as is but we have other languages we would like to use it with, say French, Spanish or German. As is, the tokenizer for English cuts on every accented character which renders the generated tokens unusable. Would it be possible to integrate a more relaxed tokenizer as an option. We would also like to be able to use your plugin for exercises about phonetic transcriptions. The IPA alphabet uses the standard characters plus a specific table in the Unicode space. Rather than making a complex tokenizer for each and every case, would it be possible to allow question writers to tokenize manually?