Moodle in English: algebra evaluator added to cloze type question

(I posted this in the mathematics tools forum, but I am told it should be posted in this Quiz forum) Sorry if you get it twice...

After being bugged by me for some time, my husband (Greg Mushial – a software writer of 43 years, and member of this Moodle community) has put together an algebraic expression evaluator such that my students can enter math expressions in different formats and be appropriately graded by Moodle if the answer is equivalent. Although he develops most of his software in Delphi, he accepted the challenge to work in PHP, and put together a module that hooks into the “SA” of the cloze type questions and evaluates and compares expressions. This question type is my absolute favorite since it gives me freedom from the multiple choice answer format, and allows the student “creativity” in how they answer mathematical expressions.

The code was developed in Moodle 1.9.15 and I have used it in my courses that run on 1.9.10, but given its structure (hooked to the existing SA cloze type question) we believe it should also run in version 2.* (which we will be doing shortly). It is not as powerful as what I have seen of STACK, but it has allowed me a new “dimension” for my math and physics quizzes. It works by interpreting arithmetic expressions, and algebraic expressions in ONE variable (which must be entered as “x”).

I have tested the code extensively in my limited universe of math (and some physics) expressions but we are always worried that some “undocumented features” (bugs) may be too well disguised for us and have escaped out notice. So it would be extremely helpful if some interested instructors could try their own questions and let us know of any problems they observe.

The “equivanswer.php” module hooks into the “SA” (Short Answer) module that serves the “cloze” type questions in Moodle. There is a couple line hook that is added to the short answer module to identify an escape sequence immediately following the equal sign for the expected correct answer. When the “SA” question type contains said sequence, the “equivanswer.php” module is called to perform the evaluation of the entered expression at the numerical values specified by the instructor (when they build the question).

When the student responds to the question, the evaluation (at specified points of the expression’s domain) of the correct answer (entered by the instructor) and that entered by the student, are performed and compared. If the values match within a tight epsilon, the answer is marked as correct.

The person that designs the question must therefore give sufficient and well selected evaluation points within the domain of the expression for the comparison to be of value.

For example, if the expected answer involves a square root, the selection of points for evaluation must produce non-negative radicands; a point of evaluation must not produce a zero in the denominator of a rational expression; etc.

The syntax in this modified form of “Short Answer” questions therefore includes: the escape sequence to identify the answer as an equiv answer type; the expected algebraic expression; the symbol “@” followed by the evaluation points for the algebraic expression (these are comma separated). An example of the syntax is given below:

equivanswer syntax

The student enters the algebraic expression in the “fill the blank” box using the same notation he/she uses in a calculator or graphing tool, and as long as the algebra is correct/equivalent, Moodle will recognize it as such :

student response

I am adding a link to a pdf document with more details on the recognized functions and screen captures on different answers accepted by Moodle as correct.

If there are more instructors interested and willing to try their questions and provide feedback to us, please let me know, and I will add you as a “teacher” in our testing course site. We would like to release a well tested module to the open source Moodle community, thus reducing unexpected results for the users.

Thank you in advance to all those willing to help.

Cristina Berisso and Greg Mushial

Average of ratings: Useful (7)

Re: algebra evaluator added to cloze type question

by Marcus Green - Tuesday, 10 January 2012, 5:49 AM

This makes me wish I taught maths

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Joseph Rézeau - Tuesday, 10 January 2012, 6:56 AM

This makes me pleased I never taught maths.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Philippe Decloitre - Saturday, 28 January 2012, 7:15 PM

This makes me wish you'd answer MY questions

Cheers

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Pierre Pichet - Tuesday, 10 January 2012, 9:09 AM

It is always a pleasure to see such new ways or new tools that are created by real teachers.

This reflects also the vitality of the Moodle community.

Pierre

P.S.-1 This is somehow the math version of the Joseph regexp project.

P.S. -2 Sorry , my todo list does not allow we to help you in the testing

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Bernard Boucher - Tuesday, 10 January 2012, 9:27 AM

Hi Cristina and Greg,

great idea and excellent work

I will be pleased to test some questions in your testing course site and give you feedback.

Salutations from Québec,

Bernard

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Colin Fraser - Tuesday, 10 January 2012, 8:54 PM

Well I do teach maths, so maybe I can give some sort of assistance that might actually be useful, but don't count on it....

I take yrs 7, 8, 9, 10, but not all in the same year, and rarely more than 2 Maths classes a semester, more General Science than anything else.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Bernat Martinez - Tuesday, 10 January 2012, 11:45 PM

Congratulations!! this is very tool for maths teaching. Yes, I would like to test it

We have tried to do something like this using the SWF module and the "free" AS3 classes from flashandmath

Some examples can be seen at http://itemspro.net/test/ using test/test

Obviously your soultion is more general and easy to use.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Joseph Rézeau - Wednesday, 11 January 2012, 12:42 AM

Hi Cristina,

What I don't understand is why you went straight for a short-answer sub-question type in the Cloze question type.

Does your "algebra evaluator" question work on its own, i.e. as a stand-alone question similar to the Short-answer, MCQ, match, REGEXP, PREG etc. questions types?

It seems to me a more logical way to first develop a standalone question type, and then try to integrate it into Cloze. Unfortunately, and this has been a long-standing request of mine, even in the moodle new 2.1 question engine, it is still not possible to automatically integrate a standalone 3rd-party question type into Cloze; you have to hack the Cloze files to do do that.

Joseph

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Cristina Berisso - Wednesday, 11 January 2012, 1:22 AM

Hi Joseph,

I'll leave the technical part of the answer to Greg (I am quite ignorant of the code behind this module), but I can tell you how the thing developed:

Greg asked me which type of question I had in mind for my courses. I answer: "something similar to short answer, but that would process the math expression entered by the student, evaluate it over a finite number of points and check against the correct answer given by the instructor."

I told him that for the time I had been using short answer questions in Moodle to ask student to enter very simple linear expressions of the form "mx+b" (with "m" and "b" integers or in decimal form and requesting the student not to type any spaces with the answer) and that it ended up working quite well (the students promptly adapted). But I wanted to go to somehow more complicated algebraic expressions. My absolute favorite format is the "cloze" question type. It is more flexible for math settings, for example you can have text in line with and immediately before and after the answer box. This gives wonderful posibilities like preceding the answer box by : "f(x)=" (so the student understands that the "f(x)" part should not be included in the answer). Also for physics questions, the box can be followed by the appropriate units required in a numerical answer, like in:

"Resultant force = [answer box] Newtons"

Cloze also allows to have several different type of answers included in one single question. I can combine answers that required not only the magnitude of a physical quantity but a direction; for example an angular velocity magnitude (numerical "NM" type )followed by a drop-down menu given the options "clockwise" or "counter-clockwise." I didn't want to loose all these wonderful options with a completely separate module.

Greg looked at the code and came back saying that there could be a simple way of hooking another piece of code to it to get something close to what I wanted.

Again, I'll ask him to reply to the technical aspects of your question.

Have a wonderful day!

Cristina

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Thursday, 12 January 2012, 8:49 AM

Joseph - the top of the day to you.

I have to agree with the tone of your post; but I have to laugh a bit also... there are well thought out, carefully crafted, carefully implemented pieces of code... then there are those: I need it yesterday implementations.

In this case the idea went from talking about "it would be nice if.." after dinner on the 29th of December, to my downloading a 1.9.15+ distribution that evening before going to bed; to reading through the code over the next couple of days and coming to the conclusion that what she wanted was possible (understanding that it would have to be working and trusted well before the beginning of the next semester - 17th Jan); to setting up a server and installing 1.9.15+ on the 2nd of Jan; to having a first working copy on the 5th; and after her testing it for a couple of days and my fixing two bugs; to her concluding that the code/function might be of use for others and posting as she did on the 9th - or, 1300 lines of code in a petite and non-obtuse system (Moodle) in a couple three days, in addition to carrying on with nominal 9-5 endeavors.

W/re the locale and implementation of the function - for something that was intended and designed for her use (not thinking beyond that) - I don't think where the hook was put in was mischosen: it simply parallels the call to compare_string_with_wildcard function, and likewise returns T or F. W/re alternative implementations - sure: creating an new question type - equivanswer might be a choice - though it might carry an excessive amount of duplicate baggage - extending the question default class; or one might add a comparator option to shortanswer question type: use match with wildcard or use numerically equivalent comparisons. Given the amount of (positive) feedback she's gotten - yes, probably a more distributable implementation should be considered. I just setup an additional server to run 2.2+ where I'll add the EA type. Given what I find there, integrated with what we've leaned from 1.9, then one might have enough information to decide on a future implementation. [though today we were talking about extending the function to multivariant, which again will have to be factored into any future versions.]

W/re the actual implementation: it adds an "include_once" to the shortanswer questiontype.php; adds an if statement choosing btwn the existing call to compare_string... and possibly calling the equiv code, ie, about 5 lines of change to existing Moodle code; everything else is implemented in a separate module (pointed to by the include_once).

I/we'll keep all posted as this progresses.

enjoy, and happy Moodling - greg

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Joseph Rézeau - Thursday, 12 January 2012, 4:55 PM

Greg,

Thanks for your detailed reply. I can see that your Xmas break was well put to use.

Unfortunately my maths skills are too basic for my understanding the intricacies of your algebra evaluator. In the examples provided by Cristina, I do not understand at all what the following mean:

"The person that designs the question must therefore give sufficient and well selected evaluation points within the domain of the expression for the comparison to be of value."

What are those "evaluation points"?

"When the “SA” question type contains said sequence, the “equivanswer.php” module is called to perform the evaluation of the entered expression at the numerical values specified by the instructor (when they build the question)."

Does this mean that, in the "triangle area" problem, a value for "x" is provided? where?

How are those mysterious (to me) 0.3,1.5,4.1 "evaluation points" related to the algebra formula?

I'd be grateful if you could shed some light on these points, but do not spend too much time trying to raise my abysmally low maths level. It is enough if maths-savvy moodlers understand how your algebraic expression evaluator works and can test it and find it useful.

Joseph

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Cristina Berisso - Friday, 13 January 2012, 12:53 AM

Hello Joseph,

I’ll try answering your question with two examples:

Do you remember graphing “linear equations” in school? These are equations that you could write in the form:

y = m x + b ( the y-variable that you plot using a vertical scale and the x that you plot using a horizontal scale on your paper, while “m” and “b” where given numbers).

For example : y = 2x-3. The graph of such an equation would render a “line” (the reason for their name “linear equations”). You may also remember that in order to draw a line on a plane, all you need is to have two points plotted and then use a ruler and draw the straight line that connects them. That means that all you need to completely define your line is "two" well-determined points.

Well, what our evaluator does is to evaluate the right hand side of the equation above (2x-3) at "two" x-values that the instructor gives, thus finding the two y-values that go with these selected x’s. The code then uses the same x-values provided by the instructor, to evaluate now the expression that the student enters and sees if the y values match the previously calculated ones (found based on the instructor’s algebraic expression).

In the case I showed of the area of the triangle, the equation that renders the area is: “base times height divided by two,” which could be written as:

y = (3x)(2x+1)/2 = (6x^2 +3x)/2

This equation is not linear since it carries a term with the “square” of the variable x (x to the second power), so its graph is not that of a line, but of a curve that you may remember under the name “parabola.” In order to define perfectly well the graph of your parabola on a plane, you need now to provide “three” well defined points. Now, our evaluator will evaluate the right hand side of the expression above (that contains the x squared) at (hopefully) three x-values provided by the instructor, to find the three associated correct y-values. These are the three “points for evaluation” that I show as: 0.3, 1.5, and 4.1. If you place these x-values in the formula, you get: y = 0.72, 9, and 56.58 respectively.

If the response of the student is not the same (or equivalent) to the algebraic expression provided by the instructor, the y values will not match.

In the image below I show the graph that illustrates such a case. The “correct” answer for this “parabola” is that represented by the blue graph. You can see two of the three points found in the paragraph above and how they are right on top of the correct curve (blue dots on the blue curve). The orange curve is the parabola that results from the algebraic expression entered by a student who forgot to “divide by two” when typing the area formula. When the calculator evaluates the student’s expression at the x-value “0.3” it will find the y-value represented by the red cross on the graph, which is clearly different from the one associated with the blue dot right below, and will mark the answer as Incorrect.

two different area expressions

It is important that the instructor provides enough number of points where to test the algebraic expression entered by the student, so no matching by pure coincidence occurs.

Let me know if this message made things a bit more clear or if it contributed to create more mental entropy…

Have a wonderful day!

Cristina

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Friday, 13 January 2012, 1:18 AM

Joseph,

Again, the very top of a fine day to you.

The most basic tenant of the equivanswer code is: two functions (answers) are equivalent, if their graphs are coincident/indentical... so what the EA code tries to do is to establish this as economically as possible. While it could exhaustively compare the graphs (values) of the answer function and the response functions for all possible values; or it could compare them at random (x) values - what it requests of the instructor/question creator is: to provide a set of x values at which to compare the two, such that to a high degree of confidence (4 or 5 sigma) the evaluation would attest to the equivalence of the answer and responce functions. Those x values are listed after the "@" in the answer definition, by the question creator (QC). The number of them is left to the QC - in Cristina's examples she uses 3, which if well chosen should provide 4 sigma confidence; but there is no reason why she couldn't have used 5 or 10. [the implementation of EA "compiles" the functions, so the evaluation of each at each x value is very quick, so more than three testpoints is probably less expensive than a single regexp call.] In the multivariant follow on, where the instructor will be able to specifiy the values of any number of variables and their evaluation points - although the code hasn't been written yet - the expected syntax is: to express the set of variables for which substitutions are to be made, as a tuple, followed by tuples expressing the values for those variables at each evaluation point.

W/re maths skills for a man that could have written Language and Mind or Aspects such is without liability

W/re time well spent: as an absolute believer in education and especially on-line education: couldn't have spent the time more productively

again, the best to you - greg

Average of ratings: -

Re: algebra evaluator added to cloze type question

by khaled baloul - Friday, 13 January 2012, 7:10 AM

Interested, please add me for the testings.

Khaled

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Jeff Forssell - Monday, 16 January 2012, 1:49 AM

This looks great!

Does it still support the usual features of CLOZE: alternative answers and feedback? Like, would this work?

Area={1:SA:=(=)3x*(2x1)/2@-0.3,1.5,4.1#That's great!
~%20%(=)3x*(2x1)@-0.3,1.5,4.1#You seem to be on the right track, but missed the important ½ factor. Try again!}

(Yes, I am so wimpy that I'd consider giving partial points for a bad answer like that! At least the student has tried and given me something to react to.)

What determines the acceptable precision in the answer? An example in the pdf file has the answer 7/4. What happens if the student answers 1.7 ? Can one add other answer options in SA form?

~%50%1.7*#That's close, but we'd like an unrounded answer. Try again!

I'm assuming one can't use CLOZE numerical format since this is declared as SA.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Cristina Berisso - Monday, 16 January 2012, 7:45 AM

Hello Jeff,

Quite cogent and timely observations! Actually, yesterday I was talking to Greg with regards to the inclusion of partial credit as the standard SA has (the current added code does give the feedback to the correct question, but does not interpret any partial credit), and also of some symbol to use to control the length of the answer box in the case he is investigating right now: multiple variables (which I will be very happy to use in my physics classes and in “solving for one of the variables in a formula” for basic algebra classes).

Your observation is very important and it would be wonderful to have it implemented for this algebra evaluator. Since I am not the code writer, I cannot answer how difficult it would be to implement such, but Greg will probably give you a thorough answer after looking at the code.

Thank you again for your valuable comments!

Cristina

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Monday, 16 January 2012, 10:30 AM

Jeff - the top of the day to ya...

To your question w/re precision: the code takes two approaches to deciding if the two answers "are the same" - if the average of the abs of the correct answer and response are less than 10e-5, then they need to be within 10e-8 of each other, ie, abs(answer-response) less than or equal 10e-8 to be judged the same (correct); conversely, if they are less close to zero than that, then they need to match within 10e-8 * the average of the abs of the answer and response. All math internally is done in double precision. This "funny" way of checking answers is an attempt to mitigate machine representation errors and function library inaccuracies resulting in correct answers being marked incorrect.

W/re the other functions working as before - and understanding that I'm new to this code - the only function that I changed was: how a response and a correct answer are compared, so, if higher level code is already parsing the instructor supplied correct or partially correct answers, unless that code is getting confused by our (=) or @, then everything should be as before - if that's not the case, then it is probably getting confused by the new escape sequence and @. If that is the case, then after I get the multivariant working, and as part of a more general answer, I'll see if I can't restore the munged functionality.

Bigger picture: as stated above: this code was initially developed so that Cristina could get on with her teaching load, and although the idea of sharing it was always in the background, her needs were primary. Now that the function/feature has been exposed to the community at large, and seems to have been well received/might be of use to others than just her... I think, in consultation with others more knowledgeable than myself, I need to rethink and revisit the code to: continue to meet Cristina's needs, but also see if it can't be generalized for the needs of the community at large... so: over the next week or so, I'm hoping to get the multivariant version working... then I should take a couple steps back from this, listen to the feedback we/she's gotten, listen to the "it would be really nice if it did x, y and z" from others, and then look to reimplement the code. As such: I would suggest to you, and the others testing/using the code: think about this type of function in terms of: what would you like for it to do for you. For those that want for it to make water run uphill, or turn lead into gold - sorry, but that I'm pretty sure I won't be able to do... but for those ideas/features that are within the realm of possible, I'll see if my codewriting skills can't produce something that continue to meet Cristina's needs, but likewise satisfy the needs of a larger community.

w/re numeric type questions - I didn't touch that code, so anything/everything should be unchanged.

the best to you - greg

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Cristina Berisso - Monday, 16 January 2012, 1:02 PM

I understand that Jeff was not worried about the standard SA working as before (it is, and I have many questions that use the standard SA and give feedback for different possible answers and are working perfectly right). What he was asking was about implementing the "equivalent answer" code so that instructors can give partial credit and feedback for some "close but not exactly right" algebraic responses from the students. I have communicated the misunderstanding to Greg. I belive he needs to look at the upstream code to give an answer on how simple (or complex) such implementation would be.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Jeff Forssell - Monday, 16 January 2012, 11:27 PM

I have now been in and tested AE and am very impressed. Pierres good work with the CLOZE code and this new addition work very well together!

I'm a "feedback freak" and wanted to test that it could display answer specific feedback texts and points for alternative answers. At first I thought it wasn't working, but had temporarily forgotten about CLOZE displaying answer feedback as popups so they are displayed near the various answers.

The only bug I've seen so far is when previewing a question (as teacher, so no show_stopper) you can choose [fill with correct] and the right expression is put in, but with all the (=) and @domain samples. And that answer is not accepted. Though for a teacher understanding the format, it is not difficult to clean away those parts.

What I tried, that does work:

alternative answers are given "partially correct" and partial points
feedback written after #, is displayed (even with HTML like bold and line breaks).
The first alternative doesn't have to be 100% right and both = and %100% are accepted markers for fully right
you can mix AE expressions (both correct and incorrect variants) and usual SA expressions in the same partial question.
~* wildcard feedback is supported
general feedback works as usual
you can have a mix of AE(SA), SA and NM partial questions in one question
you can make a question with just numbers (no variables) and have a student answer with a proper expression and either give it full points OR a little less and have a straight SA answer for 100% that rewards one that is properly rounded off. (In the feedback for the proper numerical expression, I include the "exact" answer and ask them to round it off)

Anyhow I made a 3 question quiz in the test site.

The most important features (for a feedback junkie like me) of CLOZE seem to work fine!

Wishlist:

ability to set tolerance for numerical expressions
ability to set size of answer space (or have a dynamic draggable size)
ability to recognize , as decimal (for Swedes and other countries with that convention)
anything else that might be implemented from my old numerical wishlist

Hey it's 5 pm! What happened to lunch?!

Average of ratings: Useful (1)

Re: algebra evaluator added to cloze type question

by greg mushial - Tuesday, 17 January 2012, 2:11 AM

Jeff -

The very top of a fine day to ya... and many thanks for the quality feedback - hugely appreciated.

w/re the 10,5 for what I'd call 10.5 ... I'm thinking about alternative answer list formats, ie, for us that use "." xxxxx@1.4,1.8,2.9 but for those that use "," xxxxx@1,4;1,8;2.9 I think I can handle that automatically, ie, I can teach my code to see if there are ;'s and it there are, assume they are the item separators and that any commas are internal to an individual value.. I'm assuming the localized version of PHP supports 10,5 as a valid number... correct? Or do I have to translate 10,5 to 10.5 also (if so, not a problem... just need to know what it is I need to do to make this work for a larger audience.)

again - many many thanks for both the kind words and excellent feedback,

greg

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Tuesday, 17 January 2012, 5:00 AM

Jeff -

I did in fact add the code mentioned in my previous reply, ie, if the x values after the @ use "," to denote the decimal point, then separate the values with semi-colons ";" , eg, @10,5;10,7;12,3 ... but for us that use a period to denote the decimal point, the old format continues per before. If you have a chance and could test this, such would be much appreciated (use the testsite Cristina emailed to you).

many thanks - greg

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Jeff Forssell - Wednesday, 18 January 2012, 6:13 AM

Hi Greg,

I should have described my wish more clearly. The most important part for me is the interface for students so that they can use the kind of decimal character they are used to. (Though computer savvy students have usually learned that if , doesn't work, try . )

If I make an expression with the domain samples in 2,3; 0,8; 9,7 form, it seems to work, but if I use and expression with a decimal comma in it, that doesn't work.

This expression works:

{5:SA:%90%(=)3.8*100/(20*pi*5^2/4)@,1;0,2;0,33#That's a good expression for figuring it out, or an all too precise answer considering the significant digits given. 0.96766205399872 would be what, if rounded off to 2 significant digits? Try again!~%100%0,97#Good answer! (Also properly rounded off!)~%90%0,96#Pretty good answer and it has the right number of significant digits. But it isn't rounded off properly. Try again from 0,96766205399872~*#We haven't recognized your answer as correct. Remember: density = mass/volume cylinder volume = pi*d^2/4*h Try again!}

but this doesn't:

(=)3,8*100/(20*pi*5^2/4)@,1;0,2;0,33

I actually think it is easier on a Swedish keyboard to write @.1,0.2,0.33 than @,1;0,2;0,33
But more important for me is the treatment of the student input answers. I would like 3,8*100 to be interpreted as 3.8*100
This is pretty straight forward, if no one is using . or , as thousand separators. (I didn't see any 2 variable functions supported that would need commas.)

I would like pi() to be interpreted as pi. When testing it seemed to work sometimes and sometimes not. I think I even got an implied multiplication (without *) when using pi(). Maybe that's implemented for parentheses. Not sure what would happen in a sum like pi()+2. I haven't had time to check that. It's bed time here.

Another small wish: ** also as power (like ^)

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Wednesday, 18 January 2012, 7:57 AM

Jeff - the very top of a fine evening to ya...

W/re 3,8 ... neither a problem with your expressing yourself clearly, nor a problem with myself understanding what you needed/desired/wanted... but what is best called "programmer error", ie, I put the one line of code to do such in the wrong place. It has been corrected - my apologies.

W/re pi being a function of no arguments - it should be already, though I'd rather not advertise such... but out of curiosity why would you want such? Maybe I'm missing something here.

w/re implicit multiplication: it should be handled most/all places, ie, 3x is 3*x, x ln(x) is x * ln(x), 3pi/2 is 3*pi/2, 2(sin(x)+cos(x)) is 2*( etc etc... if you find situations where this is not working, please let me know.

w/re ** = ^ ... per chance are you an old Fortran writer? But let me look into such. And again, out of curiosity: why?

enjoy, and many thanks for the feedback and time/effort in testing,

greg

ps. I'll be uploading and debugging the multivariant code tonight... along with that will come the fix for 3,8 ... let me get the mv code stable, then give it a try and see if 3,8 doesn't work as expected.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Cristina Berisso - Wednesday, 18 January 2012, 3:18 PM

The algebra evaluator for "multiple variables" seem to be working.

I just tested a few cases and we will sincerely appreciate extra tester help (classes start this week and I will be loaded with classroom responsibilities especially during the first weeks of the semester and with very little time for intensive testing of the code). Just let me know if you feel you have some spare time to play with it, and I’ll get you in the test site.

The syntax for a single variable “x” has not changed from our original posting to the forum, but this addition allows the instructor to use several variables in the algebraic expression (and not necessarily “x”). The algebraic expression is again entered after the sequence (=) , then the “@” symbol but now, instead of just the evaluation points, the instructor defines the variables being used in an n-tuple format [for example (u,v,w)], followed by n-tuples with the respective evaluation points [for example (1.3,0.8,2.1) requesting the evaluation to be performed for u = 1.3, v = 0.8, and w = 2.1]. The instructors may enter as many n-tuples of evaluation points as he/she deems necessary. A few screen captures with actual examples are shown below:

multiple variable

Student's response:

response

Another example:

several variables

This new addition to the code also allows the instructor to use a single variable different from “x”, as in the example below, when the variable “t” is defined:

one variable different from x

Student's response:

student's response

Jeff’s “decimal comma” request (which may be of value also to other European and Latin American instructors) has been re-implemented. It also needs testers. Greg’s next goal will be more into the esthetics of the question’s lay-out: to study the code upstream in order to find a way to control the size of the answer box which can become quite large when one has a large number of variables and requests several points of evaluation for each. Tell you the truth, I am quite happy even with the long boxes for the multiple variables, this is like paradise for me after struggling for years to have something a little more adequate and attractive for algebraic answers than multiple choice. But I do understand that others may find the box’s length important.

Thank you so much for all the feedback, reports, suggestions, etc. that volunteers have provided. As always, the larger the number of testers the better.

Have a wonderful day!

Cristina and Greg

Average of ratings: Useful (1)

Re: algebra evaluator added to cloze type question

by Jeff Forssell - Friday, 20 January 2012, 11:49 PM

Hi Christina

Finally started testing the multivariable.

I seem to have had a misunderstanding of the implicit multiplication. When I try the PV=nRT and try the same answer you have given ( PV/(nR) I missed the spaces!) I get incorrect. I have to use P*V/(n*R). I also tried p*V/(n*R) (since some books I've used have pressure as small p) And that was accepted. Do we want the question to always be case insensitive? If not we should have some way of choosing case sensitivty?

I tried making a variant expression (changing (=)P*V/(n*R) to (=)PV/(nR) but then it doesn't even accept P*V/(n*R) but if I use spaces (=)P V/(n R) it works for P V/(n R) P*V/(n*R) and even P* V/(n*R)

in the "mult04" by "wabbit berisso" (=you?) I tried all these:

s(t) = 1/2*3.5*t^2+(-8)+4*t 1/2*3.5*t^2-8+4t 1/2*3.5*t^2-8+4*t .5*3.5*t^2-8+4*t
but got incorrect on all. Then I checked the expression in the question and it says {10:SA:=(=)-8+4*t+3.5*t^2@(t)(.6)(-.4)(1.7)} in spite of on the page saying a(t)=3.5. In your text above it says a(t)=7 so there seems to have been some changes made on the way to writing the question, that weren't coherent. If I assume a(t)=7 I get correct for 3.5*t^2+(-8)+4*t 3.5*t^2-8+4*t 1/2*7*t^2+(-8)+4*t but not 1/2*7*t^2+-8)+4*t and that's not a way of writing that should be encouraged.

When we talk about "writing your answer like in the calculator" I wonder if we should have the option for RAD DEG in some way. In Excel they have a RAD() funktion for getting the mathematical angle from a technical degrees angle. Or perhaps one could have sing(x) = sin(x*pi/180), cosg() , tang() etc. I think it can be important in more technical courses. Or that we choose behind the scenes when we want angles to be able to be entered in trig functions (sin() etc) in degrees. (and warn students in the question text)

Not sure if I should be posting this here or Priv messaging.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Cristina Berisso - Saturday, 21 January 2012, 1:04 PM

Hi Jeff,

In the very first version of the code we did not have any implicit multiplication, and as it evolved we included a bunch of different options. Now with the multiple variable code, we restricted a little the implicit product to allow for the definition of variables using more than just one character (for example to allow for the typing of initial velocity as "vi" or "v0", final velocity as "vf", and things of the sort).

Let me insert here a summary of the basic conventions the code currently uses:

Variables can be defined consisting of one or more characters, but the first character of a variable must be a letter.
All letters are internally converted to capitals by the code, so students will NOT be penalized if for example they typed an “X” instead of an “x” in their response.
Variables separated from other variables (or functions or special characters like “pi” or “e”) by a space are interpreted as the multiplication of such variables. [ Example: a*b*pi = a b pi ] That is not true with numbers: Multiplication symbols * or parentheses as delimiters of factors must be used to express products of numbers.
The second character in a variable can be a number, therefore although “2x” is interpreted by the code as 2*x (implicit multiplication of a number times a variable), “x2” is not. Implicit multiplication is recognized for numerical factors if the number is typed immediately BEFORE a variable (or function, or special character like “pi” or “e”)
Every time the user wants to type a numerical factor AFTER an algebraic expression or a grouping parentheses in order to perform a multiplication, the number must be typed surrounded by parentheses or a * (multiplication symbol) should be typed between the expression and the numerical factor.

Some expressions that would be interpreted as equivalent to: $$12\pi\,a\,b$$ (twelve times pi times a times b) are shown below. Please notice the use of spaces in between letters:

12*pi*a*b = 12 pi a b = 12pi a b = 12pi(a)(b) = 12pi (a b) = pi(a*b)*12 = pi a b(12) = pi a b*2*6 = 2pi a b(6) = 3pi a b*4 = pi (12a) b = pi a (12b) = pi a*12b = pi*12a b

The absolute safest is to request students to type multiplication symbol * every time they want to express a product.

Implicit multiplication between a number and a variable (or function or special characters as “pi” and “e”) is interpreted when the number precedes the variable, the function, or the special character. This allows for easy typing of expressions like:

-4x^3+7x^2-9x+2

3x^4+6x^3 y-17x^2 y^2-4x y^3+10y^4

5sin(x+2pi)-3cos(4y)

8ln(3x)-6arccos(0.4y)

With regards to the question with 3.5*t^2+(-8)+4*t , I had been playing with it with different values of the acceleration to shorten the length of the answer box, and most likely forgot to change the question title... I did try the "+-" combination you mention, and did not have any problem, the code interpreted it correctly. Could the problem of your answer not been accepted as correct be due to the fact that you have a closing parentheses that have no open counterpart? [you typed 1/2*7*t^2+-8)+4*t with a ")" after the 8]

The comment on allowing degrees apart from the default radians for trig functions and using a different function name sounds very appealing. I asked Greg about the possibility and he said that there will be no problem to include sind, cosd, and tand as you mention. Are those enough or you are also thinking about their inverses?

Please keep on testing the code, we do appreciate the feedback since this is the right time to get it polished...

Have fun!

Cristina

Average of ratings: Useful (1)

Re: algebra evaluator added to cloze type question

by Colin Fraser - Saturday, 21 January 2012, 9:53 AM

Hi Christina and Greg, I think I get it now, and this is way beyond what my kids will need or be able to do. For the most part, a simple Cloze or MC or straightforward calculation is probably more appropriate to my need than this. Please, I am not devaluing what you have done, it is fantastic, but only beyond my need. Do you intend on updating it to Moodle 2.x?

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Pilar Garcia Souto - Sunday, 26 July 2015, 7:33 PM

Dear Cristina and Greg,

It would be great to have this multi-variable algebraic option added to the close type questions. I am now, like Cristina, trying to develop some questions for our maths module and I face the same problem.

I notice that the last posts date from 2012. How did the testing go? Would you say it is ready to be incorporate it? (no that it is in my power, but it would be good to see this done).

Thanks for all your work on this! Great stuff.

Best regards,

Pilar

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Pilar Garcia Souto - Monday, 3 August 2015, 11:08 PM

Dear Cristina and Greg,

My institution is keen to test your plugin in our Moodle (great news) - I have the documentation and plugin that you provided here www.gmdr.com/MoodleStuff/AlgebraicEquivalentAnswerMoodleUpdate.zip.

However the documentation says it was last tested with version 2.2, and we are now on 2.8 and soon to be moving to 2.9 or 3.0. Do you have a later version of your plugin or should we test this one?

I would be delighted to see this implemented, and it would be a pity if it did not work because of the Moodle version

Thanks for your work!

Pilar

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Jeff Forssell - Thursday, 19 January 2012, 1:02 AM

I've sent this once before, but it seems to have disappeared. So I've resurrected with Lazarus:

Morning to ya, Greg!

I've never worked in Fortran, but have seen ** used in several connections where I think the first was a very nice interactive question script, I found and used/modified, called tutorial that was written in PERL. I think ** is a nice symbol (power as a kind of repeated multiplication) and easily accessable on the num pad.

pi() is the standard in Excel and other spreadsheets.

One other little thing that I used in that tutorial.pl script was translate ½ to "+.5" which made expressions like 3½ and ½*12*3.2^2 able to work. Even though it is important to tell students that ½ isn't understood by most programs, it seems a shame to give right for (1/2)*m*v^2 and wrong for ½mv^2. Do you think you could make a translator for ½? (Though I fear there will be expressions that wouldn't "survive translation", I believe there is the possibility of false negatives for proper expressions otherwise.)

how would your implied multiplication handle 1/2mv^2, =(1/2)*mv^2 ?

how would your implied multiplication handle xln(x) = x*ln(x) or unrecogninzed?

I would like mixed fractions like 1 3/4 to be recognized (like in Excel) as (1+3/4)

As you may have noticed, I like that this question can be used for purely numerical answers (that evaluates expressions). I would like to have a chooseable tolerance and intervals. (Give me a finger and I'll take your arm!! ) Let's say:
(=)0.5*13*12^2@p2% (no variables, proportional tolerance2% 0.02 2E-2)
(=)0.5*13*12^2@a0.4 (absolut tolerance +/- 0.4)
limit1..limit2 is a format used in Moodle GIFT format for interval. I would like to use that and refine it with ( or ) meaning not = to boundary, Like
(=)2.1..4.3 accepts answers 2.1<=ans<=4.3
(=)2.1..4.3) accepts answers 2.1<=ans<4.3
(=)..4.3) accepts answers ans<4.3
(=)2.1.. accepts answers 2.1<=ans
I know that hardcore programmers will say that these boundary conditions are not absolutely guaranteed, but I used something similar in PERL and didn't have any real (=school?) world problems. (We aren't landing on Saturn!) I don't know if there would be any sense adding some kind of precision key to this kind of expression, if a boundary wasn't being included or if too much near a boundary was included like:
(=)2.1..4.3@p10**-9 accepts answers proportionally within 0.000 000 001 of 2.1

When you get around to Moodle 2.2 version, it would be good to see if CBM, Certainty Based Marking, could be connected to this.

I haven't had time yet to try your latest changes. I'm afraid it may not happen until this evening or tomorrow.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Jeff Forssell - Thursday, 19 January 2012, 1:50 AM

Now I've been trying to see how , decimal is working. It doesn't seem to work for me. If I have a decimal in the "right expression", nothing gets recognized.

If I have a functioning expression using . as decimal, it doesn't intrepret a student input answer with , decimal.

At the same time I tried some other expressions that did or didn't work.

These are based on the question with
{5:SA:%90%(=)380/(20*pi*5^2/4)@.1,.2,.3# (in case I've written wrong here)

3.8*10^2/(20*3.141593*(5/2)^2)   no
3.8*10^2/(20*3.14159265*(5/2)^2)   yes
3.8*10^2/(20*3.14159265358979*(5/2)^2)   yes
380/(20*3.1416*5^2/4)   no (Example of when I would like to control tolerance)
380/(20*pi5^2/4)   no (I don't know if I really want this to work, but wondering how implicit mult would be here)
380/(20*pi*5^2/4)   yes
380/(20*pi()*5^2/4)   no   This is important I think to fix
3.8*100/(20*pi*5^2/4)   yes
3,8*100/(20*pi*5^2/4)   no (dec comma)
3.8E2/(20*pi*5^2/4)   no A common number format that would be good to support including E+99 E+09 E99 E-89. Sometimes one sees e in stead of E, but that can lead to problems with e=2,7.... So not sure how I would vote on e99, though I think it would usually be unambiguous.
3.8*10^2/(20*pi*5^2/4)   yes
3.8*10^2/(20*pi*5^2/2^2)   yes
3.8*10^2/(20*pi*(5/2)^2)   yes

I haven't tried multivariable yet.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Cristina Berisso - Thursday, 19 January 2012, 4:25 AM

Hello Jeff,

This is in reply to the part of E+99 notation in the algebra evaluator. When Greg was working on it he asked me implicitly about it and my answer was: I don't want it, it creates real problems with my students because of the use of "e" as natural base. I told him explicitly that I wanted "e" and "exp" to exclusively represent such. I normally ask students to write scientific notation in the form "1.27 10^28".; just a space in between (or a multiplication "*" symbol), which is how they normally write it explicitly showing what the base (ten) is. I would not like that to change. The E+ notation starts interfering not only with the base e but messes up the plus or minus signs with the implicit basic operations, and I would like to have them well separated.

The other important thing is that since I am using this as a tool to creating questions where variables can be "x0, x1, x2, x3, etc." the number immediately following a letter should not be understood as an implicit product. With the code as it is right now, if you do include a space in between or use a "*", it will be understood as a product and that does not create conflict with the nice wide flexibility of using several characters to represent a variable. I really prefer to keep this feature as it is.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Jeff Forssell - Thursday, 19 January 2012, 5:02 AM

I can live with that! But I must say that "1.27 10^28" (without any multiplication sign, looks very unfinished to me).

I don't think the implicit mult space would have to make any problem for "1 3/4" to be interpreted like in Excel a "no go" too? (Not a BIG issue, but I think it would be nice to be able to accept that kind of answer.)

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Thursday, 19 January 2012, 6:47 AM

Jeff -

Just a quick interjection (still wearing 9/5 hat)... but 1.27 10^28 is not a supported syntax... implicit multiplication works for 3(... 3pi, 3ln, 3x etc but not 3 3.

W/re 1/4, 1/2, and 3/4 (as single characters) - doesn't it take special knowledge to type those? ie, outside the realm of nominal student use? You know, and I know they're Ascii hex BC, BD and BE, respectively... but is a student [that's what Moodle is all about, right??] going to use them? ... that's the focus I'd like to maintain: the student.

W/re 10e-28 ... again, my fear is that is a specialty expression, beyond being too familiar from the Fortran days of the late 1960's and early 1970's... maybe we all need to hear more input from the community at large... but at this point I'm disinclined to go down that road (though since this is only software, any/everything is possible).

Beyond that: my fear is that the fine-tuning details are becoming of less and less interest to most of the forum readers... as such, edicate suggests: any chance of taking this to email? (mine's in my profile and you have Cristina's)

Though one point I would like to hear from the users at large: when does one call these mods a done deal and make them generally available? When no changes have been made for a while? Or, are users willing to entertain changes and have to re-install them? Since I'm the keeper for Cristina's sites, I have no problem keeping them up to date... but for the rest: what's the best solution/answer (that is of course presuming that anyone would want to make use of them)?

many thanks - greg

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Sunday, 29 January 2012, 11:47 AM

Jeff - the top of the evening to ya...

Just a heads up: the degree'd version of the trig functions have been added: sind, cosd, tand, secd, cscd and cotd all accept/expect degrees as the arguments; likewise, arcsind, arccosd, arctand, arcsecd, arccscd and arccotd will return degree valued function results. These changes are for both the 1.9 and 2.2 testsites. Many thanks for suggesting such.

the best to you - greg

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Cristina Berisso - Monday, 23 January 2012, 9:42 AM

Hi Jeff,

In revisiting your decimal comma request after a attending an excellent Tchaikovsky V Symphony concert (good to relax, detoxify and charge the brain), we realized that you have been testing the code in our server, which has a North American locale setting. We believe that for a real testing of that part of the code, it should be installed in a server where PHP is set to interpret commas as decimals as you probably have in your Moodle website in Sweden. Could you wait for one to two weeks until we have the Moddle 2.2 work done, to test the comma decimal functionality there under appropriate conditions?

Cristina

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Saturday, 28 January 2012, 12:26 PM

Just the briefest of posts: as of about an hour ago, the EA (equivalent answer) / Algebra Evaluator, is up and running under 2.2.1+ ... Cristina will post more w/re details, and how to try/test it.

[though just an aside comment: the code in 2.2 is sooo much cleaner and more professional than anything in 1.9 - my hat's off to the core developers and their efforts and accomplishments - very very well done people, very well done! ]

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Saturday, 28 January 2012, 12:29 PM

A quick follow up: as in 1.9, EA adds 5 lines of code to an existing .php file, and adds one new .php file, ie, the difficulty of integrating for most should not be a problem. [and in reply to another post - don't see how EA could be implemented as a block, unfortunately.]

Average of ratings: -

Re: algebra evaluator added to cloze type question

by John Rodgers - Thursday, 19 January 2012, 11:20 AM

This is a much needed addition to the question type, and seems to be a "more simple" integration than other systems. Good work. Please take it over the goal line.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Cristina Berisso - Saturday, 28 January 2012, 3:12 PM

The algebra evaluator module has been installed in a moodle 2.2+ site for testing. As far as I can tell it is performing correctly, it recognizes equivalent algebraic expressions including multiple variable expressions. One detail that may be relevant for the users is to notice that in Moodle 2.2 the cloze questions can be set to show the “correct answer” to the student along with the feedback as he/she places the cursor over the answer box. Please be aware that even if you set the option “right answer” to SHOW while editing the quiz, the correct answer will not be shown for our algebra evaluation type questions; only a * symbol will appear after the words: “correct answer:” when the student places the cursor over the answer box with the submitted response.

I have tested a very limited number of expressions on this new setting (not much of the trig functions yet) and would appreciate volunteers that can fly a moodle 2.2 with ease. I must confess with extreme embarrassment that I felt very lost in the moodle 2.2 environment after many years of driving version 1. I don’t want to make public in this forum how long it took me to simply add a “teacher” to the test course… So please, if you are a moodle 2.2 experienced pilot, feel free to test the algebra evaluator in our moodle 2.2 testing site:

http://www.bigfatturkey.com/moodle2

and let us know at your earliest convenience if you find any incongruence, expression that is not interpreted, function that is not evaluated correctly, etc.

The visitor can enter in “teacher” roll and create questions and quizzes. We included links to examples, syntax conventions, etc. for both: single “x” variable, and multiple variable. A sample quiz with questions extracted from those created by testers in our moodle 1.9.15 site is also available. The username and password to use are shown within the course description for the only available course of this site.

Thank you again for volunteering!

Cristina

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Saturday, 28 January 2012, 5:30 PM

"Please be aware that even if you set the option “right answer” to SHOW while editing the quiz, the correct answer will not be shown for our algebra evaluation type questions; only a * symbol will appear after the words: “correct answer:” when the student places the cursor over the answer box with the submitted response." ....

Not wishing to leave the impression that this implementation is a done-deal... I'll be working toward getting the correct answers to appear, as expected, ie, more code to read to understand how the question engine in ver 2 works... as with all mods to Moodle: one step at a time ;-(

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Cristina Berisso - Sunday, 12 February 2012, 5:05 PM

Dear Moodlers,

As our confidence grows with the success of the current algebra evaluator users/testers, we would like to make the code available to the Moodle community at large. Greg just finished putting together all the necessary files for installing it, including his detailed instructions according to the version of Moodle you are running. They are in the attached zip file.

The module allows for the usage of a single variable “x” and of multiple user defined variables as described in the postings above. We believe this addition to the cloze question type could serve many math and science instructors and their students.

Instructions on the usage of the code and the functions it supports can be found via:

http://www.collegelevel.org/tempo/equivanswerFull.pdf

May it serve you well!

Cristina and Greg

P.S. Since this module requires modification of the core Moodle code, we’ve been told to file a new issue as an "improvement" in the tracker, and wait for feedback from the Quiz component leader. We’ll be posting the appropriate links when this is done.

equivanswer.zip

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Tim Hunt - Monday, 13 February 2012, 3:19 AM

This is a very cool bit of work. I hope people find it useful and give you good feedback.

I am, however, sure it does not belong in the standard Moodle release yet. The cloze question type is already a big complicated mess, and adding this would just make matters worse, however nice the functionality is. People have been saying for more than 5 years that the cloze qtype should be more plug-in-able, and they are right, and your code just adds weight to that case (MDL-6371). However, it still does not make it easy.

As it happens, I am working with Chris Sangwin on STACK at the moment. STACK questions are built up from separate components a bit like Cloze questions are. Once I have learned how STACK implements its questions, and more importantly, how it presents an interface to teachers to let them author questions in this flexible way, perhaps I will then be able to steal some of the ideas for Moodle.

In the mean time, there is a good place to share your code. In the plugins database, there is a category for changes that are not self-contained plugins. I suggest you upload your code there for now. I am sorry that there is not a better answer than that at the moment.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Monday, 13 February 2012, 7:59 AM

Just as a follow up: the .zip which Cristina uploaded, for size reasons, does not include a copy of the .pdf documentation, but merely a link to it. I have uploaded a complete .zip to www.gmdr.com/MoodleStuff/AlgebraicEquivalentAnswerMoodleUpdate.zip for those that would like to download a complete distribution including a copy of the documentation.

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Joseph Rézeau - Monday, 13 February 2012, 6:34 PM

Hi Greg,

Have you considered using a github repository for your algebra question type? See http://docs.moodle.org/22/en/Git_for_Administrators.

Joseph

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Tuesday, 14 February 2012, 2:03 AM

Joseph, sir,

Had not considered it, but will look into it.

greg

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Tuesday, 14 February 2012, 2:08 AM

Joseph, sir,

BTW in following the link to the git page, in the 2nd paragraph there is this minor typo:

To get the most of of Git it is worth... => To get the most out of Git it is worth...

greg

Average of ratings: -

Re: algebra evaluator added to cloze type question

by khaled baloul - Monday, 13 February 2012, 7:50 PM

Hello Greg and all,

After installation . A SA question preview.

I got the following error:

Help please!

I am using moodle 2.2+ version.

khaled

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Tuesday, 14 February 2012, 2:01 AM

Khaled -

My apologies - we hadn't/haven't seen such here (otherwise we would have fixed it before releasing it)... all our testing and actually production has been on LAMP servers. What are you running on? Beyond that, I looked at the code pointed to... generally in PHP references to variables which have not been defined are defined to be null or false... in this case there was a reference to the $debug variable which in your case was undefined (because you had not called AE in debug mode which is correct, ie, not a mistake on your part)... anyway, I put an explicit $debug = false ahead of the reference in question. This is a stab in the dark, but I'm hoping it'll be a fix. I updated the .zip on gmdr.com to reflect this change. If you could try that updated version and get back to us, I would be much appreciative. Cristina will update the zip attached to this forum thread this evening (she's currently teaching a physics class).

again, my apologies,

greg

Average of ratings: -

Re: algebra evaluator added to cloze type question

by Cristina Berisso - Wednesday, 15 February 2012, 12:20 AM

Hello all,

The code with the fix for what Greg perceives of Khaled problem (we haven't heard back from him yet) is included in the updated zip on Greg's site:

www.gmdr.com/MoodleStuff/AlgebraicEquivalentAnswerMoodleUpdate.zip

I created also an entry in the Tracker as suggested by Martin D. so we now have a place where to add info and updates . Here is the link:

http://tracker.moodle.org/browse/MDL-31601

Thank you Tim, Martin, and all for your guidance on how to make the code available without having to clutter the forum and without tight size constraints!

Cristina

Average of ratings: -

Re: algebra evaluator added to cloze type question

by khaled baloul - Wednesday, 15 February 2012, 5:17 AM

Hello Christina, Hi Greg and All,

Thanks a lot for your quick answer and sorry for my late answer!

I tried the new code and now i got this error:

I will do more testing soon, and will come with new feebacks

Thanks again

Regards

Khaled

Average of ratings: -

Re: algebra evaluator added to cloze type question

by khaled baloul - Wednesday, 15 February 2012, 5:21 AM

Hi Greg,

For information:

Using LAMP server provider and last moodle version for testing!

Regards

khaled

Average of ratings: -

Re: algebra evaluator added to cloze type question

by greg mushial - Wednesday, 15 February 2012, 7:14 AM

Khaled -

Again, my apologies... but this one might be harder to fix in that it's a blowup in the cloze code and not in EA... I'm willing to take a look at it, but it might take more time to resolve in that I know nothing of that code

sorry this is giving you problems,

greg

Average of ratings: -

Re: algebra evaluator added to cloze type question

by khaled baloul - Wednesday, 15 February 2012, 4:47 PM

Greg,

Thanks anyway, I Will be doing more tests later on.

I seems that Stack is available only for 1.9 plateformes. May be had u any experience with it to share with us.

Regards

Khaled

Average of ratings: -

Re: algebra evaluator added to cloze type question

by khaled baloul - Wednesday, 15 February 2012, 5:12 PM

Hello Cristina,

It seems like nothing has changed after unchecking 3 "Right Answers" (one was already unchecked) from the settings!

Khaled

Average of ratings: -