openwebwork-devel — WeBWorK development discussion

Re: [WWdevel] SetMaker is slow!

[Sam H. <sh...@ma...>]

On Jul 28, 2006, at 12:16 AM, Sam Hathaway wrote:

> Does it bother anyone else how slow SetMaker is? It's not that  
> module's fault, it just takes a long time to render problems, and  
> rendering 20 of them gets pretty insane.
>
> I was thinking that a solution to this would be to cache the entire  
> output of the problem processor. For non-interactive situations  
> like SetMaker, the only thing we'd need to cache on is the problem  
> seed -- other data (like num_correct, num_incorrect, etc.) are  
> always the same.

This isn't the whole story. There are a lot of things that can  
invalidate the cache:

- changing the pg source
- changing an auxiliary file
- changing a macro file
- changing a pg module (don't worry about that though)
- changing certain values in global.conf or course.conf that are  
propagated through to the problem environment

For equation image caching, we ignore everything but the first one.  
It is possible to end up with stale images, but it hasn't been a big  
deal since the math munging code is pretty static and most people  
aren't making changes that would disrupt that. If the whole  
translation process is being cached, many more things could change  
the output, and it would be pretty confusing if SetMaker.pm was  
showing older output than Problem.pm.

> The cache could simply be a file containing the problem TEXT, named  
> based on the source file name and the problem seed, or we could get  
> a little more complex and also store the names of any temporary  
> files that the problem depends on. That way, deleting, say, a  
> generated graph would invalidate the cache and cause the problem to  
> be re-rendered.
>
> This would speed up problem browsing TREMENDOUSLY, since PG  
> wouldn't even run once ANYONE had browsed a problem set. This might  
> even be a good candidate for pre-caching -- run through the problem  
> library once with seed 1234 to save a lot of instructor time down  
> the road.

[WWdevel] moved default courses directory to courses.dist

[Sam H. <sh...@ma...>]

Hi,

I moved the webwork2/courses directory to webwork2/courses.dist in  
the repository.

Mike and Arnie have decided to treat the default courses directory as  
a template, and have users copy it to /opt/webwork/courses on  
installation.

When you cvs update, you'll see an error like this one:

cvs update: Updating courses
cvs update: cannot open directory /webwork/cvs/system/webwork2/ 
courses: No such file or directory
cvs update: skipping directory courses

To fix it, remove the line

D/courses////

from your webwork2/CVS/Entries file.

Make sure you pass -d to cvs update to get the new courses.dist  
directory.

If you've been using the webwork2/courses directory for real courses  
(I think most of us are not), you should make a copy of the new  
courses.dist => ../courses and move your existing courses from the  
OLD webwork2/courses directory to that directory. You can then delete  
the old webwork2/courses directory.

Sorry for the inconvenience.
-sam

Re: [WWdevel] Possible replacement for EV3

[Michael G. <ga...@ma...>]

Hi Davide,

This stuff all looks very nice.

One thing we can do is to redefine BEGIN_TEXT/END_TEXT so that it  
uses EV3P with the default parameters.
There is no reason not to allow the new formula inputs as well as  
long as things remain backward compatible with EV3.

We can make the other versions of EV3P available in the webwork core  
as well for cases where folks want
to fine tune the evaluation abilities. (Just as EV2 is still  
available if you want to interpolate $ variables before
command execution instead of afterwards.)

I'd like to tune up translator a bit in any case and structure it so  
that there is a plugin preprocessor and post processor
surrounding the main rendering.  Some of this may already be in place  
-- I don't have the code in front of me -- but I know
that I never completed all of the plans.  The preprocessor handles  
things like  BEGIN_TEXT/END_TEXT  and could handle
BEGIN_HINT  BEGIN_SOLUTION as well.  I've been resisting putting too  
much stuff in these processors but I think these
two would be ok -- particularly if we optimize the grep search.   A  
post processor would handle the work of Latex2HTML (RIP)
but it could also do things such as color answer blanks red or green  
depending on whether the answer is right -- this would
be useful for matrices and problems with large numbers of answer  
blanks.  At least some of these hooks are already in
place but it would be nice to finish the job.

Take care,

Mike

On Aug 27, 2006, at 11:57 AM, Davide P. Cervone wrote:

> Folks:
>
> While working on my essay-answer tools, I ended up making a modified
> version of EV3 that I think might be more generally useful.  My
> original intent was to include the ability to enter mathematics not
> in TeX form but in the answer-checker-style text strings, so that
> something like sin(x/(x+1)) would produce \sin\!\left(\frac{x}{x+1}
> \right).  There are times when that would be a more convenient form.
> So I made a copy of EV3 called EV3P and added a feature like \(...\)
> and \[...\] where a text string could be passed to the Parser, and
> the Parser would produce the TeX form of the result.  I used `...`
> for in-line math and ``...`` for display math (and if either is used
> within math mode already, it uses whatever that mode is).  So you can
> say
>
> 	Suppose \(f(x) = `sin(x/(x+1))`\) is a function of `x`
>
> for example.
>
> As I was using this, I found that I wanted to be able to control
> which substitutions were being performed (for example, when writing
> an answer box that allowed students to preview an essay answer that
> includes mathematics, I didn't want them to be able to do command
> substition via \{...\}).  Rather than have yet another copy of EV3, I
> decided to allow my new EV3P to have parameters where you could
> control which substitutions are done.  If the first argument to EV3P
> is a HASH reference, this is taken to be the list of options.  These
> include processCommands, processVariables, processMath and
> processParser.  Using EV3P({processCommands=>0},$string) would not do
> command substitution, for example.  Setting processParser=>0 would
> make EV3P functionally equivalent to EV3.
>
> The code for EVP3 is in my answerEssay.pl file (see http://
> cvs.webwork.rochester.edu/viewcvs.cgi/union_problib/examples/
> answerEssay/?cvsroot=Union+College).  I think EV3P could be used to
> replace EV3 to allow another simple means of entering mathematics.
> (I'm also thinking of defining a parser Context in which more
> mathematical notation is available, like limits, sums, integrals, and
> so on, to make this more useful.)  What do you think?  You can try
> essay5.pg in the directory linked above to experiment with the
> results of EV3P.
>
> Davide
>
>
> ---------------------------------------------------------------------- 
> ---
> Using Tomcat but need to do more? Need to support web services,  
> security?
> Get stuff done quickly with pre-integrated technology to make your  
> job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache  
> Geronimo
> http://sel.as-us.falkag.net/sel? 
> cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> OpenWeBWorK-Devel mailing list
> Ope...@li...
> https://lists.sf.net/lists/listinfo/openwebwork-devel

[WWdevel] Filtering of characters in student answers

[Davide P.C. <dp...@un...>]

Folks:

Most of the routines that build answer boxes filter the student  
answers so that the text they produce can be passed through EV3.   
This is because the result of something like ans_rule(10) includes  
the previous answer, when available, as part of its text, and if the  
ans_rule() occurs within BEGIN_TEXT/END_TEXT, then that text will be  
passed to EV3() for further processing.  This means that if the  
student answer included something like \(...\), then it would be  
processed as mathematics, producing garbage in the answer blank for  
the student.

The current solution is to remove characters like \, @, $ and ` from  
the student's answer to prevent further processing, but I think there  
is a better solution.  If these characters were changed to their HTML  
unicode references (\, @, $, and `), then they  
would pass through EV3() unchanged, but would still appear correctly  
in the browser.  This would allow the full preservation of student  
answers.  (I needed something like this for the essay-answer-with- 
preview feature that I was working on, and it does seem to work.)

Comments?

Davide

[WWdevel] Possible replacement for EV3

[Davide P. C. <dp...@un...>]

Folks:

While working on my essay-answer tools, I ended up making a modified  
version of EV3 that I think might be more generally useful.  My  
original intent was to include the ability to enter mathematics not  
in TeX form but in the answer-checker-style text strings, so that  
something like sin(x/(x+1)) would produce \sin\!\left(\frac{x}{x+1} 
\right).  There are times when that would be a more convenient form.   
So I made a copy of EV3 called EV3P and added a feature like \(...\)  
and \[...\] where a text string could be passed to the Parser, and  
the Parser would produce the TeX form of the result.  I used `...`  
for in-line math and ``...`` for display math (and if either is used  
within math mode already, it uses whatever that mode is).  So you can  
say

	Suppose \(f(x) = `sin(x/(x+1))`\) is a function of `x`

for example.

As I was using this, I found that I wanted to be able to control  
which substitutions were being performed (for example, when writing  
an answer box that allowed students to preview an essay answer that  
includes mathematics, I didn't want them to be able to do command  
substition via \{...\}).  Rather than have yet another copy of EV3, I  
decided to allow my new EV3P to have parameters where you could  
control which substitutions are done.  If the first argument to EV3P  
is a HASH reference, this is taken to be the list of options.  These  
include processCommands, processVariables, processMath and  
processParser.  Using EV3P({processCommands=>0},$string) would not do  
command substitution, for example.  Setting processParser=>0 would  
make EV3P functionally equivalent to EV3.

The code for EVP3 is in my answerEssay.pl file (see http:// 
cvs.webwork.rochester.edu/viewcvs.cgi/union_problib/examples/ 
answerEssay/?cvsroot=Union+College).  I think EV3P could be used to  
replace EV3 to allow another simple means of entering mathematics.   
(I'm also thinking of defining a parser Context in which more  
mathematical notation is available, like limits, sums, integrals, and  
so on, to make this more useful.)  What do you think?  You can try  
essay5.pg in the directory linked above to experiment with the  
results of EV3P.

Davide

Re: [WWdevel] Duplicate problem libraries in CVS

[John J. <jj...@as...>]

Hi,

I think the end result should be to close other libraries where we can, 
and to provide a file saying where to find problems which are not in the 
same location.

I think the diff situation is much worse than it seemed.  I tried diff 
-r on current cvs checkouts, and they gave tons of results.  I think 
this is a difference in how diff operates on different types of 
systems.  Every problem in database_problems has library tags in it, so 
it should be different from the version under rochester_problib.

The reason some files/directories are not under database_problems is 
that we were trying to minimize duplication.  In retrospect, it might 
have been best to start with the Rochester library, but we didn't.  Then 
as problems were added, we tried to remove as many duplicates as we 
could before adding the problems.

The current policy on duplicate problems under database_problems is to 
avoid them unless the same problem has rightfully been tagged into two 
different places.  If a problem happens to fit to spots in the tagging 
heierarchy, then we keep both copies both because that's the most 
natural solution given how tagging and the current database setup works, 
and because future changes to the problem might make sense for its role 
in one place in the heierarchy but not in another.  I should say that 
having duplicates like this is currently only theoretical - I don't know 
of any problem which is correctly tagged to 2 different spots of the 
heierarchy.

For missing problems, I count a little over 800 by running wc on the 
captured errors from

  for j in `find . -name \*.pg` ; do diff -b $j 
../database_problems/Rochester/$j | grep -v \# >> Diffout ; done  >& Differr

I have tried finding problems, and it isn't very hard.  The bigger part 
of the problem is that you have to repeat the process 800 times.

There are also differences inside files beyond the tags.  These are 
captured by the command above in Diffout (although without file paths).  
Looking through that file, the first 5 files where differences were 
found came from corrections only done in the library versions.

I had e-mailed Jeff about the possibility of having some of the student 
labor work on the 800 "moved" problems.  Maybe that can help get this 
fixed up reasonably soon.

John


Arnold Pizer wrote:

>Hi,
>
>Currently we have duplicates of most libraries, e.g.
>devel ...lock/cvs/database_problems/Rochester/
>and
>devel ...lock/cvs/rochester_problib/
>
>I think this will soon lead to difficulties (e.g. when fixing buggy
>problems and adding new ones).  One solution would be to only use the
>libraries under database_problems and to close off the other libraries
>in such a way that one gets a message that they have bben moved if one
>tries to access them.  This would involve updating the CVS
>documentation. E.g. basically one would have to download
>database_problems to get any of the libraries but that is probably
>what most people will do anyway.  It would also mean that places using
>e.g. rochester_problib will have to change the path to this library in
>global.conf but that's a pretty easy change.
>
>Here are a few questions for John.  I did a diff -r Rochester
>../rochester_problib/ (results below) and at this point everything is
>identical except for one missing problem
>(../rochester_problib/setTrig05Graphs: c5s3p1_9) and a number of
>missing directories.  If we go with my solution above, we should add
>all the missing directories. John, is there a reason you didn't copy
>these? I haven't looked in email or macros but the others are things
>that we should keep.  Did you copy everything from the other
>libraries?
>
>I think we should make a decision on how to proceed with this soon
>before these libraries start diverging (e.g. I want to fix a few
>Rochester problems but will hold off until this is settled).
>
>Any ideas or comments?
>
>Here's the result of the diff:
>
>devel ...lock/cvs/database_problems diff -r Rochester ../rochester_problib/
>Only in ../rochester_problib/: email
>Only in ../rochester_problib/: macros
>Only in ../rochester_problib/: set0
>Only in ../rochester_problib/: setAppletExamples
>Only in ../rochester_problib/: setDiffEQold2Separable
>Only in ../rochester_problib/: setDiffEQold3Mixing
>Only in ../rochester_problib/: setDiffEQold4Growth
>Only in ../rochester_problib/: setDiffEQold5Logistic
>Only in ../rochester_problib/: setFunctions0Evaluate
>Only in ../rochester_problib/: setFunctions1FunctionsGraphs
>Only in ../rochester_problib/: setFunctions2Composition
>Only in ../rochester_problib/: setFunctions3Transforms
>Only in ../rochester_problib/: setFunctions4Inverse
>Only in ../rochester_problib/: setGenericQuestions
>Only in ../rochester_problib/: setGeometry1Points
>Only in ../rochester_problib/: setGeometry2Lines
>Only in ../rochester_problib/: setGeometry3Conics
>Only in ../rochester_problib/: setIntegrationProjects
>Only in ../rochester_problib/: setLogExp1
>Only in ../rochester_problib/: setLogExp2Applications
>Only in ../rochester_problib/: setMAAtutorial
>Only in ../rochester_problib/: setNewQuestions
>Only in ../rochester_problib/: setSampleAnswers
>Only in ../rochester_problib/: setSampleGraders
>Only in ../rochester_problib/: setSampleGraphs
>Only in ../rochester_problib/: setSampleQuestionnaires
>Only in ../rochester_problib/: setSequentialProblems
>Only in ../rochester_problib/: setStatistics9ProcessControl
>Only in ../rochester_problib/setTrig05Graphs: c5s3p1_9
>Only in ../rochester_problib/: setTrigonometry1
>Only in ../rochester_problib/: setTrigonometry2Waves
>Only in ../rochester_problib/: setTrigonometry3WordProblems
>Only in ../rochester_problib/: setTrigonometry4Inverse
>Only in ../rochester_problib/: setTrigonometry5Hyperbolic
>devel ...lock/cvs/database_problems
>
>Arnie
>
>-------------------------------------------------------------------------
>Using Tomcat but need to do more? Need to support web services, security?
>Get stuff done quickly with pre-integrated technology to make your job easier
>Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
>http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
>_______________________________________________
>OpenWeBWorK-Devel mailing list
>Ope...@li...
>https://lists.sf.net/lists/listinfo/openwebwork-devel
>  
>

Re: [WWdevel] location of Union macros

[Davide P. C. <dp...@un...>]

> Currently if one wants to use the Union problems, one has to link in
> the macros.  E.g. the instructions in
>
> http://devel.webwork.rochester.edu/twiki/bin/view/Webwork/ 
> InstallationManualV2pt2forSuSE10pt1
>
> read
>
>  Search for macrosPath and scroll down several lines to the line
> $pg{directories}{macros},
>  After this line add the two lines:
> '/opt/webwork/library/union_problib/macros',
> '/opt/webwork/library/union_problib/parserOrientation',
>
> Is there any reason not to put these under macros in the standard  
> distribution?
>

Yes.  Many of the Union macro files are pre-Parser and are no linger  
needed, and I'm not planning to continue supporting them.  Some are  
pretty early work, and need to be redone.  I do not think they are  
appropriate for the core distribution.  The things that I am ready  
for permanent installation are already added to macros.

Now that the macros path can be easily modified, there is much less  
reason to include these in pg/macros than before.  I would not want  
to see that directory cluttered up with macro files that I'm not  
fully satisfied with.


> I just looked and these  (and some examples)  are not included under
> database_problems/Union
>
> devel ...lock/cvs/database_problems diff -r Union/ ../union_problib/
> Only in ../union_problib/examples: AnswerFile-pl
> Only in ../union_problib/examples: AnswerFile-pm
> Only in ../union_problib/examples: answerEssay
> Only in ../union_problib/examples: compoundProblem
> Only in ../union_problib/examples: multiProblem
> Only in ../union_problib/: macros
> Only in ../union_problib/: parserOrientation
> Only in ../union_problib/: setOrientation
>

The examples files are not actual problems for use, but simply  
samples of programming techniques.  They illustrate the use of some  
macro packages for doing some complicated and non-standard things  
(like problems with several parts, or writing answers to files).   
They are relatively new and I'm not sure that the packages are stable  
yet.  They are experimental and I don't want them to be thought of as  
standard parts of WeBWorK yet.

Davide

[WWdevel] location of Union macros

[Arnold P. <arn...@gm...>]

Hi,

Currently if one wants to use the Union problems, one has to link in
the macros.  E.g. the instructions in

http://devel.webwork.rochester.edu/twiki/bin/view/Webwork/InstallationManualV2pt2forSuSE10pt1

read

 Search for macrosPath and scroll down several lines to the line
$pg{directories}{macros},
 After this line add the two lines:
'/opt/webwork/library/union_problib/macros',
'/opt/webwork/library/union_problib/parserOrientation',

Is there any reason not to put these under macros in the standard distribution?

I just looked and these  (and some examples)  are not included under
database_problems/Union

devel ...lock/cvs/database_problems diff -r Union/ ../union_problib/
Only in ../union_problib/examples: AnswerFile-pl
Only in ../union_problib/examples: AnswerFile-pm
Only in ../union_problib/examples: answerEssay
Only in ../union_problib/examples: compoundProblem
Only in ../union_problib/examples: multiProblem
Only in ../union_problib/: macros
Only in ../union_problib/: parserOrientation
Only in ../union_problib/: setOrientation


Arnie

[WWdevel] Duplicate problem libraries in CVS

[Arnold P. <arn...@gm...>]

Hi,

Currently we have duplicates of most libraries, e.g.
devel ...lock/cvs/database_problems/Rochester/
and
devel ...lock/cvs/rochester_problib/

I think this will soon lead to difficulties (e.g. when fixing buggy
problems and adding new ones).  One solution would be to only use the
libraries under database_problems and to close off the other libraries
in such a way that one gets a message that they have bben moved if one
tries to access them.  This would involve updating the CVS
documentation. E.g. basically one would have to download
database_problems to get any of the libraries but that is probably
what most people will do anyway.  It would also mean that places using
e.g. rochester_problib will have to change the path to this library in
global.conf but that's a pretty easy change.

Here are a few questions for John.  I did a diff -r Rochester
../rochester_problib/ (results below) and at this point everything is
identical except for one missing problem
(../rochester_problib/setTrig05Graphs: c5s3p1_9) and a number of
missing directories.  If we go with my solution above, we should add
all the missing directories. John, is there a reason you didn't copy
these? I haven't looked in email or macros but the others are things
that we should keep.  Did you copy everything from the other
libraries?

I think we should make a decision on how to proceed with this soon
before these libraries start diverging (e.g. I want to fix a few
Rochester problems but will hold off until this is settled).

Any ideas or comments?

Here's the result of the diff:

devel ...lock/cvs/database_problems diff -r Rochester ../rochester_problib/
Only in ../rochester_problib/: email
Only in ../rochester_problib/: macros
Only in ../rochester_problib/: set0
Only in ../rochester_problib/: setAppletExamples
Only in ../rochester_problib/: setDiffEQold2Separable
Only in ../rochester_problib/: setDiffEQold3Mixing
Only in ../rochester_problib/: setDiffEQold4Growth
Only in ../rochester_problib/: setDiffEQold5Logistic
Only in ../rochester_problib/: setFunctions0Evaluate
Only in ../rochester_problib/: setFunctions1FunctionsGraphs
Only in ../rochester_problib/: setFunctions2Composition
Only in ../rochester_problib/: setFunctions3Transforms
Only in ../rochester_problib/: setFunctions4Inverse
Only in ../rochester_problib/: setGenericQuestions
Only in ../rochester_problib/: setGeometry1Points
Only in ../rochester_problib/: setGeometry2Lines
Only in ../rochester_problib/: setGeometry3Conics
Only in ../rochester_problib/: setIntegrationProjects
Only in ../rochester_problib/: setLogExp1
Only in ../rochester_problib/: setLogExp2Applications
Only in ../rochester_problib/: setMAAtutorial
Only in ../rochester_problib/: setNewQuestions
Only in ../rochester_problib/: setSampleAnswers
Only in ../rochester_problib/: setSampleGraders
Only in ../rochester_problib/: setSampleGraphs
Only in ../rochester_problib/: setSampleQuestionnaires
Only in ../rochester_problib/: setSequentialProblems
Only in ../rochester_problib/: setStatistics9ProcessControl
Only in ../rochester_problib/setTrig05Graphs: c5s3p1_9
Only in ../rochester_problib/: setTrigonometry1
Only in ../rochester_problib/: setTrigonometry2Waves
Only in ../rochester_problib/: setTrigonometry3WordProblems
Only in ../rochester_problib/: setTrigonometry4Inverse
Only in ../rochester_problib/: setTrigonometry5Hyperbolic
devel ...lock/cvs/database_problems

Arnie

Re: [WWdevel] LaTeXMathML mode

[Arnold P. <arn...@gm...>]

Hi,

Please hold off on removing the latex2html related code.  People at
Utah have written over a 1000 problems and we are told many of them
have TeX code outside of math mode.  This means that they can only be
viewed in latex2html mode which is why Utah usually uses this mode.
Peter Alfred has just given us the Utah problems and hopefully we can
get an intern to clean up the ones with offending TeX code.  It's
possible other problem writers have done the same.

I think what we should do is to just by default comment out the almost
never used modes.

Arnie

On 8/23/06, John Jones <jj...@as...> wrote:
>
>  Hi,
>
>  It seems reasonable to keep it just as Mike suggests - present in case an
> expert really wants it, but disabled by default.
>
>  I would think latex2html related code could be cleaned out.
>
>  John
>
>
>
>  Michael Gage wrote:
>  In the beginning I used plain text a lot for debugging, but I confess
> i haven't needed
> to do that in years.
>
> Could one keep it around for debugging purposes only by leaving it in
> the list of display modes
> but disabling it by default? or do people think we should get rid of
> it all together?
>
> -- Mike
>
>
> On Aug 23, 2006, at 3:48 PM, Davide P. Cervone wrote:
>
>
>
>
>  I use tth occasionally in development, or if I'm getting students who
> complain that they're having trouble with jsMath or image modes.
>
>  I've used it, too, but not very often. What I DON'T use is "plain
> text", which I have disabled here at Union. Most problems are not
> written to work with this mode, and so I don't want students trying
> to use it. It is often awkward to make a problem that works with
> this mode and still looks good in the other modes, and it requires
> using macros pretty much everywhere (e.g., use $GE rather than \ge),
> which no one does.
>
>
>
>  I've never used ASCIIMathML, and so see no reason to eliminate it.
>
>  I hope he means "no reason NOT to eliminate it." I'd be fine
> removing ASCIIMathML as well.
>
> Has anyone tried this with MSIE? I can't do that, and I'm concerned
> about it, because there is extra setup that it needs, and I'm not
> sure it is being done properly.
>
> Davide
>
>
>
> ----------------------------------------------------------------------
> ---
> Using Tomcat but need to do more? Need to support web services,
> security?
> Get stuff done quickly with pre-integrated technology to make your
> job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache
> Geronimo
> http://sel.as-us.falkag.net/sel?
> cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> OpenWeBWorK-Devel mailing list
> Ope...@li...
> https://lists.sf.net/lists/listinfo/openwebwork-devel
>
>
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job
> easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> OpenWeBWorK-Devel mailing list
> Ope...@li...
> https://lists.sf.net/lists/listinfo/openwebwork-devel
>
>
>
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job
> easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
>
> _______________________________________________
> OpenWeBWorK-Devel mailing list
> Ope...@li...
> https://lists.sf.net/lists/listinfo/openwebwork-devel
>
>
>

[WWdevel] rel-2-3-dev branch created

[Sam H. <sh...@ma...>]

Hi,

Last week I created a pre-release branch for WeBWorK 2.3. It's called  
rel-2-3-dev. It will eventually be released as WeBWorK 2.3.0 and then  
continue accumulating fixes for subsequent 2.3.x releases.

This code is very similar to HEAD (perhaps identical) at the moment,  
but it will diverge as we start playing with experimental code in  
HEAD. It should be fairly stable at this point.

Please don't commit stuff to this directly, but if you think a commit  
should be included in this branch make a not of it in your commit  
message or drop me an email. I've been monitoring the cvs log emails  
and including bugfix-type commits.

If you'd like to help test, please update your systems to the branch:

cd /path/to/webwork2
cvs up -dP -r rel-2-3-dev
cd ../pg
cvs up -dP -r rel-2-3-dev

If you've been running HEAD, the update should be painless. If you're  
running rel-2-2-dev or a 2.2.x release, you'll need to apply updates  
to global.conf, database.conf, and webwork.apache-config. Another  
thing to remember is to run wwdb_check and wwdb_upgrade after you  
install 2.3 for the first time.

Draft release notes:
<http://devel.webwork.rochester.edu/twiki/bin/view/Webwork/ 
WeBWorKRelease2pt2pt2>
<http://devel.webwork.rochester.edu/twiki/bin/view/Webwork/ 
PGLanguageRelease2pt2pt2>

Draft installation manual:
<http://devel.webwork.rochester.edu/twiki/bin/view/Webwork/ 
InstallationManualV2pt3>

Enjoy.
-sam

Re: [WWdevel] LaTeXMathML mode

[John J. <jj...@as...>]

Hi,

It seems reasonable to keep it just as Mike suggests - present in case 
an expert really wants it, but disabled by default.

I would think latex2html related code could be cleaned out.

John


Michael Gage wrote:

>In the beginning I used plain text a lot for debugging, but I confess  
>i haven't needed
>to do that in years.
>
>Could one keep it around for debugging purposes only by leaving it in  
>the list of display modes
>but disabling it by default?  or do people think we should get rid of  
>it all together?
>
>-- Mike
>
>
>On Aug 23, 2006, at 3:48 PM, Davide P. Cervone wrote:
>
>  
>
>>>I use tth occasionally in development, or if I'm getting students who
>>>complain that they're having trouble with jsMath or image modes.
>>>      
>>>
>>I've used it, too, but not very often.  What I DON'T use is "plain
>>text", which I have disabled here at Union.  Most problems are not
>>written to work with this mode, and so I don't want students trying
>>to use it.  It is often awkward to make a problem that works with
>>this mode and still looks good in the other modes, and it requires
>>using macros pretty much everywhere (e.g., use $GE rather than \ge),
>>which no one does.
>>
>>    
>>
>>>I've never used ASCIIMathML, and so see no reason to eliminate it.
>>>      
>>>
>>I hope he means "no reason NOT to eliminate it."  I'd be fine
>>removing ASCIIMathML as well.
>>
>>Has anyone tried this with MSIE?  I can't do that, and I'm concerned
>>about it, because there is extra setup that it needs, and I'm not
>>sure it is being done properly.
>>
>>Davide
>>
>>
>>
>>---------------------------------------------------------------------- 
>>---
>>Using Tomcat but need to do more? Need to support web services,  
>>security?
>>Get stuff done quickly with pre-integrated technology to make your  
>>job easier
>>Download IBM WebSphere Application Server v.1.0.1 based on Apache  
>>Geronimo
>>http://sel.as-us.falkag.net/sel? 
>>cmd=lnk&kid=120709&bid=263057&dat=121642
>>_______________________________________________
>>OpenWeBWorK-Devel mailing list
>>Ope...@li...
>>https://lists.sf.net/lists/listinfo/openwebwork-devel
>>
>>    
>>
>
>
>-------------------------------------------------------------------------
>Using Tomcat but need to do more? Need to support web services, security?
>Get stuff done quickly with pre-integrated technology to make your job easier
>Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
>http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
>_______________________________________________
>OpenWeBWorK-Devel mailing list
>Ope...@li...
>https://lists.sf.net/lists/listinfo/openwebwork-devel
>  
>

Re: [WWdevel] LaTeXMathML mode

[Davide P. C. <dp...@un...>]

>> I don't have MSIE with the MathPlayer plugin,
>> so I can't test it there.  Could someone else check that?
>
> I just tested this and it didn't work at all:

Crap!  Thanks for testing it, though.  I'll check into it here and  
see what I can figure out.

Davide

Re: [WWdevel] LaTeXMathML mode

[Michael G. <ga...@ma...>]

In the beginning I used plain text a lot for debugging, but I confess  
i haven't needed
to do that in years.

Could one keep it around for debugging purposes only by leaving it in  
the list of display modes
but disabling it by default?  or do people think we should get rid of  
it all together?

-- Mike


On Aug 23, 2006, at 3:48 PM, Davide P. Cervone wrote:

>> I use tth occasionally in development, or if I'm getting students who
>> complain that they're having trouble with jsMath or image modes.
>
> I've used it, too, but not very often.  What I DON'T use is "plain
> text", which I have disabled here at Union.  Most problems are not
> written to work with this mode, and so I don't want students trying
> to use it.  It is often awkward to make a problem that works with
> this mode and still looks good in the other modes, and it requires
> using macros pretty much everywhere (e.g., use $GE rather than \ge),
> which no one does.
>
>> I've never used ASCIIMathML, and so see no reason to eliminate it.
>
> I hope he means "no reason NOT to eliminate it."  I'd be fine
> removing ASCIIMathML as well.
>
> Has anyone tried this with MSIE?  I can't do that, and I'm concerned
> about it, because there is extra setup that it needs, and I'm not
> sure it is being done properly.
>
> Davide
>
>
>
> ---------------------------------------------------------------------- 
> ---
> Using Tomcat but need to do more? Need to support web services,  
> security?
> Get stuff done quickly with pre-integrated technology to make your  
> job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache  
> Geronimo
> http://sel.as-us.falkag.net/sel? 
> cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> OpenWeBWorK-Devel mailing list
> Ope...@li...
> https://lists.sf.net/lists/listinfo/openwebwork-devel
>

Re: [WWdevel] LaTeXMathML mode

[Davide P. C. <dp...@un...>]

>> Peter Jipsen announcing the modification to ASCIIMathML.  This works
>> a lot better for our needs than the original ASCIIMathML (though I
>> didn't remove that mode), as it handles much more of the LaTeX
>> commands and syntax.  I don't have MSIE with the MathPlayer plugin,
>> so I can't test it there.  Could someone else check that?
>
> Thanks for the contribution. The output looks kind of dodgy with  
> Camino on my Mac, but I haven't looked into it. Specifically,  
> digits are overlapping and some symbols seem to be missing. This is  
> with the math fonts from the Mozilla site installed.

The MathML support on the Mac has never been very satisfactory.  I've  
never really gotten it to work well enough to use, so I don't have  
much experience with it.  My understanding is that the output is not  
up to the quality of TeX (but what is), so it will not look as good  
to us as either images or jsMath, but it has some advantages (like  
being able to copy and past the mathematics into things like  
Mathematica, and for the visually impaired, it can be read aloud by  
the computer (with appropriate software).  That might come in handy  
for students with disabilities.

Davide

Re: [WWdevel] LaTeXMathML mode

[Davide P. C. <dp...@un...>]

> I use tth occasionally in development, or if I'm getting students who
> complain that they're having trouble with jsMath or image modes.

I've used it, too, but not very often.  What I DON'T use is "plain  
text", which I have disabled here at Union.  Most problems are not  
written to work with this mode, and so I don't want students trying  
to use it.  It is often awkward to make a problem that works with  
this mode and still looks good in the other modes, and it requires  
using macros pretty much everywhere (e.g., use $GE rather than \ge),  
which no one does.

> I've never used ASCIIMathML, and so see no reason to eliminate it.

I hope he means "no reason NOT to eliminate it."  I'd be fine  
removing ASCIIMathML as well.

Has anyone tried this with MSIE?  I can't do that, and I'm concerned  
about it, because there is extra setup that it needs, and I'm not  
sure it is being done properly.

Davide

Re: [WWdevel] LaTeXMathML mode

[P. G. L. <gl...@um...>]

Hi Sam,

I use tth occasionally in development, or if I'm getting students who 
complain that they're having trouble with jsMath or image modes.  I'd 
leave that in, because it gives a distinctly different way of rendering 
mathematics.  I've never used ASCIIMathML, and so see no reason to 
eliminate it.

Gavin

-- 
P Gavin LaRose, PhD  |  gl...@um... | 734.764.6454  |  ...you have
Program Manager, Instructional Technology |   to respect someone who can
Mathematics Dept,  University of Michigan | spell  Tuesday, even if they
http://www.math.lsa.umich.edu/~glarose/   | can't spell it right. -Milne

On Wed, 23 Aug 2006, Sam Hathaway wrote:

> Thanks for the contribution. The output looks kind of dodgy with
> Camino on my Mac, but I haven't looked into it. Specifically, digits
> are overlapping and some symbols seem to be missing. This is with the
> math fonts from the Mozilla site installed.
>
> Is there any reason not to remove ASCIIMathML? It never worked well
> in my experience. Speaking of which, is anyone still using TtH?
> -sam

Re: [WWdevel] LaTeXMathML mode

[Sam H. <sh...@ma...>]

On Aug 17, 2006, at 8:08 PM, Davide P. Cervone wrote:

> I have added a LaTeXMathML mode based on the message we received from
> Peter Jipsen announcing the modification to ASCIIMathML.  This works
> a lot better for our needs than the original ASCIIMathML (though I
> didn't remove that mode), as it handles much more of the LaTeX
> commands and syntax.  I don't have MSIE with the MathPlayer plugin,
> so I can't test it there.  Could someone else check that?

Thanks for the contribution. The output looks kind of dodgy with  
Camino on my Mac, but I haven't looked into it. Specifically, digits  
are overlapping and some symbols seem to be missing. This is with the  
math fonts from the Mozilla site installed.

Is there any reason not to remove ASCIIMathML? It never worked well  
in my experience. Speaking of which, is anyone still using TtH?
-sam

Re: [WWdevel] Sequential problems and hardcopy

[P. G. L. <gl...@um...>]

Hi Davide,

It should be easy enough to make this apply to regular problem sets.  I 
have it set up so that it's concealed in a conditional---if the set is 
versioned and correct answers are asked for, it grabs the problems for the 
student and inserts the answers.  That's probably not what we'd want for 
regular sets, though.  I wanted it there so that I can have students print 
out tests with their solutions on them to go over them with a tutor, but 
for regular homeworks it seems it should either automatically put in any 
answers that have been submitted, or have a new checkbox to allow it.

Gavin

-- 
P Gavin LaRose, PhD  |  gl...@um... | 734.764.6454  |  ...you have
Program Manager, Instructional Technology |   to respect someone who can
Mathematics Dept,  University of Michigan | spell  Tuesday, even if they
http://www.math.lsa.umich.edu/~glarose/   | can't spell it right. -Milne

On Mon, 21 Aug 2006, Davide P. Cervone wrote:

>> Re: Hardcopy generation and access to answers, this summer I put a hack 
>> into Hardcopy.pm to make it grab student answers for versioned (gateway) 
>> sets so that I can output those on a graded set.  Without having extensive 
>> testing to determine the net impact of this on the penalty thereby 
>> incurred, it seems that this at least works.  I suspect that this is not an 
>> answer to your question, but hopefully it's at least a step away from a non 
>> sequitor.
>> 
>> I'll add that I didn't bother to try and insert the student's answer in the 
>> answer blanks, instead simply listing it in the same manner the Correct 
>> Answer is included.
>
> It sounds like you already have done basically what is needed, then.  Would 
> what you have done be able to work with standard homework sets as well?
>
> Davide
>
>
>

Re: [WWdevel] Sequential problems and hardcopy

[Davide P. C. <dp...@un...>]

> Re: Hardcopy generation and access to answers, this summer I put a  
> hack into Hardcopy.pm to make it grab student answers for versioned  
> (gateway) sets so that I can output those on a graded set.  Without  
> having extensive testing to determine the net impact of this on the  
> penalty thereby incurred, it seems that this at least works.  I  
> suspect that this is not an answer to your question, but hopefully  
> it's at least a step away from a non sequitor.
>
> I'll add that I didn't bother to try and insert the student's  
> answer in the answer blanks, instead simply listing it in the same  
> manner the Correct Answer is included.

It sounds like you already have done basically what is needed, then.   
Would what you have done be able to work with standard homework sets  
as well?

Davide

Re: [WWdevel] Hardcopy from Library Browser and Problem Editor

[Davide P. C. <dp...@un...>]

>> Speaking of hardcopy issues, often when I am editing a problem, I'd
>> like to be able to test its hardcopy output.  To do this, I have to
>> make a test homework set, assign the problem to that, and then get
>> hardcopy of that set.  It would be very convenient to be able to get
>> hardcopy directly from the problem editor.  How hard would that be?
>
> I had once looked at this for the Library Browser.  There was nothing
> insurmountable to it.  The first step looked like it would be to  
> extract
> the guts of the hardcopy code from Hardcopy.pm to a utility so that it
> could be called from a variety of modules.

Sounds like a reasonable approach.  Anyone interested in doing it.  :-)

I don't think I'm going to have time for a while, I'm afraid.

Davide

Re: [WWdevel] Sequential problems and hardcopy

[P. G. L. <gl...@um...>]

Hi Davide,

Re: Hardcopy generation and access to answers, this summer I put a hack 
into Hardcopy.pm to make it grab student answers for versioned (gateway) 
sets so that I can output those on a graded set.  Without having extensive 
testing to determine the net impact of this on the penalty thereby 
incurred, it seems that this at least works.  I suspect that this is not 
an answer to your question, but hopefully it's at least a step away from a 
non sequitor.

I'll add that I didn't bother to try and insert the student's answer in 
the answer blanks, instead simply listing it in the same manner the 
Correct Answer is included.

Gavin

-- 
P Gavin LaRose, PhD  |  gl...@um... | 734.764.6454  |  ...you have
Program Manager, Instructional Technology |   to respect someone who can
Mathematics Dept,  University of Michigan | spell  Tuesday, even if they
http://www.math.lsa.umich.edu/~glarose/   | can't spell it right. -Milne

On Mon, 21 Aug 2006, Davide P. Cervone wrote:

> Folks:
>
> There has been some questions on the discussion board recently about
> multi-part or sequential problems, and I have been thinking about
> some issues that came up when I worked on a solution for one of these.
>
> First, the text that is displayed for sequential problems is
> dependent on the answers that the student has given so far, but since
> the hardcopy processing does not have access to those answers, the
> hardcopy will never show the later parts of the problem, on the first
> part.  How hard would it be to pass the answers to the hardcopy
> processor so that an accurate version of the problem can be
> generated?  (I know that there will be lookup time, and since the
> hardcopy can show many problems, how significant would that be?)
>
> One interesting possibility that this would create is that the
> hardcopy could (optionally) include the answers the student has
> typed, so that she could get a permanent record of her solutions, and
> they could even be printed within the problem, on the answer blanks
> (though that might cause formatting problems, I suppose).
>
> Second, when a sequential problem goes on to the second part, it
> would be nice to be able to have the option of preventing the student
> from altering the answers to the first part.  This could cause the
> second part to disappear again, for example, or if the second part
> depends on the values of the first part, you might not want the
> student to go back and change the previous answers to make the later
> part easier.
>
> This could be done by disabling the answer blanks for the first part,
> so they would still show up (with the student answer showing), but
> not be editable.  You could have something similar to Mike's
> STOP_PROCESSING()/RESUME_PROCESS() commands, say
> BEGIN_DISABLED_ANSWERS()/END_DISABLED_ANSWERS() or something like
> that.  (In addition to disabling them, you would need to include a
> hidden field so that the data is still passed to the form on the next
> submission.)
>
> Comments?
>
> Davide
>
> -------------------------------------------------------------------------
> Using Tomcat but need to do more? Need to support web services, security?
> Get stuff done quickly with pre-integrated technology to make your job easier
> Download IBM WebSphere Application Server v.1.0.1 based on Apache Geronimo
> http://sel.as-us.falkag.net/sel?cmd=lnk&kid=120709&bid=263057&dat=121642
> _______________________________________________
> OpenWeBWorK-Devel mailing list
> Ope...@li...
> https://lists.sf.net/lists/listinfo/openwebwork-devel
>
>
>

Re: [WWdevel] Hardcopy from Library Browser and Problem Editor

[John J. <jj...@as...>]

Davide P. Cervone wrote:

>Folks:
>
>Speaking of hardcopy issues, often when I am editing a problem, I'd  
>like to be able to test its hardcopy output.  To do this, I have to  
>make a test homework set, assign the problem to that, and then get  
>hardcopy of that set.  It would be very convenient to be able to get  
>hardcopy directly from the problem editor.  How hard would that be?
>  
>
I had once looked at this for the Library Browser.  There was nothing 
insurmountable to it.  The first step looked like it would be to extract 
the guts of the hardcopy code from Hardcopy.pm to a utility so that it 
could be called from a variety of modules.

>Of course, having just one problem in the (two-column) output usually  
>makes a bad break in the middle of the problem, so it would probably  
>be good to output TWO copies of the problem, so you get two side-by- 
>side unbroken problems for testing purposes.
>  
>
I imagine that it would not be too hard to add a flag to the hardcopy 
function to show two copies of the problem.  Maybe it could 
automatically pick a different seed for the second copy.

John

[WWdevel] Hardcopy from Library Browser and Problem Editor

[Davide P. C. <dp...@un...>]

Folks:

Speaking of hardcopy issues, often when I am editing a problem, I'd  
like to be able to test its hardcopy output.  To do this, I have to  
make a test homework set, assign the problem to that, and then get  
hardcopy of that set.  It would be very convenient to be able to get  
hardcopy directly from the problem editor.  How hard would that be?

Of course, having just one problem in the (two-column) output usually  
makes a bad break in the middle of the problem, so it would probably  
be good to output TWO copies of the problem, so you get two side-by- 
side unbroken problems for testing purposes.

Also, it would be really nice to be able to get hardcopy directly  
from the Library Browser. This would make it easy to print reference  
pages for the various libraries, for example.  If the problems to be  
printed were selected by checkbox, that would be even better.

What do you think?

Davide

[WWdevel] Sequential problems and hardcopy

[Davide P. C. <dp...@un...>]

Folks:

There has been some questions on the discussion board recently about  
multi-part or sequential problems, and I have been thinking about  
some issues that came up when I worked on a solution for one of these.

First, the text that is displayed for sequential problems is  
dependent on the answers that the student has given so far, but since  
the hardcopy processing does not have access to those answers, the  
hardcopy will never show the later parts of the problem, on the first  
part.  How hard would it be to pass the answers to the hardcopy  
processor so that an accurate version of the problem can be  
generated?  (I know that there will be lookup time, and since the  
hardcopy can show many problems, how significant would that be?)

One interesting possibility that this would create is that the  
hardcopy could (optionally) include the answers the student has  
typed, so that she could get a permanent record of her solutions, and  
they could even be printed within the problem, on the answer blanks  
(though that might cause formatting problems, I suppose).

Second, when a sequential problem goes on to the second part, it  
would be nice to be able to have the option of preventing the student  
from altering the answers to the first part.  This could cause the  
second part to disappear again, for example, or if the second part  
depends on the values of the first part, you might not want the  
student to go back and change the previous answers to make the later  
part easier.

This could be done by disabling the answer blanks for the first part,  
so they would still show up (with the student answer showing), but  
not be editable.  You could have something similar to Mike's  
STOP_PROCESSING()/RESUME_PROCESS() commands, say  
BEGIN_DISABLED_ANSWERS()/END_DISABLED_ANSWERS() or something like  
that.  (In addition to disabling them, you would need to include a  
hidden field so that the data is still passed to the form on the next  
submission.)

Comments?

Davide