aimmath-developers — Development related messages

[Aimmath-developers] [ aimmath-Bugs-798640 ] Sporadic Validation error: bad name x

[SourceForge.net <no...@so...>]

Bugs item #798640, was opened at 2003-09-01 16:14
Message generated for change (Comment added) made by npstrick
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798640&group_id=44411

Category: Marking
Group: None
Status: Open
Resolution: None
Priority: 8
Submitted By: Ken Monks (monks)
Assigned to: Nobody/Anonymous (nobody)
Summary: Sporadic Validation error: bad name x

Initial Comment:
When Trying certain questions or quizzes, AiM 

reports "Validation error: bad name x." where x can be 

any single letter variable name that occurs in the answer 

that is provided by AiM (which is the correct answer).  

The error is not repeatable.  When a question in a Tried 

quiz produces the error, Trying that same question 

separately with the same answer will not reproduce the 

error.  I have only noticed this with AiM supplied 

answers when Trying a quiz or question, not when 

entering answers by hand as a registered student (but I 

haven't tested that as frequently, so it is possible it 

happens there also).

----------------------------------------------------------------------

>Comment By: Neil Strickland (npstrick)
Date: 2003-09-01 21:56

Message:
Logged In: YES 
user_id=430681

This error means that when `aim/SafeParse` is called to parse

the student's answer, the global variable x has somehow 

acquired a value.  This is not supposed to happen, even if

the relevant question file assigns a value to x, because of the

way that the functions `Local/Save` and `Local/Restore` are

used.  However, it is possible that this mechanism is faulty.



If you are seeing this error, please let me know whether the

relevant variable occurs in the local variable list for your

question, and whether it gets assigned a value.



Please also replace the line



 report['BadName'] := nam;



in SafeParse.mpl by the line 

   

 report['BadName'] := sprintf("%s has value %A",nam,eval

(convert(nam,name)));



This modification is not safe against hacking, so it should only

be inserted temporarily.  However, it will at least report what

value has been assigned to the bad name, which should help

to pinpoint where the assignment occurred.

----------------------------------------------------------------------

Comment By: Gustav W Delius (gustav_delius)
Date: 2003-09-01 19:56

Message:
Logged In: YES 
user_id=737246

This bug is hitting me too. The problem is reproducible as long 

as the quiz is not changed. However deleting other unrelated 

questions of the same quiz can make the problem go away.



The bug shows up when trying a whole quiz as administrator. 

For me it is still there if I click on "Try separately". However if 

I then press "New Version" or "Mark" on the single-question 

page the error goes away. Pressing "Mark" or "New Version" 

on the page with the whole quiz does not make the error go 

away.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798640&group_id=44411

[Aimmath-developers] [ aimmath-Bugs-798640 ] Sporadic Validation error: bad name x

[SourceForge.net <no...@so...>]

Bugs item #798640, was opened at 2003-09-01 16:14
Message generated for change (Comment added) made by gustav_delius
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798640&group_id=44411

Category: Marking
Group: None
Status: Open
Resolution: None
Priority: 8
Submitted By: Ken Monks (monks)
Assigned to: Nobody/Anonymous (nobody)
Summary: Sporadic Validation error: bad name x

Initial Comment:
When Trying certain questions or quizzes, AiM 

reports "Validation error: bad name x." where x can be 

any single letter variable name that occurs in the answer 

that is provided by AiM (which is the correct answer).  

The error is not repeatable.  When a question in a Tried 

quiz produces the error, Trying that same question 

separately with the same answer will not reproduce the 

error.  I have only noticed this with AiM supplied 

answers when Trying a quiz or question, not when 

entering answers by hand as a registered student (but I 

haven't tested that as frequently, so it is possible it 

happens there also).

----------------------------------------------------------------------

>Comment By: Gustav W Delius (gustav_delius)
Date: 2003-09-01 19:56

Message:
Logged In: YES 
user_id=737246

This bug is hitting me too. The problem is reproducible as long 

as the quiz is not changed. However deleting other unrelated 

questions of the same quiz can make the problem go away.



The bug shows up when trying a whole quiz as administrator. 

For me it is still there if I click on "Try separately". However if 

I then press "New Version" or "Mark" on the single-question 

page the error goes away. Pressing "Mark" or "New Version" 

on the page with the whole quiz does not make the error go 

away.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798640&group_id=44411

[Aimmath-developers] [ aimmath-Bugs-798637 ] Another error in Class/IndexFunction

[SourceForge.net <no...@so...>]

Bugs item #798637, was opened at 2003-09-01 16:01
Message generated for change (Comment added) made by npstrick
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798637&group_id=44411

Category: None
Group: None
Status: Open
Resolution: None
Priority: 8
Submitted By: Ken Monks (monks)
Assigned to: Nobody/Anonymous (nobody)
Summary: Another error in Class/IndexFunction

Initial Comment:
The following error message:



"Error, (in Class/IndexFunction) Type error in 

assignment: field InertFunction in class InertExpr should 

have type Inert"



That occurs quite frequently when trying quizzes, 

though I can't reproduce the exact conditions to make it 

repeatable.  If I change the 3 lines labeled "WAS:" in

Class/IndexFunction in Class.mpl back to what they were 

the error seems to go away. I don't really understand 

the Class/IndexFunction proc or why those changes 

were made in the first place so I don't want to just 

change it back since someone obviously changed them 

for a reason.



Is this related to Gustav's Class/IndexFunction error 

[796266 sporadic prompt field error]?

----------------------------------------------------------------------

>Comment By: Neil Strickland (npstrick)
Date: 2003-09-01 17:27

Message:
Logged In: YES 
user_id=430681

It was me that made these changes, in response to bug

number 796262 ("c> type error").  Previously, if x was an

object created by the OOP framework, then an

assignment like x['Foo'] := bar; would fully evaluate bar.

In particular, questionversion['MapleType'] := vector; 

would set questionversion['MapleType'] to the vector()

procedure, whereas you really need it to be set to the

unevaluated name 'vector'.



Anyway, the probable explanation of Ken's bug is that

there is an assignment x['InertFunction'] := bar, where 

bar evaluates to something of type Inert, but bar itself

does not have type inert.  This assignment should be

changed to x['InertFunction'] := eval(bar);  I hope that

will fix the problem.  I will let Ken look at this, as I have

not understood the Inert stuff yet.



----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798637&group_id=44411

Re: [Aimmath-developers] installlin.html

[Manolis M. <ma...@ma...>]

> Manolis wrote:
>
> > The problem is that substituting ROOT with AiM for instance .. will not
do
> > the work (at least it seems so, I am still trying and let you know in a
> > while). If you have an old AIM you probably need to keep your old
> > options as
> > they are preaty hardcoded in every .mpl file.
>
> Can you say that again? I don't understand what you mean.

Was it my typos :-D ? Neil can hopefully say more on that. Once I asked him
and said :

>Currently there is no good way to copy information from an old AIM setup to
a new clean
>installation; you just have to set up the subjects and quizzes again by
hand and copy the
>question files  across manually.

Neil was referring to copying files The same problem exists when you install
a new AIM and slightly change your options (such as the directories). The
question, quizzes .m files have paths and emails and passwords stored in
them and if you copy them along then AIM fails to find them ... That is if
you want to keep your questions you just have to use your previous settings
(I have a feeling that even your password needs to be the same) then as usal
run Make that will create index.html etc and then unhide the subjects that
will apper at the admin page.

Maybe I am wrong in some of these...


Manolis

[Aimmath-developers] [ aimmath-Bugs-798642 ] ln formats in italics

[SourceForge.net <no...@so...>]

Bugs item #798642, was opened at 2003-09-01 12:18
Message generated for change (Comment added) made by monks
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798642&group_id=44411

Category: Student Interface
Group: None
>Status: Closed
>Resolution: Fixed
Priority: 4
Submitted By: Ken Monks (monks)
Assigned to: Ken Monks (monks)
Summary: ln formats in italics

Initial Comment:
The natural log symbol ln, now formats in italics in Maple 

9 with the latest version of AiM.

----------------------------------------------------------------------

>Comment By: Ken Monks (monks)
Date: 2003-09-01 12:36

Message:
Logged In: YES 
user_id=766386

I restored the code for `latex/latex/symbol` that was 

recently removed from aim/LaTeX.mpl, and that fixed this 

problem.  If that causes formatting problems for anyone else, 

let me know.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798642&group_id=44411

[Aimmath-developers] [ aimmath-Bugs-798642 ] ln formats in italics

[SourceForge.net <no...@so...>]

Bugs item #798642, was opened at 2003-09-01 12:18
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798642&group_id=44411

Category: Student Interface
Group: None
Status: Open
Resolution: None
Priority: 4
Submitted By: Ken Monks (monks)
Assigned to: Ken Monks (monks)
Summary: ln formats in italics

Initial Comment:
The natural log symbol ln, now formats in italics in Maple 

9 with the latest version of AiM.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798642&group_id=44411

[Aimmath-developers] [ aimmath-Bugs-798640 ] Sporadic Validation error: bad name x

[SourceForge.net <no...@so...>]

Bugs item #798640, was opened at 2003-09-01 12:14
Message generated for change (Settings changed) made by monks
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798640&group_id=44411

Category: Marking
Group: None
Status: Open
Resolution: None
>Priority: 8
Submitted By: Ken Monks (monks)
Assigned to: Nobody/Anonymous (nobody)
Summary: Sporadic Validation error: bad name x

Initial Comment:
When Trying certain questions or quizzes, AiM 

reports "Validation error: bad name x." where x can be 

any single letter variable name that occurs in the answer 

that is provided by AiM (which is the correct answer).  

The error is not repeatable.  When a question in a Tried 

quiz produces the error, Trying that same question 

separately with the same answer will not reproduce the 

error.  I have only noticed this with AiM supplied 

answers when Trying a quiz or question, not when 

entering answers by hand as a registered student (but I 

haven't tested that as frequently, so it is possible it 

happens there also).

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798640&group_id=44411

[Aimmath-developers] [ aimmath-Bugs-798637 ] Another error in Class/IndexFunction

[SourceForge.net <no...@so...>]

Bugs item #798637, was opened at 2003-09-01 12:01
Message generated for change (Settings changed) made by monks
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798637&group_id=44411

Category: None
Group: None
Status: Open
Resolution: None
>Priority: 8
Submitted By: Ken Monks (monks)
Assigned to: Nobody/Anonymous (nobody)
Summary: Another error in Class/IndexFunction

Initial Comment:
The following error message:



"Error, (in Class/IndexFunction) Type error in 

assignment: field InertFunction in class InertExpr should 

have type Inert"



That occurs quite frequently when trying quizzes, 

though I can't reproduce the exact conditions to make it 

repeatable.  If I change the 3 lines labeled "WAS:" in

Class/IndexFunction in Class.mpl back to what they were 

the error seems to go away. I don't really understand 

the Class/IndexFunction proc or why those changes 

were made in the first place so I don't want to just 

change it back since someone obviously changed them 

for a reason.



Is this related to Gustav's Class/IndexFunction error 

[796266 sporadic prompt field error]?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798637&group_id=44411

[Aimmath-developers] [ aimmath-Bugs-798640 ] Sporadic Validation error: bad name x

[SourceForge.net <no...@so...>]

Bugs item #798640, was opened at 2003-09-01 12:14
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798640&group_id=44411

Category: Marking
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Ken Monks (monks)
Assigned to: Nobody/Anonymous (nobody)
Summary: Sporadic Validation error: bad name x

Initial Comment:
When Trying certain questions or quizzes, AiM 

reports "Validation error: bad name x." where x can be 

any single letter variable name that occurs in the answer 

that is provided by AiM (which is the correct answer).  

The error is not repeatable.  When a question in a Tried 

quiz produces the error, Trying that same question 

separately with the same answer will not reproduce the 

error.  I have only noticed this with AiM supplied 

answers when Trying a quiz or question, not when 

entering answers by hand as a registered student (but I 

haven't tested that as frequently, so it is possible it 

happens there also).

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798640&group_id=44411

[Aimmath-developers] [ aimmath-Bugs-798637 ] Another error in Class/IndexFunction

[SourceForge.net <no...@so...>]

Bugs item #798637, was opened at 2003-09-01 12:01
Message generated for change (Tracker Item Submitted) made by Item Submitter
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798637&group_id=44411

Category: None
Group: None
Status: Open
Resolution: None
Priority: 5
Submitted By: Ken Monks (monks)
Assigned to: Nobody/Anonymous (nobody)
Summary: Another error in Class/IndexFunction

Initial Comment:
The following error message:



"Error, (in Class/IndexFunction) Type error in 

assignment: field InertFunction in class InertExpr should 

have type Inert"



That occurs quite frequently when trying quizzes, 

though I can't reproduce the exact conditions to make it 

repeatable.  If I change the 3 lines labeled "WAS:" in

Class/IndexFunction in Class.mpl back to what they were 

the error seems to go away. I don't really understand 

the Class/IndexFunction proc or why those changes 

were made in the first place so I don't want to just 

change it back since someone obviously changed them 

for a reason.



Is this related to Gustav's Class/IndexFunction error 

[796266 sporadic prompt field error]?

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=798637&group_id=44411

[Aimmath-developers] [ aimmath-Bugs-797333 ] Guest user no longer able to access system

[SourceForge.net <no...@so...>]

Bugs item #797333, was opened at 2003-08-29 10:54
Message generated for change (Comment added) made by monks
You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=797333&group_id=44411

Category: Student Interface
Group: None
>Status: Closed
>Resolution: Fixed
Priority: 8
Submitted By: Chris Sangwin (sangwinc)
Assigned to: Ken Monks (monks)
Summary: Guest user no longer able to access system

Initial Comment:
It appears that guest users are no longer able to gain 

access to quizzes.  As far as my system seems to 

behave, any attempt to access the system by a guest 

user results in the error



Error, (in Class/IndexFunction) Type error in assignment: 

field InviteNewVersion in class `aim/Quiz/Context` 

should have type boolean 



Is this connected with Ken's modifications?



Chris



----------------------------------------------------------------------

>Comment By: Ken Monks (monks)
Date: 2003-09-01 11:50

Message:
Logged In: YES 
user_id=766386

Guests have access once again.

----------------------------------------------------------------------

You can respond by visiting: 
https://sourceforge.net/tracker/?func=detail&atid=439479&aid=797333&group_id=44411

Re: [Aimmath-developers] installlin.html

[Greg G. <gr...@ma...>]

On Sun, 31 Aug 2003, Manolis Mavrikis wrote:
> > I'm not sure how much of this we want to tell an installer of AiM.
> > I guess I'll write some of it and if it starts looking too long and
> > complicated, just suggest that he/she read the appropriate docs of
> > tomcat.
>
> yes .. maybe just what you wrote at the email before briefly or not at all.
>
> I 'll wait for your changes and after you commit I 'll comment

Sorry, for the delay ... I only (just now) committed my further changes.

  Regards,
  Greg

[Aimmath-developers] authoring tools and other mods

[Ken M. <mo...@pt...>]

I just committed all of the stuff I had been working on for the past couple
of weeks.  Most of it consists of question authoring tools.  Most if it is
also rather harmless, in that if you don't use it in writing your questions
it shouldn't break anything.  However, there are a few things that could
affect the LaTeX formatting of the mathematics in existing questions (though
I hope in most cases it will be for the better). If there is some change you
don't like, let me know and I'll do a feature-ectomy on that item.  See
below for details.

COMMITTED CHANGES: Sep 1, 2003
------------------------------

* Added the following files:

1. aim/Inert.mpl

This file contains a package called 'Inert'. (It has nothing to do with
ToInert and FromInert.)  It is an authoring tool that is designed to provide
inert forms of Maple expressions whose formatted appearance is independent
and separate from their mathematical value (somewhat analogous to the inert
operators Int or Diff in Maple). It is difficult to explain this in a short
email message, so I wrote up a quick explanation/introduction and
temporarily put it in the file AuthoringGuide.pdf in the docs directory.

This package is intended to be a first step towards solving the problem of
how to overcome the default simplification and formatting choices that are
made by Maple and AiM in order to give question authors the ability to
customize the notation used for a given expression according to their
lecture style and/or textbook.  This also solves the problem raised by Greg
recently regarding specifying the display format of the teacher's answer,
since using an Inert Expression as the teacher's answer puts the choice of
formatting under the question author's control.  Finally, it makes it
possible for me to donate some of the AiM questions I've written to the AiM
community, since almost all of my AiM questions are written using the Inert
package and its associated utilities.

Many of the changes in this round of updates are geared toward putting the
"hooks" for the Inert package throughout the system so that AiM becomes
'Inert'-aware.  Thus you can use Inert Expressions just like ordinary Maple
expressions with @-substitution with the t> flag, or as the teacher's answer
with the a> flag, etc.

If all of this sounds a bit vague, I apologize, but if you read the
AuthoringGuide.pdf, it should all make much more sense.

2.  aim/Decimal.mpl

This is a package for dealing with the decimal representations of rational
numbers in a rigorous manner that avoids the round off problems inherent in
working with floats.

3.  aim/SET.mpl

This defines a package that defines an AiM class called SET, which
represents and ordinary mathematical set. Ordinary Maple sets cannot be used
in AiM questions for the most part, because there is no guaranteed order to
the elements of a Maple set (so e.g. if you display the set to the student,
the order of the elements can change each time he reloads the page).  This
package replaces a Maple set with a SET object whose elements do have a
fixed order (but unlike a list, there cannot be duplicate elements). All of
the typical elementary set operations are also defined for SET's in this
package.

4.  aim/Number.mpl

This is a package of utilities for dealing with numbers and their
representations. It has routines for converting a positive integers to a
Roman numerals, Maple integers to their English word names, insert commas in
a large integer (e.g. 1234567 -> 1,234,567), produce the prime factorization
of an integer as a nicely formatted LaTeX string, and a routine similar to
Greg's nth() operation except that it puts the suffix as an exponent in a
LaTeX string.

5. aim/admin/CompileSubject.mpl

I got sick of going around to each of my quizzes and recompiling all of the
questions when I made significant changes to the code to see if the
questions all still compile, so I wrote a new admin command called
CompileSubject that compiles all of the files in all of the question source
directories of all of the non-hidden quizzes in a given subject.  I also put
a link to this on the subject admin page.  If you think this is too
dangerous of a command for the ordinary AiM user to have access to, we can
always delete the link to it from the subject page and just call it from a
shortcut externally or something.

6. AuthoringGuide.pdf (in the docs directory)

This is a temporary rough draft of what I hope will eventually become a
question authoring guide.  Currently it only contains a brief tutorial on
using the Inert package, but I hope to add sections about the Random package
and the Rand command, and the SET, Decimal, and Number packages.  Perhaps
others will want to eventually add sections about the Int, Diff, Trig, etc
packages... i.e. anything that is primarily a question authoring tool.  I
think this would be a big help for new question authors.  I often find
myself writing a particular utility, only to find out later that it already
exists somewhere in the AiM code.  This would be a more direct way to learn
what tools are available.

* In addition to adding the above packages, this round of changes also
includes the following

1. Changes to aim/Random.mpl
   + added several new features and options to Rand (documentation to follow
later)
   + added `Nice/Monomial`(), `Nice/Binomial`(), `Nice/Trinomial`() commands
to produce "Nice" versions of binomials, etc. For example, most textbooks
avoid using binomials like -2x^2-5 whereas something like x^2+1 is used all
of the time.  See the code if you want the exact definition of "Nice".

2. aim/Util.mpl
   + added the Keys() command and two minor utilities.  The Keys() command
is a utility for easily making question selector functions. See the docs for
the Keys() command.

3. aim/admin/Subject.mpl
   + added the link to the CompileSubject command

4. Aim.mpl
   + fixed the bug that was preventing guest access

5. aim/Test.mpl
   + upgraded `aim/Test` so it can handle Inert, InertExpr, and SET objects

6. aim/SyntaxHints
   + modified `aim/EntryHint`() so it can deal with InertExpr's.

7. aim/LaTeX
   + changed the default definition `latex/latex/**` to prevent it
     from making undesirable simplifications. We handle that
     manually in Inert.
   + changed the default definition of `latex/exp` so it formats exp(x) as
     "{{\\boldmath e}^{`latex/print`(x)}}" and exp(1) as {\\boldmath e}
   + changed `latex/latex/*` so it doesn't override the AiM default for
     _LaTeXSmallFractionConstant
   + defined a global variable 'LaTeXDigits', that is analogous to the Maple
variable 'Digits', except that it only affects the number of significant
digits displayed when converting a float to a LaTeX string, but does not
affect the number of digits used for doing calculations. The default value
is 3.
   + defined `latex/latex/float` to provide better conversion of floating
point numbers to LaTeX.
   + upgraded `aim/LaTeX` to handle InertExpr objects
   + defined a TeX() proc that behaves exactly like `aim/LaTeX` except that
strings are returned without adding extra double quotes around them (which
is almost always what you want in AiM).
   + fixed a bug in aim/LaTeX/Subs
   + added five new helper utilities to the aim/LaTeX packaged (mostly
needed to implement other things)

8. aim/MultiQuestion.mpl
   + changed the default time limit in the initialize routine from 2.0
seconds to the value of Config['TimeLimit'].

9. AliceServer.mpl
   + no longer unassign `latex/exp` since we define it in aim/LaTeX
   + define e:=exp(1) globally. Note that this means you probably can't use
e as a local variable in your AiM questions. On the other hand both you and
the students can use e^x for exp(x).
   + define _LaTeXSmallFractionConstant := 5000 (TODO: implement a more
flexible way to set this value).

10. Util.mpl
   + added two technical utilities I needed elsewhere

-----------
Having installed both the above changes and the most recent changes posted
to the archive, I can report that most things seem to be working for me
under Maple 9.  The only things I have noticed so far are:

  + ln is now formatting incorrectly for some reason. It is coming out in
italics. I think latex/latex/symbol is messed up in Maple 9.  I'll try to
debug this.

  + there is an error message:

    Error, (in Class/IndexFunction) Type error in assignment: field
InertFunction in class InertExpr should have type Inert

  That occurs quite frequently, though I can't reproduce the exact
conditions to make it repeatable.  If I change the 3 lines labeled WAS: in
Class/IndexFunction in Class.mpl back to what they were the error to goes
away. I don't really understand that proc or why those changes were made in
the first place so I don't want to just change it back since someone
obviously changed them for a reason. I'll post this to the official bug site
list tomorrow.

  + I still get a sporadic error:

    Validation Note: bad name x.

    when Trying certain questions.  It is non-repeatable. It is always with
one-letter variable when it does occur.  It only ever happens when I Try the
questions or Try a quiz as administrator and the system is filling in the
answer box for me, never  when I attempt the question as a student.  It is
very strange.

OK,... time for bed.  I'll post the bugs and continue testing and working on
documentation tomorrow...

ciao,
KEN

-----------------------------------------
Ken Monks - Professor of Mathematics
            University of Scranton
            Scranton, PA 18510
email: mailto:mo...@sc...
web:   http://www.scranton.edu/~monks
-----------------------------------------

Re: [Aimmath-developers] admin.html and defaultZone (was: Re: installlin.html)

[Greg G. <gr...@ma...>]

On Sun, 31 Aug 2003, Manolis Mavrikis wrote:
> Greg said:
> > Are you saying that if I had run AutoConf.mpl having forgotten to set the
> > port and so generated a DefaultZone.mpl file, that it would have been
> > sufficient to delete the DefaultZone.mpl file, edit ManualConfig.mpl
> > with the right port and the run AutoConf.mpl again, that admin.html,
> > index.html and DefaultZone.mpl would be regenerated?
> 
> If we are talking about a fresh installation then yes (I think) deleting the
> DefaultZone.mpl will do the job. This is from my understanding of what Neil
> told me before and from what I see in the code. ... 
> Neil could perhaps explain this once more ... (I attach his previous email
> at the bottom)

Thanks for that ... I recall having trouble finding exactly where the 
writing of these files happened.

> also I made I change that I want to consult you about. At AutoConf.mpl I
> changed the line
> 
> setconfig('ServletURL', cat("http://",Config['FullContext'],"/Alice")): to
> setconfig('ServletURL', "/Alice");
> 
> I think this is better as it define the link relative to the current
> installation anyway ...

Ah! ... that's an interseting idea, change the definition of ServletURL
to be relative instead of absolute ... that ought to work oughtn't it,
since one never descends into deeper directories. How far have you tested
it?

  Regards,
  Greg

Re: [Aimmath-developers] browser detection warnings etc

[Greg G. <gr...@ma...>]

On Sun, 31 Aug 2003, Manolis Mavrikis wrote:
> Greg proposed to have some way of warning students about their browser and
> the math display.
> 
> I propose the following solution. Instead of the index.html page create an
> index.jsp page that can detect the browser and we can do several things
> about that.
> 
> Along these lines have a look at this page
> 
> http://82.41.200.217:8088/AIM/index.jsp

Sounds excellent ... I haven't looked into the java server pages stuff
... if it can do all that we want that's great.

> at the bottom of the screen there is a formula that will appear using
> unicode or symbol font according to your browser. We can then add a link to
> a page that describes problems if they appear. In a similar way we can bring
> a pop-up that says about the IE bug with 2^(e^2) although I have a feeling
> that this is a tth bug. 

No, it's a definite deficiency of IE 5, nothing to do with tth.

> ... Look how the html code looks for this:
> ...
> 2<sup>(</sup>e<sup>i</sup>) ...

What was your (La)TeX for this? ... did you remember to enclose the
bracketed term in {} ... this is what I got for: $2^{(e^i)}$ 

2<sup>(e<sup>i</sup>)</sup>

Just in case, I checked if that changed when enclosed in \[ \] instead
of $ $. Of course if you have a very old version of tth, it is possible
it may have had that bug. I uses version 3.30 above ... and the current
version bundled with AiM is 3.40.
 
  Regards,
  Greg

Re: [Aimmath-developers] tomcat and AIM (RE:was installlin.html)

[Greg G. <gr...@ma...>]

On Sun, 31 Aug 2003, Manolis Mavrikis wrote:
> Greg asked:
> > Can you explain that please? Have you created a group or user for tomcat4?
> > ... so that it need not run as root, or something? Up 'til now tomcat here
> > has been running as root and hence writes various things like compiled AIM
> > questions with root permissions. Do you avoid that?
> ...
> What happened is at some point I installed tomcat 4.1.something that came
> with an RPM. This creates a tomcat4 user and a tomcat4 group who has
> privildges to start, stop tomcat etc. Also when an application starts it
> gives ownership of all files to tomcat4 user. ...
> Perhaps all of these work with any created user and for instance one could
> create a user (perhaps a special one, I am not a good linux admin sorry) on
> their own and chown -R the whole dir to it and have the same effect... why
> don't you try it ? I would really want to know.

Ahh! ... all makes complete sense now. Thanks Manolis for all the explanations,
(sorry for the delay of my reply). Indeed, I'm sure there is nothing
special about the `tomcat4' user and/or group. It could just as easily be
`tomcat' or `aim' for that matter. I just checked the tomcat doc file
RUNNING.txt and it explains that ports <= 1024 require root permissions
to bind to (hence why with port 80, I've actually needed to run tomcat as
root, as many including Gustav had suggested). What had confused me 
was that with 8080 being the default port that tomcat binds to, I had some
sort of muddled notion that 80 was somehow an abbreviation (since that 
used to be the default port used in previous incarnations of AIM).

Anyway, I'm sure that if you leave server.xml specifying port 8080 then
all that one needs to do is add the manager to the group for tomcat 
(whatever was decided) ... I think I'd recommend (given the usual scenario
that tomcat is used exclusively for AiM), that the user and group for
the whole /usr/local/tomcat directory be set to aim. Then the manager
could login as aim to start up tomcat and edit anything in the webapps
directory as well. Also, if the manager as an ordinary user is in the
group for aim then he/she will be able to read/write/execute anything
that has those group permissions, which may actually mean that he/she
will need to login as aim and do a `chmod g+..' on files (where .. is the
appropriate substring or rwx needed) occasionally.  

> > Is this with any Tomcat e.g. 4.1.18 (what I currently have) ... and with
> > port 80, I guess that becomes
> >
> > http://localhost/manager/stop?path=/AIM
> 
> the manager application is usually disabled you have to add something at the
> server.xml see
> 
> http://jakarta.apache.org/tomcat/tomcat-4.1-doc/manager-howto.html
> 
> ...
> it works remotely but it's pretty safe as it needs a password (which you set
> at the users xml in conf directory BTW if you do the tomcat4 user chmod then
> it's better to leave this directory to root)

That was easy ... as the howto explained I added a line 

  <user name="..."  password="..." roles="manager" />

for appropriate ...s in /usr/local/tomcat/conf/tomcat-users.xml
and the lines

  <Context path="/manager" debug="0" privileged="true" 
           docBase="/usr/local/tomcat/server/webapps/manager"/>

in server.xml to activate the manager, and restarted tomcat. Then

  http://.../manager/stop?path=/AiM

where ... is our server's URL, etc. worked fine. The beauty of this is 
that it closes down just AiM and leaves the other parallel installations
alone ... very nice!

BTW I find the stopaim.sh and startaim.sh scripts in the original
webapps/ROOT/WEB-INF still happily (shuts down)/(starts up) tomcat
... I think you suggested that you needed to use 

/usr/local/tomcat/bin/shutdown.sh 

etc. ... as far as I can work out stopaim.sh and startaim.sh are
simply simplified versions of shutdown.sh and startup.sh.

I'm not sure how much of this we want to tell an installer of AiM.
I guess I'll write some of it and if it starts looking too long and 
complicated, just suggest that he/she read the appropriate docs of
tomcat.

  Regards,
  Greg

RE: [Aimmath-developers] -u flag

[Neil S. <N.P...@sh...>]

> On a related topic, Neil responded to a post about an issue with the 
> seed in Safari with "A quick check with Google shows that there are 
> many known problems with javascript support in Safari.  So, I propose 
> to ignore this issue." This is actually a big issue. IE is no longer 
> supported on OS X. Microsoft has made a public statement that there 
> will be no further development of IE for the Mac. Thus whatever we do 
> with AIM needs to work correctly in browsers other than IE. Everything 
> I have tested indicates the desire of this group to support multiple 
> browsers. I would really like to see AIM continue to embrace all 
> platforms while maintaining W3 standards compliance.

I don't think this is such a big issue.  As I understand it there are 
versions of IE (and probably also Mozilla) that work fine under OS X, and 
that will surely continue to be the case for a while, even without active
support from Microsoft.  The javascript problems in Safari are clearly
bugs rather than disputable incompatibilities, and they embarrassing in
number and importance, so they will surely be fixed in the next release.

Neil

RE: [Aimmath-developers] browser detection warnings etc

[Gustav W D. <gu...@ma...>]

> invloves MathML for Mozilla and XSLTs with stylesheets for IE) and I 'll
> wait for Gustav's moodle-MathML solutions to see how all these can be used
> together.

It is actually Alex's solution more than mine. He is working on it at the
moment and expects to have something working at the beginning of the week.
What we do is of course written in PHP because Moodle is written in PHP. I
don't know how to write jsp pages.

Gustav

[Aimmath-developers] browser detection warnings etc

[Manolis M. <ma...@ma...>]

Greg proposed to have some way of warning students about their browser and
the math display.

I propose the following solution. Instead of the index.html page create an
index.jsp page that can detect the browser and we can do several things
about that.

Along these lines have a look at this page

http://82.41.200.217:8088/AIM/index.jsp

(I don't have a static IP but the last few months it's being like that so
hopefully it will work)

at the bottom of the screen there is a formula that will appear using
unicode or symbol font according to your browser. We can then add a link to
a page that describes problems if they appear. In a similar way we can bring
a pop-up that says about the IE bug with 2^(e^2) although I have a feeling
that this is a tth bug. Look how the html code looks for this:

<table align="center"><tr><td nowrap="nowrap" align="center">
2<sup>(</sup>e<sup>i</sup>)</td></tr></table>
</td></tr></table>

This is wrong the /sup for the first parenthesis should be after it closes !
(not to mention that it shouldn't have been done with sups but tables but
anyway ..this is how tth works. I am working in a much more elaborate
solution that we can probably use in AIM but this is not ready yet (it
invloves MathML for Mozilla and XSLTs with stylesheets for IE) and I 'll
wait for Gustav's moodle-MathML solutions to see how all these can be used
together.

Manolis

[Aimmath-developers] admin.html and defaultZone (was: Re: installlin.html)

[Manolis M. <ma...@ma...>]

Greg said:
>>I was going to add some comments about this, as if one
>>already has an installation using port 80, one has to remember to set
>>that in ManualConfig.mpl ... otherwise admin.html, index.html and
>> DefaultZone.mpl (in root/) are set with :8080 in their paths, and if one
>> adds port 80 to ManualConfig.mpl afterwards and reruns AutocConfig.mpl
>> these files are not rewritten (and you have to manually change them).

Manolis said
> > they are not rewritten because you already had a DeafultZone...
otherwise
> > for new users its fine.

Greg said:
> Are you saying that if I had run AutoConf.mpl having forgotten to set the
> port and so generated a DefaultZone.mpl file, that it would have been
> sufficient to delete the DefaultZone.mpl file, edit ManualConfig.mpl
> with the right port and the run AutoConf.mpl again, that admin.html,
> index.html and DefaultZone.mpl would be regenerated?

If we are talking about a fresh installation then yes (I think) deleting the
DefaultZone.mpl will do the job. This is from my understanding of what Neil
told me before and from what I see in the code. The problem comes when you
have previous subjects in this. But I 've noticed that when you use first
time admin.html the subejcts will be hiden so one just needs to unhide them.
Everything will work provided you kept your previous seetings ... it's ok
for now. Also you have to use your previous password and email as I think
that changes are not propagated to all subjects.

Neil could perhaps explain this once more ... (I attach his previous email
at the bottom)

also I made I change that I want to consult you about. At AutoConf.mpl I
changed the line

setconfig('ServletURL', cat("http://",Config['FullContext'],"/Alice")): to
setconfig('ServletURL', "/Alice");

I think this is better as it define the link relative to the current
installation anyway and apart from helping if you tunnnel connections like I
do and myserver:8080 is not the correct URL anymore it could have also
reduced the problem now as one could just use the same old admin.html that
she had. I think the same could happen to the DocumentURL and maybe some
paths in the .m files ?

Manolis

Neil said:

index.html and admin.html are written by the MakeLoginPage method of the
DefaultZone
object.  This method is defined in aim/Zone.mpl, and it is invoked from
aim/admin/Zone.mpl,
aim/admin/ZoneAdminHome.mpl and aim/admin/ZoneStudentHome.mpl.   The
index.html
file is constructed by gluing together the files studenthead.html and
studentfoot.html, and
inserting a table of subjects built from the Subjects field in the
DefaultZone object.  The
web interface allows you to edit studenthead.html and studentfoot.html
directly.  Similarly,
admin.html is built from zone{head,foot}.html, which can be edited directly.

Currently there is no good way to copy information from an old AIM setup to
a new clean
installation; you just have to set up the subjects and quizzes again by hand
and copy the
question files  across manually.  Of course this is not satisfactory.

[Aimmath-developers] Safari (Re: was -u flag)

[Manolis M. <ma...@ma...>]

> What would Manolis need to have it set the -u for Safari?

added anotherer line in an if statement.. it's really simple. This is now
commited and it seems it works

>Thus whatever we do with AIM needs to work correctly in browsers other than
IE. Everything
> I have tested indicates the desire of this group to support multiple
> browsers. I would really like to see AIM continue to embrace all
> platforms while maintaining W3 standards compliance.

I agree with you ... perhaps the work that Gustav and Alex with MathML are
doing could help towards that

>So any changes I make and commit will have no impact on the
> Windows or Linux installations.

good...then we can work together to fix JavaScript problems in Safari. The
problem really interests me as I am already making changes to WALLIS to work
with Macs and anymore problems would be nice to know.

Manolis

[Aimmath-developers] tomcat and AIM (RE:was installlin.html)

[Manolis M. <ma...@ma...>]

Greg asked:

> Can you explain that please? Have you created a group or user for tomcat4?
> ... so that it need not run as root, or something? Up 'til now tomcat here
> has been running as root and hence writes various things like compiled AIM
> questions with root permissions. Do you avoid that?

The linux I was testing things was not so clean as I thought it would be and
now
I am quite confused and I can't test more things here at the moment.

What happened is at some point I installed tomcat 4.1.something that came
with an RPM. This creates a tomcat4 user and a tomcat4 group who has
privildges to start, stop tomcat etc. Also when an application starts it
gives ownership of all files to tomcat4 user. There is no RPM for 4.1.27 so
I am not sure they still want things to happen this but when I installed
tomcat4.1.27 I gave ownership to tomcat4 and my user (manolis) belongs to
the tomcat4 group as well so I am able to copy, move, delete files etc.

Perhaps all of these work with any created user and for instance one could
create a user (perhaps a special one, I am not a good linux admin sorry) on
their own and chown -R the whole dir to it and have the same effect... why
don't you try it ? I would really want to know.

> Is this with any Tomcat e.g. 4.1.18 (what I currently have) ... and with
> port 80, I guess that becomes
>
> http://localhost/manager/stop?path=/AIM
>

the manager application is usually disabled you have to add something at the
server.xml see

http://jakarta.apache.org/tomcat/tomcat-4.1-doc/manager-howto.html

and then you can do lots of goodies

> I guess that won't work remotely will it? i.e. with
> localhost replaced with the full URL. Do you have to run this as
> the tomcat4 user you mentioned?

it works remotely but it's pretty safe as it needs a password (which you set
at the users xml in conf directory BTW if you do the tomcat4 user chmod then
it's better to leave this directory to root)

I usually have it enabled during development and then I disable it.

Manolis

Re: [Aimmath-developers] -u flag

[Jon B. <jbr...@wo...>]

What would Manolis need to have it set the -u for Safari? The browser 
shows as Browser: Mozilla/5.0 (Macintosh; U; PPC Mac OS X; en-us) 
AppleWebKit/85 (KHTML, like Gecko) Safari/85. Safari is also Unicode 
capable.

On a related topic, Neil responded to a post about an issue with the 
seed in Safari with "A quick check with Google shows that there are 
many known problems with javascript support in Safari.  So, I propose 
to ignore this issue." This is actually a big issue. IE is no longer 
supported on OS X. Microsoft has made a public statement that there 
will be no further development of IE for the Mac. Thus whatever we do 
with AIM needs to work correctly in browsers other than IE. Everything 
I have tested indicates the desire of this group to support multiple 
browsers. I would really like to see AIM continue to embrace all 
platforms while maintaining W3 standards compliance.

Once the beta is released, I will make the necessary changes in the 
configuration files so that AIM will run under OS X. For the most part 
this means copying and pasting the Linux code and making some minor 
changes. So any changes I make and commit will have no impact on the 
Windows or Linux installations. I will also try to create some install 
documents for OS X. They will be very similar to the Linux install 
documents except for the Maple installation.

On Saturday, August 30, 2003, at 06:44  PM, Gustav W Delius wrote:

> Manolis' modified servlet that is setting the -u tag automatically 
> depending
> on browser type works brilliantly with Netscape and Internet Explorer. 
> I
> have committed it to CVS.
>
>
>
>
> -------------------------------------------------------
> This sf.net email is sponsored by:ThinkGeek
> Welcome to geek heaven.
> http://thinkgeek.com/sf
> _______________________________________________
> Aimmath-developers mailing list
> Aim...@li...
> https://lists.sourceforge.net/lists/listinfo/aimmath-developers

Jon Breitenbucher
The College of Wooster
Department of Mathematics and Computer Science
Wooster, Ohio 44691

(330)263-2207
jbr...@wo...
http://jbreitenbuch.wooster.edu/~jonb

[Aimmath-developers] -u flag

[Gustav W D. <gu...@ma...>]

Manolis' modified servlet that is setting the -u tag automatically depending
on browser type works brilliantly with Netscape and Internet Explorer. I
have committed it to CVS.

[Aimmath-developers] bugs - executive summary

[Neil S. <N.P...@sh...>]

I think that all reported bugs are now fixed, but:

1) Please look out for the sporadic "Field Prompt should have type string"
   bug, and let me know if it recurs.

2) It is possible that some old latex/tth problems may have been 
   reintroduced - please keep an eye open for this.

Neil