mathlib-develop — This is a mailing list for discussing the development of the JMathLib program.

[Mathlib-develop] Changes to user functions, and bug-fix

[mark <msp...@ya...>]

Hi

At the moment I'm working on a change to m-file based user functions so t=
hat=20
when the m-file is parsed the syntax tree is serialised as a p-file. This=
=20
would then be loaded in preference to the origional m-file so that it won=
't=20
need to be reparsed.

While doing this I found out that I introduced some bugs when I changed t=
he=20
VariableList to use a hash, so I'm going to have to fix that first.

Best Regards

Mark Sparshatt
=20

Re: [Mathlib-develop] Change to MathLibObject/VariableList implementation

[mark <msp...@ya...>]

On Wednesday 23 Jul 2003 7:46 pm, Stefan Mueller wrote:
> HI Mark,
> I aggree. Please go ahead.
> Stefan.
>
> mark wrote:
> > Hi,
> >
> > Both the MathLibObject and the VariableList classes implement lists o=
f
> > named variables. They are both implemented using ArrayLists.
> >
> > This is ineficient since when finding a variable by name it is necess=
ary
> > to iterate through the list until we find the correct variable.
> >
> > I think that it would be better to implement these classes using a
> > HashMap since it would speed up variable access and is a better fit t=
o
> > the problem.
> >
> > The only downside I can see is that the code needed to iterate throug=
h a
> > hashes key/value pairs is, to put it politely, suboptimal.
> >
> > What I will most likely do is work on this in a local copy of the cod=
e
> > and only commit it if there aren't any problems.
> >
> > Any comments?
> >

Okay I've changed the MathLibObject and VariableList to use a hash and=20
committed them. You'll probably need to do an ant clean  before rebuildin=
g.

Best regards

Mark Sparshatt
=20

Re: [Mathlib-develop] Re: MathLib

[mark <msp...@ya...>]

On Wednesday 23 Jul 2003 7:47 pm, Stefan Mueller wrote:
> Hi Geoff,
>
> JMathLib is written in java and is running on all platforms which
> support Java 2.
>
> A new version of JMathLib was realeased some days ago. Please
> check it out at http://mathlib.sourceforge.net/download.php
>
> Hope this helps,
>    Stefan M.
>
> Geoff Byrns wrote:
> > Stefan,
> >
> > I've seen JMATHLib on the net and it looks very interesting.  Is ther=
e a
> > version that runs on the MAC OS X system?  I am not very familiar wit=
h
> > JAVA but have used MATLAB before and so would like the opportunity to
> > use JMATHLib.  many Thanks
> >
> > Geoff

Maybe it would be a good idea to include a link for people to download th=
e=20
Java runtime on the website.

Best regards

Mark Sparshatt

[Mathlib-develop] Re: MathLib

[Stefan M. <st...@he...>]

Hi Geoff,

JMathLib is written in java and is running on all platforms which
support Java 2.

A new version of JMathLib was realeased some days ago. Please
check it out at http://mathlib.sourceforge.net/download.php

Hope this helps,
   Stefan M.

Geoff Byrns wrote:

> Stefan,
> 
> I've seen JMATHLib on the net and it looks very interesting.  Is there a 
> version that runs on the MAC OS X system?  I am not very familiar with 
> JAVA but have used MATLAB before and so would like the opportunity to 
> use JMATHLib.  many Thanks
> 
> Geoff
> 
> 

-- 
------------------------------------
Dr.-Ing. Stefan Mueller
email: St...@he...
------------------------------------

Re: [Mathlib-develop] Change to MathLibObject/VariableList implementation

[Stefan M. <st...@he...>]

HI Mark,
I aggree. Please go ahead.
Stefan.

mark wrote:
> Hi,
> 
> Both the MathLibObject and the VariableList classes implement lists of named 
> variables. They are both implemented using ArrayLists.
> 
> This is ineficient since when finding a variable by name it is necessary to 
> iterate through the list until we find the correct variable.
> 
> I think that it would be better to implement these classes using a HashMap 
> since it would speed up variable access and is a better fit to the problem.
> 
> The only downside I can see is that the code needed to iterate through a 
> hashes key/value pairs is, to put it politely, suboptimal.
> 
> What I will most likely do is work on this in a local copy of the code and 
> only commit it if there aren't any problems.
> 
> Any comments?
>  
> Best Regards
> 
> Mark Sparshatt
> 
> 
> 
> -------------------------------------------------------
> This SF.Net email sponsored by: Free pre-built ASP.NET sites including
> Data Reports, E-commerce, Portals, and Forums are available now.
> Download today and enter to win an XBOX or Visual Studio .NET.
> http://aspnet.click-url.com/go/psa00100006ave/direct;at.asp_061203_01/01
> _______________________________________________
> Mathlib-develop mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/mathlib-develop
> 

-- 
------------------------------------
Dr.-Ing. Stefan Mueller
email: St...@he...
------------------------------------

[Mathlib-develop] Change to MathLibObject/VariableList implementation

[mark <msp...@ya...>]

Hi,

Both the MathLibObject and the VariableList classes implement lists of na=
med=20
variables. They are both implemented using ArrayLists.

This is ineficient since when finding a variable by name it is necessary =
to=20
iterate through the list until we find the correct variable.

I think that it would be better to implement these classes using a HashMa=
p=20
since it would speed up variable access and is a better fit to the proble=
m.

The only downside I can see is that the code needed to iterate through a=20
hashes key/value pairs is, to put it politely, suboptimal.

What I will most likely do is work on this in a local copy of the code an=
d=20
only commit it if there aren't any problems.

Any comments?
=20
Best Regards

Mark Sparshatt

Re: [Mathlib-develop] New release

[Stefan M. <st...@he...>]

Hi,
I've updated the webpages. A new announcement is pointing to the new 
release.

Kind regards,
   Stefan.

mark wrote:

> Hi
> 
> Version 0.4.4 has been released.
> 
> Before the next release (which I'm hoping will be in the next couple of w=
> eeks)=20
> I'm hoping to clean up the system for producing the releases.
> 
> Best Regards
> 
> Mark Sparshatt
> 
> 
> 
> -------------------------------------------------------
> This SF.net email is sponsored by: VM Ware
> With VMware you can run multiple operating systems on a single machine.
> WITHOUT REBOOTING! Mix Linux / Windows / Novell virtual machines at the
> same time. Free trial click here: http://www.vmware.com/wl/offer/345/0
> _______________________________________________
> Mathlib-develop mailing list
> Mat...@li...
> https://lists.sourceforge.net/lists/listinfo/mathlib-develop
> 

-- 
------------------------------------
Dr.-Ing. Stefan Mueller
email: St...@he...
------------------------------------

Re: [Mathlib-develop] New release

[mark <msp...@ya...>]

Hi

Version 0.4.4 has been released.

Before the next release (which I'm hoping will be in the next couple of w=
=3D
eeks)=3D20
I'm hoping to clean up the system for producing the releases.

Best Regards

Mark Sparshatt

[Mathlib-develop] New release

[mark <msp...@ya...>]

Hi

I'll be doing a new file release tommorow evening. This will be version 0=
=2E4.4.

As we discussed with this release we'll change the license to be the LGPL=
=2E

I've just committed a couple of changes.

I've commented out the lines from VariableToken.java that returned null a=
fter=20
evaluating a function, so now  paramaterless functions work as expected i=
f=20
they are called without brackets.

Now the debug flag is set to false when the interpreter is run. I've adde=
d a=20
-d option to the AWT GUI which sets the flag to true. Also I've modified =
the=20
build.xml to pass through the -d option. Which means that when running th=
e=20
program using ant we'll still get the debug info but people using the she=
ll=20
scripts won't.

Best Regards

Mark Sparshatt

[Mathlib-develop] Function call glitch

[mark <msp...@ya...>]

Hi

I've been look into why calling a function without brackets stops it from=
=20
displaying any results.

I found that the VariableToken.java contains the following lines

ErrorLogger.debugLine("VariableToken: var " + name + " not found: check=20
functions");

// If it is not a variable maybe it's a function or script-file
FunctionToken function =3D new FunctionToken(name, limitTokens);
OperandToken  retValue =3D function.evaluate(null);=20
           =20
// if the "function" is really a script it won't return anything
if (function.isMScript())
     return null;
              =20
>>>// check if a function has been evluated=20
>>>//(in case this functions returns null by default)
>>>if (function.isEvaluated())
>>>       return null;
           =20
if (retValue !=3D null)=20
      return retValue;

since after evaluating a function function.isEvaluated() is always true, =
so=20
this always returns null.

Does anyone know if there is a reason for this check or if I can remove i=
t?

Also at one point we talked about allowing functions with parameters to b=
e=20
called without brackets - sin 2 rather than sin (2). However after thinki=
ng=20
about it there are two disadvantages

1 it makes the parser more complicated
2 it increases the chances of errors in user scripts

so my feeling (for this week ;-) ) is to only allow function calls withou=
t=20
brackets when there are no parameters.

What does everyone else think.

Best Regards

Mark Sparshatt
        =20

Re: [Mathlib-develop] Documentation updates

[Stefan M. <st...@he...>]

mark wrote:

> Hi
> 
> I'm about to start on an update of the MathLib Documentation and wanted to go 
> over a couple of ideas.
> 
> Firstly I'm thinking about splitting the docs into three parts
> User Manual
> Implementation Manual
> Function Reference
> 
> Secondly I think that the implementation manual should be included in the code 
> using Javadoc. I've found a doclet which allows the JavaDoc tool to output 
> docbook xml. I think doing it this way will help keep the docs up to date 
> with the code.
> 
> Does anyone have any comments?
I aggree.

Kind regards,
   Stefan

------------------------------------
Dr.-Ing. Stefan Mueller
email: St...@he...
------------------------------------

[Mathlib-develop] Documentation updates

[mark <msp...@ya...>]

Hi

I'm about to start on an update of the MathLib Documentation and wanted t=
o go=20
over a couple of ideas.

Firstly I'm thinking about splitting the docs into three parts
User Manual
Implementation Manual
Function Reference

Secondly I think that the implementation manual should be included in the=
 code=20
using Javadoc. I've found a doclet which allows the JavaDoc tool to outpu=
t=20
docbook xml. I think doing it this way will help keep the docs up to date=
=20
with the code.

Does anyone have any comments?

Best Regards

Mark Sparshatt

Re: [Fwd: Re: [Mathlib-develop] Status update]

[mark <msp...@ya...>]

Hi

On Monday 07 Jul 2003 9:49 am, mark wrote:
> On Sunday 06 Jul 2003 6:21 pm, Stefan Mueller wrote:
> > mark wrote:
> >   > Improve documentation.
> >
> > I tried to build the documentation some time ago, but couldn't get th=
e
> > tools to work which are mentioned in ant's makefile.
> > Could you please commit the jar-files to CVS.
>
> Okay I'll commit them in the next couple of days.
>

Now all the jar files used to generate the docs are within the=20
Documentation/src/tools directory.
=20
> > Right now the generated documentation files committed to CVS. What ab=
out
> > keeping them out of CVS like it is done with class-files? Most of the
> > documentation is generated from m- or java-files. This would allow us
> > to change the appearance without committing new files. The static fil=
es
> > should be kept in CVS tough.
>
> I agree. There's no need to store any files that are generated dynamica=
lly.
>

I've cleared out all the dynamic files. Or at least I think I have. If yo=
u=20
spot any you miss.

A note about creating the docs. When building the pdf/txt versions you'll=
 get=20
a lot of errors about properties not being implemented. This shouldn't af=
fect=20
the build process.

More importantly generating the pdf anf txt versions takes a long time (a=
bout=20
40 minutes on my machine) so it's not something you want to do often.

Best regards

Mark

Re: [Fwd: Re: [Mathlib-develop] Status update]

[mark <msp...@ya...>]

Hi
On Sunday 06 Jul 2003 6:21 pm, Stefan Mueller wrote:
> mark wrote:
>   > Improve documentation.
>
> I tried to build the documentation some time ago, but couldn't get the
> tools to work which are mentioned in ant's makefile.
> Could you please commit the jar-files to CVS.

Okay I'll commit them in the next couple of days.=20

> Right now the generated documentation files committed to CVS. What abou=
t
> keeping them out of CVS like it is done with class-files? Most of the
> documentation is generated from m- or java-files. This would allow us
> to change the appearance without committing new files. The static files
> should be kept in CVS tough.
>

I agree. There's no need to store any files that are generated dynamicall=
y.
  =20
Best Regards

Mark Sparshatt

[Fwd: Re: [Mathlib-develop] Status update]

[Stefan M. <st...@he...>]

-------- Original Message --------
Subject: Re: [Mathlib-develop] Status update
Date: Sun, 06 Jul 2003 20:06:00 +0200
From: Stefan Mueller <st...@he...>
To: mark <msp...@ya...>
References: <200...@ya...>

mark wrote:

  > Hi
  >
  > Well development seems to have stalled for some time. Given the time
since the
  > last release I think we should work on producing another release
within the
  > next month.

I strongly aggree.


  > The two things I want to do before releasing are
  > remove the Meditor code. This should be easy since it isn't being
used at the
  > moment.

I aggree


  > Improve documentation.

I tried to build the documentation some time ago, but couldn't get the
tools to work which are mentioned in ant's makefile.
Could you please commit the jar-files to CVS.
Right now the generated documentation files committed to CVS. What about
keeping them out of CVS like it is done with class-files? Most of the
documentation is generated from m- or java-files. This would allow us
to change the appearance without committing new files. The static files
should be kept in CVS tough.


  > Is there anything else that needs to be added before the next release?

No. I don't think so. Since the last release has been in September 2002
(which is REALLY a long time ago) any release at all is good!!!!!!!


  > I suggest that we should start using the SourceForge Bug tracking
facilities.
  > I think It'll make keeping track of bugs easier.

Again, I aggree. This is a more professional approach.


  > I've already added a few bugs to the the list. There was one there to
do with
  > the parser which I've closed since the examples are working now.
  >
  > I'll try to devote more time to the project in order to get
development back
  > on track.

I'd also like to devote more time to mathlib again. I've been very very
busy over
the last few month. Hopefully I'll be able to start again.

Kind regards,
     Stefan.

------------------------------------
Dr.-Ing. Stefan Mueller
email: St...@he...
------------------------------------





-- 
------------------------------------
Dr.-Ing. Stefan Mueller
email: St...@he...
------------------------------------

[Mathlib-develop] Status update

[mark <msp...@ya...>]

Hi

Well development seems to have stalled for some time. Given the time sinc=
e the=20
last release I think we should work on producing another release within t=
he=20
next month.

The two things I want to do before releasing are
remove the Meditor code. This should be easy since it isn't being used at=
 the=20
moment.
Improve documentation.

Is there anything else that needs to be added before the next release?

I suggest that we should start using the SourceForge Bug tracking facilit=
ies.=20
I think It'll make keeping track of bugs easier.=20

I've already added a few bugs to the the list. There was one there to do =
with=20
the parser which I've closed since the examples are working now.

I'll try to devote more time to the project in order to get development b=
ack=20
on track.

Best Regards

Mark Sparshatt

Re: [Mathlib-develop] Path Function & problems with funcion calls

[Stefan M. <st...@he...>]

mark wrote:
> On Thursday 20 Feb 2003 7:40 pm, Stefan Mueller wrote:
> 
>>>While testing it I notived a problem with function calls
>>>
>>>entering
>>>
>>>
>>>>path
>>>
>>>runs the path function and sets the ans variable but doesn't display
>>>anything
>>>
>>>while entering
>>>
>>>
>>>>path()
>>>
>>>displays the result of the function call.
>>
>>I'll have a look into it.
>>
>>Kind regards,
>>   Stefan.
> 
> 
> 
> This seems to be connected to the function being treated as a variable rather 
> than a function when the brackets are ommitted.
> 
> There is a similar problem if there is a space between the function name and 
> the brackets. If you enter
> 
>>sin (5)
> 
> 
> then you get a null pointer error.
> 
> After checking it seems that MathLib treats VariableTokens and FunctionTokens 
> in the same way. It first checks for a variable with that name then for a 
> function.
> 
> Therefore it might be an idea to combine VariableTokens and FunctionTokens 
> into one class (Say ReferenceToken) since this would ensure consistent 
> behaviour regardless of how the function or variable is entered.
> 
> One difficulty is the fact that VariableToken descends from OperandToken while  
> FunctionToken descends from OperatorToken. RefereenceToken would have to be a 
> subclass of OperandToken which would mean that it would have to store any 
> parameters being passed to the variable or function.
> 
> Though this would have the advantage of making functions more consistent with 
> keyword functions such as IF and WHILE.

I've changed FunctionToken to be an OperandToken. I also fixed the
bugs "path" and "path   ()"

I'd like to change the opinion of one of my last emails. I didn't 
introduce a
ReferenceToken class. In my opinion the VariablenToken and FunctionToken
classes are somehow still more different, than the same.
The variable token needs to handle stuff like  some_variable.some_field
while the function token needs to handle java-functions and
script-files. These are not huge differences, but the individual purpose
is still different.
The change from operator token to operand token for the functions also
simplified some code in the Expression-class. It seems to be working
quite okay right now. I'm still into simplifying some code.

Kind regards,
   Stefan.

Re: [Mathlib-develop] change name of mathlib

[Stefan M. <st...@he...>]

Alejandro Torras wrote:
>>
>> Hi,
>>
>> what about changing the name of mathlib?
>>
>> I did a search on google and it turned out more than 40200 hits.
>>
>> A small change to Jmathlib turns out just 2 hits.
>>
>> We could  just change the name on the webpages and in all
>> mails, so the search engines will find it in the future.
>>
>> I also thought about a totally different name, like
>> jOnePlus
>> jAddAdd
>> ....
>> these words are not used anywhere on the web.
>>
>> The word "mathlib" is not very unique in my opinion and people
>> won't think of OUR mathlib when they read this word.
>>
>> What do you think about this?
> 
> 
> I agree, no objection but do we have to change everything related to the 
> word "mathlib"? The project name too?

I think it's enough to change from MathLib to JMathLib everywhere in the
text, webpages and emails. If we would start changing the project name
on sourceforge, everything old would be gone. I changed the webpages
already.

Maybe we should change just two more files:
Source/MathLib/UI/Applet/JMathLibGUI.java
Source/MathLib/TUI/JMathLibTUI.java

The distribution files should be renamed to JMathLib.jar .....

Do you all aggree?

Kind regards,
   Stefan.

------------------------------------
Dr.-Ing. Stefan Mueller
email: St...@he...
------------------------------------

Re: [Mathlib-develop] change name of mathlib

[Alejandro T. <ate...@ho...>]

>
>Hi,
>
>what about changing the name of mathlib?
>
>I did a search on google and it turned out more than 40200 hits.
>
>A small change to Jmathlib turns out just 2 hits.
>
>We could  just change the name on the webpages and in all
>mails, so the search engines will find it in the future.
>
>I also thought about a totally different name, like
>jOnePlus
>jAddAdd
>....
>these words are not used anywhere on the web.
>
>The word "mathlib" is not very unique in my opinion and people
>won't think of OUR mathlib when they read this word.
>
>What do you think about this?

I agree, no objection but do we have to change everything related to the 
word "mathlib"? The project name too?

Regards,
Alejandro.

_________________________________________________________________
Help STOP SPAM with the new MSN 8 and get 2 months FREE*  
http://join.msn.com/?page=features/junkmail

Re: [Mathlib-develop] change name of mathlib

[mark <msp...@ya...>]

On Thursday 06 Mar 2003 9:01 am, Stefan Mueller wrote:
> Hi,
>
> what about changing the name of mathlib?
>
> I did a search on google and it turned out more than 40200 hits.
>
Also searching on the SourceForge site brings up five different projects.

> A small change to Jmathlib turns out just 2 hits.
>
> We could  just change the name on the webpages and in all
> mails, so the search engines will find it in the future.
>
> I also thought about a totally different name, like
> jOnePlus
> jAddAdd
> ....
> these words are not used anywhere on the web.
>
> The word "mathlib" is not very unique in my opinion and people
> won't think of OUR mathlib when they read this word.
>
> What do you think about this?
>
I don't have any objections to changing the name. Though I prefer Jmathli=
b.=20

We discussed changing the License. If that's something that Peoople still=
 want=20
to do then it'd be a good idea to change the name to destinguish between =
the=20
two versions under different licences.

Best Regards

Mark

__________________________________________________
Do You Yahoo!?
Everything you'll ever need on one web page
from News and Sport to Email and Music Charts
http://uk.my.yahoo.com

[Mathlib-develop] change name of mathlib

[Stefan M. <st...@he...>]

Hi,

what about changing the name of mathlib?

I did a search on google and it turned out more than 40200 hits.

A small change to Jmathlib turns out just 2 hits.

We could  just change the name on the webpages and in all
mails, so the search engines will find it in the future.

I also thought about a totally different name, like
jOnePlus
jAddAdd
....
these words are not used anywhere on the web.

The word "mathlib" is not very unique in my opinion and people
won't think of OUR mathlib when they read this word.

What do you think about this?

Kind regards,
   Stefan.

------------------------------------
Dr.-Ing. Stefan Mueller
email: St...@he...
------------------------------------

[Mathlib-develop] Constants defined in operator classes

[mark <msp...@ya...>]

Hi

I've defined a set of constants in the different operator classes which=20
contain instances of those classes.

So instead of writing=20
new AddSubOperatorToken('+')

it's now possible to write
AddSubOperatorToken.add

which will reduce the number of objects that need to be created.

Best Regards

Mark

__________________________________________________
Do You Yahoo!?
Everything you'll ever need on one web page
from News and Sport to Email and Music Charts
http://uk.my.yahoo.com

Re: [Mathlib-develop] Re: Symbolic calculations

[mark <msp...@ya...>]

Hi

This is an overview of the state of the Symolic section.

When an expression containting a symbolic token is evaluated it returns a=
=20
symbolic expression. there are three types of SymbolicExpression dependin=
g on=20
the type of the operator

SymbolicExpression
=09ArithmaticSymbolicExpression
=09GeometricSymbolicExpression
=09ExponentialSymbolicExpression

When a function that contains either a SymbolicToken or a SymbolicExpress=
ion=20
is evaluated it returns itself.

ArithmaticSymbolicExpression stores an ArrayList of the operands and a nu=
meric=20
multiplier. So X + Y + Z is stored as a single ArithmaticSymbolicExpressi=
on.

GeometricSymbolicExpression stores an ArrayList of the numerator and=20
denominator as well as a numeric multiplier. So it can store expressions =
like=20
(U*V*W)/(X*Y)

ExponentialSymbolicExpression stores the mantissa and exponent.

At the moment the following functions work to varying degrees
Simplify(exp)
Expand(exp)
Derivative(exp, deriveBy)
Integral(exp, integrateBy)
Subst(exp, old, new)

The following doesn't work

Derivative for expressions of the form U^V
Derivatives only work for a small number of functions
The output from Derivative can be messy

Integrals for expressions of the form U*V, U/V, U^V and fn(U)

Factorization still doesn't work at all

Despite this I'd suggest removing the code for the meditor library since =
it's=20
not being used any more.

if there are any queries, comments or suggestions then let me know  =20

Best Regards

Mark

__________________________________________________
Do You Yahoo!?
Everything you'll ever need on one web page
from News and Sport to Email and Music Charts
http://uk.my.yahoo.com

[Mathlib-develop] moved FunctionToken to OperandToken

[Stefan M. <st...@he...>]

Hi,

I've move the moved FunctionToken from being an OperatorToken to an
OperandToken. This is a first approach to handle variables and functions
differently than before.

Kind regards,
   Stefan.
-- 
------------------------------------
Dr.-Ing. Stefan Mueller
email: St...@he...
------------------------------------

Re: [Mathlib-develop] Missing Files for SymbolicToken

[mark <msp...@ya...>]

On Friday 21 Feb 2003 5:38 pm, Stefan Mueller wrote:
> Hi Mark,
>
> there are a couple of files missing.
> Please see below,
>
> kind regards,
>    Stefan.

I've uploaded all the files to do with symbolic calculations. When runnin=
g the=20
full test I'm getting a lot of errors. I'll try to reduce the number over=
 the=20
next week.

Best Regards

Mark
=20