Nabble - ActiveMath Users

Re: Metadata inheritance by reference: frequently used? For what?

2009-12-08T09:25:21Z

Hi Paul,

2009-12-04 00:53 Paul Libbrecht <paul@...>:
> No, the inherit element is only allowed as a child of the metadata
> element to indicate that this metadata inherits from this other item's
> metadata.
> ...
> As a rule of the thumb:
> - the inherit element actually invokes inheritance
> - the inherit attribute defines the inheritance policy
> maybe we could be more explicit with the names.

Good, that makes a lot of sense to me and keeps everything safe. Also I agree
that the preview is a good feature. I can't imagine a better name than
inherit/@from, but maybe you could rename the attribute into
field/@inherit-policy or @inheritance-policy.

The question what RDF semantics such imports should have, or whether they
should be resolved before even trying to interpret metadata as RDF, is still a
separate one.

> > Right, it's not yet in those DTDs. Instead there is metadata/ @inherits,
> > something that we also have in "our" OMDoc 1.2 spec, but it is not really
> > documented there, that's why I only discovered it recently.
>
> Same for me. I never considered or used the inherit attribute of
> metadata, you taught me about its existance.

So maybe you would also like to remove it from the DTD at that occasion ;-)

> > That could be a pragmatic approach to giving inheritance an RDF semantics:
> > not explicitly representing the fact that the metadatum is inherited in
> > RDF, but first applying inheritance and then extracting the metadata to
> > RDF, where they would appear as if they had not been inherited.
>
> That'd be certainly the first way we'd implement it. Especially since we do
> not plan an RDF realization where a reasoner has everything in memory, at
> least I don't since it's dangerous for speed.

I made a similar experience before, when implementing an OMDoc→OWL
translation. I hard-coded the translation, but then additionally formalized
parts of it in OMDoc. The implementation does not currently make use of that
formalization, but still it was helpful to do the formalization, as it gave me
a better understanding of the translation. And a future implementation might
actually use it.

> We're back on that: good validation.
> Are you not rather aiming at OWL, at least?

I wonder how OWL would be useful for validation. So what is it that we would
like to validate? Suppose we wanted to make certain metadata mandatory and
wanted to check whether they have been given for all elements. We could say
something like

(\exists nameOfMetadataField) \sqsubseteq WellFormedElement

and then check whether all elements are instances of WellFormedElement. But
that could also be checked with a DTD.

Or suppose we wanted to check whether no inappropriate metadata fields are
given. For example, in the 2004 deliverable "LeActiveMath Structure and
Metadata Model" (still up to date?) I see that the "interactivitytype" field
only applies to exercises. So suppose we had in our ontology

lom:interactivitytype rdfs:domain am:Exercise . # (1)
am:Exercise owl:disjointWith am:Proof . # (2)

and then in a document an element

<proof xml:id="p">
<metadata>
<interactivitytype>T</interactivitytype>
</metadata>
</proof>

This might give us the following RDF:

<#p> a am:Proof ; # (3)
lom:interactivitytype "T" . # (4)

Due to the open world reasoning of RDFS and OWL, we'd obtain from (2) that

<#p> a am:Exercise .

However, exercise and proof are disjoint => the reasoner would report a
conflict. But what RDF triple out of (1) to (4) is causing the conflict?
From a metadata authoring point of view, it is clearly (4), but a reasoner
wouldn't be able to tell. So therefore I think that OWL is not the best
choice for _this_ kind of validation.

What else would you like to validate?

> > My plan for OMDoc 1.3 is having RDFa-compatible syntax for full
> > expressivity and extensibility, e.g. <meta
> > property="dc:date">YYYY-MM-DD</meta>, or <meta
> > property="anyOtherOntology:field">value</meta>, but the old syntax would
> > still be allowed, and some translation to the RDFa syntax would be
> > specified.
>
> Extensibility and validatability can be ennemies or?

Well, that case would no longer be validatable with a DTD, but with the OWL
method mentioned above it would work. That is an advantage of working on the
RDF level, that you can abstract from the syntax.

> > The goal for the RDFa syntax is that any RDFa parser can process it, so
> > there should not be any OMDoc-specific assumptions in it. The backwards-
> > compatible syntax, on the other hand, can be as OMDoc-specific or as
> > ActiveMath- specific as we like it to be. So I think inheritance would
> > best be handled at that level. The way you do it, by default attribute
> > values, wouldn't work with the RDFa-like syntax anyway, as AFAIK in a DTD
> > you can only have default values depending on a certain element (as, e.g.,
> > <Date/>), but not depending on an attribute value (as, e.g.,
> > property="dc:date").
>
> But that is the question: is there an RDF infrastructure to replace
> that?

Good question. However, as we discussed before, both of us are still
considering to hard-code part of the inheritance manually. With your DTD
approach, you can only automate part of it, i.e. you can model the default
inheritance policies. Default inheritance policies could also be modeled in
an ontology, e.g. by saying

dc:date am:inheritancePolicy am:InheritIfMissing .

The other aspect is that by the <inherit/> element you can locally override
this policy. That could work in RDFa as follows:

<whatever xml:id="w">
<metadata>
<meta property="dc:date"> 
<resource typeof="am:InheritedField">
<link rel="am:inheritsFrom" resource="#otherElement"/>
<link rel="am:inheritancePolicy" resource="[am:InheritIfMissing]"/>
</resource>
</meta>
</metadata>
</whatever>

... leading to the RDF

<#w> dc:date [ a am:InheritedField ;
am:inheritsFrom <#otherElement> ;
am:inheritancePolicy am:InheritIfMissing . ] .

In that step of processing the reasoner must not be too aggressive, i.e. not
complain about the value of dc:date not being a string (or xsd:date), but a
resource -- but that is feasible. Suppose we have

<#otherElement> dc:date D .

then the application of imports should yield

<#w> dc:date D .

This step can be hard-coded, or implemented as a rule. SWRL's rules are
Datalog with unary and binary predicates. So it could be done like this:

dc:date(Element, D) :- dc:date(Element, I),
am:InheritedField(I),
am:inheritsFrom(I, OtherElement),
am:inheritancePolicy(I, am:InheritIfMissing),
dc:date(OtherElement, D).

This would leave us with one or more triples where the dc:date is an
InheritedField, but we could hard-code an additional step that discards them.

After the application of imports, we can set the reasoner to a more aggressive
mode, e.g. assuming that now the range of dc:date must be xsd:date.

> So RDF is export only?

So far, yes. Of course it is being used for certain things. E.g. the
navigation links in SWiM are created from the RDF exported from OMDoc, and now
we're also putting the exported RDF back as RDFa into the rendered XHTML+PMML
documents, so that e.g. web crawlers or JavaScripts can use it (will be online
before Christmas).

> At least validation!

I agree, it would be good to have that, and therefore one should pay attention
to getting the RDF representation of metadata and inheritance right, but from
our side it won't be done before I got my Ph.D. -- There will only be some
general considerations on validation of metadata in my thesis :-/

> And indeed in all metadata specs of ActiveMath, the attributes of the
> item are part of metadata.

Ah, you mean things like proof/@for? Sure. I don't call it "metadata"
because I believe it's rather primary data in our documents, as opposed to
"metadata = data about data" -- but, as you see in the examples above, I also
extract this information to RDF, so in the end it is a similar approach as in
ActiveMath.

> Amazing, I never thought temporal was needed.
>
> Can you not say, instead, that it's the property Creator of the "first
> enclosing" ? (or is that already temporal).

Hmm, maybe internally maintain two copies of the metadata vocabulary? One
before applying inheritance, one after applying inheritance? Then it would
work as follows:

# the simple case (not sure how negation really works in SWRL, I'll write
# Prolog here)
dc:dateAfterInheritance(Element, D) :-
dc:date(Element, D),
\+ am:InheritedField(D).

dc:dateAfterInheritance(Element, D) :-
dc:date(Element, I),
am:InheritedField(I),
\+ am:inheritsFrom(I, _), # I assume: if no <inherits from="..."/> is
# given, the default is the enclosing element
am:inheritancePolicy(I, am:InheritIfMissing),
am:parentElement(I, P),
dc:dateAfterInheritance(P, D). # not dc:date here, as we only want real
# date values here; the other rule accounts
# for getting them.

Still, validators and other tools would not understand the "after inheritance"
properties, so they would have to be renamed back.

> The below tastes like a jess or prolog rule... too algorithmic!

As this is basically a join of two relations (the metadata field, and
inheritsFrom or parentElement), I suppose that with a bit of brain-twisting
this could also be achieved with OWL 2 property chain inclusions.

> I would even be tempted to say that XPath is almost able to say it in
> a far less algorithmic way (which might be better to understand for a
> human though).

BTW, OWL is certainly less algorithmic than a rule language, but emulating
rules in OWL IMHO leads to things that are even harder to understand for
humans.

> If you had some free properties available, couldn't you define the
> Creator-if-missing, Creator-merge, and and Creator-none properties
> then merge them by, again, a first choice?

That seems related to my above-mentioned set of "after inheritance"
properties. Seems intuitively feasible to me. Again it would require a
hard-coded step of renaming properties.

> I'm impressed by you saying we should stay on OMDoc-based
> infrastucture ;-)

Well, in learning a lot about the advantages of semantic web technologies, I
also learned about their limitations. So I ended up with the rule of thumb:
Use RDF and friends whenever it (and its tools) give you something for free
(e.g. integration with other semantic web applications); otherwise be
pragmatic and hard-code something.

Cheers,

Christoph

--
Christoph Lange, Jacobs Univ. Bremen, http://kwarc.info/clange, Skype duke4701

_______________________________________________
Users@... mailing list
Users@...@activemath.org
http://www.dfki.de/mailman/cgi-bin/listinfo/users.activemath.org

signature.asc (205 bytes) Download Attachment

Re: Metadata inheritance by reference: frequently used? For what?

2009-12-03T15:53:08Z

Le 03-déc.-09 à 12:32, Christoph LANGE a écrit :

>> It's defined as being the same effect as the inheritance of the
>> "closest enclosing" element meaning that the inheritance policy
>> should
>> follow that model. The policy is defined by attribute values, the
>> inherit attribute, which can be inside your current metadata or in an
>> enclosing element.
>
> Hmm, maybe I got it wrong where the inherits element has to be put.
> (Is there
> an exact specification, besides your paper?)

The DTD. that's all.

> So do you mean something like
>
> <exercise id="exercise">
> <metadata>
> <Date><inherit from="#other-exercise"/></Date>
> ...
> </metadata>
> </exercise>

No, the inherit element is only allowed as a child of the metadata
element to indicate that this metadata inherits from this other item's
metadata.

> Now that there is also an inherit _attribute_, which defines the
> mode of
> inheritance, I'm a bit confused. In the video I see that the
> attribute would
> be attached, e.g., to the <Date/> element. But where is the
> _element_ placed?
> Like this?
>
> <exercise id="exercise">
> <metadata>
> <inherit from="#other-exercise"/>
> ...
> </metadata>
> </exercise>

Right. This basically says: add all the (policy-allowing) elements
from this other metadata into here.

> Anyway, I realize that inconsistencies can potentially only be
> caused by wrong
> inheritance mode values for the Field/@inherit attribute.

Right.
So, I think that if a user is:
- able to understand how it goes
- can test it (use the preview)
then it should be acceptable.

As a rule of the thumb:
- the inherit element actually invokes inheritance
- the inherit attribute defines the inheritance policy
maybe we could be more explicit with the names.

>> With all defaults applied (they're all in the DTD), it means that
>> some
>> fields have inherit if missing, others have inherit merge...
>>
>>> What experiences do you (= the ActiveMath community) have with that?
>>
>> It's quite new to gather experience. It hasn't reached the authoring
>> ActiveMath for example.
>> (http://eds.activemath.org/en/AMinstallation/svn-stable)
>
> Right, it's not yet in those DTDs. Instead there is metadata/
> @inherits,
> something that we also have in "our" OMDoc 1.2 spec, but it is not
> really
> documented there, that's why I only discovered it recently.

Same for me. I never considered or used the inherit attribute of
metadata, you taught me about its existance.

>> I also note that the inherit by reference element is not yet
>> implemented but
>> that's just an inch away. It will require to re-parse the given
>> element when
>> an inherit element will be "applied".
>
> That could be a pragmatic approach to giving inheritance an RDF
> semantics:
> not explicitly representing the fact that the metadatum is inherited
> in RDF,
> but first applying inheritance and then extracting the metadata to
> RDF, where
> they would appear as if they had not been inherited.

That'd be certainly the first way we'd implement it.
Especially since we do not plan an RDF realization where a reasoner
has everything in memory, at least I don't since it's dangerous for
speed.

> Modeling the process of
> inheritance completely in RDF, with rules or ontologies, seems
> impossible to
> me, but I think it would not be necessary. So the only remaining
> challenge is
> to prevent documents to which metadata inheritance has not yet been
> applied
> from generating inconsistent RDF.

We're back on that: good validation.
Are you not rather aiming at OWL, at least?

> My plan for OMDoc 1.3 is having RDFa-compatible syntax for full
> expressivity
> and extensibility, e.g. <meta property="dc:date">YYYY-MM-DD</meta>,
> or <meta
> property="anyOtherOntology:field">value</meta>, but the old syntax
> would still
> be allowed, and some translation to the RDFa syntax would be
> specified.

Extensibility and validatability can be ennemies or?

> The
> goal for the RDFa syntax is that any RDFa parser can process it, so
> there
> should not be any OMDoc-specific assumptions in it. The backwards-
> compatible
> syntax, on the other hand, can be as OMDoc-specific or as ActiveMath-
> specific
> as we like it to be. So I think inheritance would best be handled
> at that
> level. The way you do it, by default attribute values, wouldn't
> work with the
> RDFa-like syntax anyway, as AFAIK in a DTD you can only have default
> values
> depending on a certain element (as, e.g., <Date/>), but not
> depending on an
> attribute value (as, e.g., property="dc:date").

But that is the question: is there an RDF infrastructure to replace
that?

>>> Does incorrect use lead to conflicts?
>>
>> everything is made to avoid them, in particular the inheritance
>> policy.
>> I would claim there's none... but I'd have to prove it (may be a fun
>> exercise ;-)).
>
> The way you did it, it seems to me that the author can only break it
> by
> defining wrong defaults for the inherit

A conflict could only be detected with a good validator.

>
>>> https://trac.omdoc.org/OMDoc/ticket/1566
>>
>> So you'd solve this as SWRL rules?
>
> Maybe. To also answer your question about "rules vs. ontology": I
> have not
> yet implemented any of these rules. Secondly, in this context I
> have a quite
> general understanding of "ontology". On the one hand, I'd like to
> reuse
> existing ontologies, such as DC, as metadata vocabularies. They are
> mostly
> implemented in RDFS or OWL, i.e. description logic, which means "no
> rules".
> Well, mostly. The property chain inclusions of OWL 2 already
> emulate certain
> rules with description logic means, consider the following rule for
> inheritance (given in a Prolog-like syntax):
>
> contributor(Section, Person) :-
> partOf(Section, Subsection),
> creator(Subsection, Person).
>
> But it can also modeled in OWL 2 as
>
> partOf \circ creator \sqsubseteq contributor
>
> And then, by what I presented at MKM 2009, we can consider any OMDoc
> theory an
> ontology (and thus a potential metadata vocabulary), so ultimately I
> would
> implement any inheritance rules that should be too complex for RDFS,
> OWL,
> etc., as "something that can be expressed in OMDoc and that some
> OMDoc-aware
> software can make sense of" ;-)

So RDF is export only?
At least validation!
>
>> I like your statement that "metadata does not exist"...
>
> Well, we decided to make it optional in OMDoc 1.3. I believe that
> it gives
> the document more structure when you read or edit it, but from a
> _semantic_
> point of view it is not required.

sure.
And indeed in all metadata specs of ActiveMath, the attributes of the
item are part of metadata.

>> Can you encode if-missing in SWRL ?
>
> I'm not an expert in SWRL but somewhat superficially familiar with
> RDF-based
> rule languages. It is certainly not possible in a straightforward
> way, as
> these languages do not have any notion of temporal aspects.

Amazing, I never thought temporal was needed.

Can you not say, instead, that it's the property Creator of the "first
enclosing" ?
(or is that already temporal).

The below tastes like a jess or prolog rule... too algorithmic!
I would even be tempted to say that XPath is almost able to say it in
a far less algorithmic way (which might be better to understand for a
human though).

> Either some
> metadata field is there or not, but there is no reasoning like "if
> it has not
> been there before, then put it there". Thanks, BTW, for making me
> think about
> that, I hadn't considered that problem in detail before.
>
> However, it might work as follows, e.g. for dc:creator (in pseudo-
> code):
>
> IF Container dc:creator C
> && (
> NOT( Child dc:creator _ )
> || Child dc:creator C
> )
> THEN
> Child dc:creator C
>
> Suppose that initially the child does not have a dc:creator set.
> The first
> application of the rule would inherit it from the container. The
> reasoner
> would continue its work until no more rule can be applied. So it
> would try to
> apply this rule once more, then seeing that container and child have
> equal
> dc:creator properties. "Application" of the rule would then leave
> dc:creator
> as it is, so we would have reached a fixed point.
>
> What I said applies to the dc:creator attribute _in general_. Locally
> modifying the inheritance policy for concrete metadata fields of
> concrete
> OMDoc elements seems impossible to me in normal RDF, unless you
> introduce
> mechanisms to talk about RDF triples, e.g. reification or named
> graphs, but
> that is no longer easy to handle.

If you had some free properties available, couldn't you define the
Creator-if-missing, Creator-merge, and and Creator-none properties
then merge them by, again, a first choice?

>
>>> and will do further investigations. However, I'm interested in how
>>> important explicit inheritance actually is.
>>
>> This is good news because Sergey Sosnovsky, in Cc, will be OWLifying
>> metadata of ActiveMath. There probably is an overlap. Our intent
>> is mostly
>> validation at this stage.
>
> That definitely sounds interesting. @Sergey: Hi, tell me more! I
> have some
> fuzzy ideas about validation in my mind, but I have not yet really
> dealt with
> it. Not sure, though, if OWL is the right language for validation.
> If you
> want to check constraints, such as, e.g. "report an error if the
> metadata
> field F is missing, or if there is a metadata field that does not
> belong
> there", that might be hard to emulate given OWL's open world
> assumption.

It'll have to be cut at some point... I see no other way.

I'm impressed by you saying we should stay on OMDoc-based
infrastucture ;-)

paul

>> I believe the inheritance can be specified as axioms in OWL but had
>> no
>> time to look at it and I preferred to implement it in simple java
>> based on the existing content-storage implementation which still only
>> loads one file at a time.
>
> As said above, this would be the pragmatic approach. My current
> objective is
> introducing RDF for the metadata to make them extensible and
> exchangeable. If
> RDF and RDF-based rule/ontology languages give us some reasoning for
> free, use
> it, e.g. for inheritance. If it is painful, don't use it, but
> assume some
> pre-processing and only extract RDF after that.
>
> Cheers,
>
> Christoph
>
> --
> Christoph Lange, Jacobs Univ. Bremen, http://kwarc.info/clange,
> Skype duke4701

_______________________________________________
Users@... mailing list
Users@...@activemath.org
http://www.dfki.de/mailman/cgi-bin/listinfo/users.activemath.org

Re: Metadata inheritance by reference: frequently used? For what?

2009-12-03T03:32:19Z

Hi Paul,

2009-12-03 00:57 Paul Libbrecht <paul@...>:
> Sorry Christoph, I was not paying attention... I could move back to
> there if you wish.

thanks for your reply! No, actually I think this is the better place for
discussing it, I just hadn't thought of posting it here initially.

> > But what kind of inheritance is that? inherit-if-missing, or
> > inherit-merge?
> > Or can you control which kind to use, e.g. by an additional attribute?
>
> It's defined as being the same effect as the inheritance of the
> "closest enclosing" element meaning that the inheritance policy should
> follow that model. The policy is defined by attribute values, the
> inherit attribute, which can be inside your current metadata or in an
> enclosing element.

Hmm, maybe I got it wrong where the inherits element has to be put. (Is there
an exact specification, besides your paper?) So do you mean something like

<exercise id="exercise">
<metadata>
<Date><inherit from="#other-exercise"/></Date>
...
</metadata>
</exercise>

Now that there is also an inherit _attribute_, which defines the mode of
inheritance, I'm a bit confused. In the video I see that the attribute would
be attached, e.g., to the <Date/> element. But where is the _element_ placed?
Like this?

<exercise id="exercise">
<metadata>
<inherit from="#other-exercise"/>
...
</metadata>
</exercise>

Anyway, I realize that inconsistencies can potentially only be caused by wrong
inheritance mode values for the Field/@inherit attribute.

> With all defaults applied (they're all in the DTD), it means that some
> fields have inherit if missing, others have inherit merge...
>
> > What experiences do you (= the ActiveMath community) have with that?
>
> It's quite new to gather experience. It hasn't reached the authoring
> ActiveMath for example.
> (http://eds.activemath.org/en/AMinstallation/svn-stable)

Right, it's not yet in those DTDs. Instead there is metadata/@inherits,
something that we also have in "our" OMDoc 1.2 spec, but it is not really
documented there, that's why I only discovered it recently.

> I also note that the inherit by reference element is not yet implemented but
> that's just an inch away. It will require to re-parse the given element when
> an inherit element will be "applied".

That could be a pragmatic approach to giving inheritance an RDF semantics:
not explicitly representing the fact that the metadatum is inherited in RDF,
but first applying inheritance and then extracting the metadata to RDF, where
they would appear as if they had not been inherited. Modeling the process of
inheritance completely in RDF, with rules or ontologies, seems impossible to
me, but I think it would not be necessary. So the only remaining challenge is
to prevent documents to which metadata inheritance has not yet been applied
from generating inconsistent RDF.

My plan for OMDoc 1.3 is having RDFa-compatible syntax for full expressivity
and extensibility, e.g. <meta property="dc:date">YYYY-MM-DD</meta>, or <meta
property="anyOtherOntology:field">value</meta>, but the old syntax would still
be allowed, and some translation to the RDFa syntax would be specified. The
goal for the RDFa syntax is that any RDFa parser can process it, so there
should not be any OMDoc-specific assumptions in it. The backwards-compatible
syntax, on the other hand, can be as OMDoc-specific or as ActiveMath-specific
as we like it to be. So I think inheritance would best be handled at that
level. The way you do it, by default attribute values, wouldn't work with the
RDFa-like syntax anyway, as AFAIK in a DTD you can only have default values
depending on a certain element (as, e.g., <Date/>), but not depending on an
attribute value (as, e.g., property="dc:date").

> The metadata inheritance preview is there, yes, this blog entry shows
> a video of it:
> http://eds.activemath.org/en/node/237

Thanks for pointing me there again. Before, I had just _read_ the blog post.
I tend to ignore videos ;-)

> > Does incorrect use lead to conflicts?
>
> everything is made to avoid them, in particular the inheritance policy.
> I would claim there's none... but I'd have to prove it (may be a fun
> exercise ;-)).

The way you did it, it seems to me that the author can only break it by
defining wrong defaults for the inherit

> > https://trac.omdoc.org/OMDoc/ticket/1566
>
> So you'd solve this as SWRL rules?

Maybe. To also answer your question about "rules vs. ontology": I have not
yet implemented any of these rules. Secondly, in this context I have a quite
general understanding of "ontology". On the one hand, I'd like to reuse
existing ontologies, such as DC, as metadata vocabularies. They are mostly
implemented in RDFS or OWL, i.e. description logic, which means "no rules".
Well, mostly. The property chain inclusions of OWL 2 already emulate certain
rules with description logic means, consider the following rule for
inheritance (given in a Prolog-like syntax):

contributor(Section, Person) :-
partOf(Section, Subsection),
creator(Subsection, Person).

But it can also modeled in OWL 2 as

partOf \circ creator \sqsubseteq contributor

And then, by what I presented at MKM 2009, we can consider any OMDoc theory an
ontology (and thus a potential metadata vocabulary), so ultimately I would
implement any inheritance rules that should be too complex for RDFS, OWL,
etc., as "something that can be expressed in OMDoc and that some OMDoc-aware
software can make sense of" ;-)

> I like your statement that "metadata does not exist"...

Well, we decided to make it optional in OMDoc 1.3. I believe that it gives
the document more structure when you read or edit it, but from a _semantic_
point of view it is not required.

> I think any formalization should consider the three inheritance
> policies and should allow each to be authored.

Indeed my plan is to model the inheritance policies that you have defined for
those metadata fields that are supported by ActiveMath (which is a finite set)
as rules in some ontology. That would cover the semantics of everything from
an RDF point of view. (I'm not talking about a backwards-compatible _syntax_
here.) if-missing might be an exception; see below.

> Even though this was a big debate, I tend to believe that authors have
> a chance to even wish metadata inheritance on "typical learning time",
> simply deciding to package "collection of 3-minutes exercises", one
> day in a file. By default it is, of course, disabled, but authors are
> still free to change that policy.

That sounds reasonable to me.

> Can you encode if-missing in SWRL ?

I'm not an expert in SWRL but somewhat superficially familiar with RDF-based
rule languages. It is certainly not possible in a straightforward way, as
these languages do not have any notion of temporal aspects. Either some
metadata field is there or not, but there is no reasoning like "if it has not
been there before, then put it there". Thanks, BTW, for making me think about
that, I hadn't considered that problem in detail before.

However, it might work as follows, e.g. for dc:creator (in pseudo-code):

IF Container dc:creator C
&& (
NOT( Child dc:creator _ )
|| Child dc:creator C
)
THEN
Child dc:creator C

Suppose that initially the child does not have a dc:creator set. The first
application of the rule would inherit it from the container. The reasoner
would continue its work until no more rule can be applied. So it would try to
apply this rule once more, then seeing that container and child have equal
dc:creator properties. "Application" of the rule would then leave dc:creator
as it is, so we would have reached a fixed point.

What I said applies to the dc:creator attribute _in general_. Locally
modifying the inheritance policy for concrete metadata fields of concrete
OMDoc elements seems impossible to me in normal RDF, unless you introduce
mechanisms to talk about RDF triples, e.g. reification or named graphs, but
that is no longer easy to handle.

> > and will do further investigations. However, I'm interested in how
> > important explicit inheritance actually is.
>
> This is good news because Sergey Sosnovsky, in Cc, will be OWLifying
> metadata of ActiveMath. There probably is an overlap. Our intent is mostly
> validation at this stage.

That definitely sounds interesting. @Sergey: Hi, tell me more! I have some
fuzzy ideas about validation in my mind, but I have not yet really dealt with
it. Not sure, though, if OWL is the right language for validation. If you
want to check constraints, such as, e.g. "report an error if the metadata
field F is missing, or if there is a metadata field that does not belong
there", that might be hard to emulate given OWL's open world assumption.

> I believe the inheritance can be specified as axioms in OWL but had no
> time to look at it and I preferred to implement it in simple java
> based on the existing content-storage implementation which still only
> loads one file at a time.

As said above, this would be the pragmatic approach. My current objective is
introducing RDF for the metadata to make them extensible and exchangeable. If
RDF and RDF-based rule/ontology languages give us some reasoning for free, use
it, e.g. for inheritance. If it is painful, don't use it, but assume some
pre-processing and only extract RDF after that.

Cheers,

Christoph

--
Christoph Lange, Jacobs Univ. Bremen, http://kwarc.info/clange, Skype duke4701

_______________________________________________
Users@... mailing list
Users@...@activemath.org
http://www.dfki.de/mailman/cgi-bin/listinfo/users.activemath.org

signature.asc (205 bytes) Download Attachment

Re: Metadata inheritance by reference: frequently used? For what?

2009-12-02T15:57:42Z

Le 03-déc.-09 à 00:04, Christoph LANGE a écrit :
> last week I asked a question on metadata inheritance in the
> ActiveMath
> variant of the OMDoc language on the OMDoc developers' list, but it is
> probably more appropriate to ask it here.

Sorry Christoph, I was not paying attention... I could move back to
there if you wish.

> In Paul's 2009 paper on collection
> management, explicit metadata inheritance by reference is explained as
> follows:
>
>> ActiveMath’s metadata inheritance […] oﬀers metadata
>> inheritance by
>> reference to indicate the intent of inheriting metadata from an
>> item at a
>> fully diﬀerent location. A potential usage may be the sharing of
>> meta- data
>> proﬁles which gather all the information pertinent to the typical
>> users in
>> an educational context. It is expressed with a child such as a
>> child of the
>> metadata element <inherit from="pointer"/> and should be
>> interpreted as the
>> meta- data of the closest enclosing element (ignoring enclosing
>> metadata of
>> enclosing elements of the referenced item).
>
> But what kind of inheritance is that? inherit-if-missing, or
> inherit-merge?
> Or can you control which kind to use, e.g. by an additional attribute?

It's defined as being the same effect as the inheritance of the
"closest enclosing" element meaning that the inheritance policy should
follow that model. The policy is defined by attribute values, the
inherit attribute, which can be inside your current metadata or in an
enclosing element.

With all defaults applied (they're all in the DTD), it means that some
fields have inherit if missing, others have inherit merge...

> What experiences do you (= the ActiveMath community) have with that?

It's quite new to gather experience. It hasn't reached the authoring
ActiveMath for example.
(http://eds.activemath.org/en/AMinstallation/svn-stable)

I also note that the inherit by reference element is not yet
implemented but that's just an inch away. It will require to re-parse
the given element when an inherit element will be "applied".

> Is it frequently used in documents? Do people use it correctly?
> Does jEditOQMath
> or any other tool have particular support for it?

The metadata inheritance preview is there, yes, this blog entry shows
a video of it:
http://eds.activemath.org/en/node/237

> Does incorrect use lead to conflicts?

everything is made to avoid them, in particular the inheritance policy.
I would claim there's none... but I'd have to prove it (may be a fun
exercise ;-)).

> How about, e.g. one resource having an educational metadatum
> difficulty=hard, but inheriting metadata from a resource having
> difficulty=easy?

the DTD says the policy is if-missing so no conflicts here.

> I am currently specifying metadata for OMDoc 1.3. As a short
> summary, I want
> to make metadata extensible, semantic web compatible, but also
> backwards
> compatible with the state of ActiveMath as well as OMDoc 1.2. The
> idea is to
> use RDFa as a syntax, at least for all new features, so that the
> metadata have
> a straightforward RDF semantics. While this will allow us to model
> implicit
> inheritance in an elegant way using ontologies (I will write
> something about
> that), it may get us into trouble w.r.t. explicit inheritance. I
> outlined the
> problem in https://trac.omdoc.org/OMDoc/ticket/1566

So you'd solve this as SWRL rules?
I like your statement that "metadata does not exist"...

I think any formalization should consider the three inheritance
policies and should allow each to be authored.
Even though this was a big debate, I tend to believe that authors have
a chance to even wish metadata inheritance on "typical learning time",
simply deciding to package "collection of 3-minutes exercises", one
day in a file. By default it is, of course, disabled, but authors are
still free to change that policy.

Can you encode if-missing in SWRL ?

> and will do further investigations. However, I'm interested in how
> important explicit inheritance
> actually is.

This is good news because Sergey Sosnovsky, in Cc, will be OWLifying
metadata of ActiveMath.
There probably is an overlap. Our intent is mostly validation at this
stage.

I believe the inheritance can be specified as axioms in OWL but had no
time to look at it and I preferred to implement it in simple java
based on the existing content-storage implementation which still only
loads one file at a time.

paul

PS: are *rules* part of an ontology? This surprises me.

_______________________________________________
Users@... mailing list
Users@...@activemath.org
http://www.dfki.de/mailman/cgi-bin/listinfo/users.activemath.org

smime.p7s (2K) Download Attachment

Metadata inheritance by reference: frequently used? For what?

2009-12-02T15:04:07Z

Dear ActiveMath users,

last week I asked a question on metadata inheritance in the ActiveMath
variant of the OMDoc language on the OMDoc developers' list, but it is
probably more appropriate to ask it here. In Paul's 2009 paper on collection
management, explicit metadata inheritance by reference is explained as
follows:

> ActiveMath’s metadata inheritance […] oﬀers metadata inheritance by
> reference to indicate the intent of inheriting metadata from an item at a
> fully diﬀerent location. A potential usage may be the sharing of meta- data
> proﬁles which gather all the information pertinent to the typical users in
> an educational context. It is expressed with a child such as a child of the
> metadata element <inherit from="pointer"/> and should be interpreted as the
> meta- data of the closest enclosing element (ignoring enclosing metadata of
> enclosing elements of the referenced item).

But what kind of inheritance is that? inherit-if-missing, or inherit-merge?
Or can you control which kind to use, e.g. by an additional attribute?

What experiences do you (= the ActiveMath community) have with that? Is it
frequently used in documents? Do people use it correctly? Does jEditOQMath
or any other tool have particular support for it? Does incorrect use lead to
conflicts? How about, e.g. one resource having an educational metadatum
difficulty=hard, but inheriting metadata from a resource having
difficulty=easy?

I am currently specifying metadata for OMDoc 1.3. As a short summary, I want
to make metadata extensible, semantic web compatible, but also backwards
compatible with the state of ActiveMath as well as OMDoc 1.2. The idea is to
use RDFa as a syntax, at least for all new features, so that the metadata have
a straightforward RDF semantics. While this will allow us to model implicit
inheritance in an elegant way using ontologies (I will write something about
that), it may get us into trouble w.r.t. explicit inheritance. I outlined the
problem in https://trac.omdoc.org/OMDoc/ticket/1566 and will do further
investigations. However, I'm interested in how important explicit inheritance
actually is.

Looking forward to your comments -- thanks in advance,

Christoph

--
Christoph Lange, Jacobs Univ. Bremen, http://kwarc.info/clange, Skype duke4701

_______________________________________________
Users@... mailing list
Users@...@activemath.org
http://www.dfki.de/mailman/cgi-bin/listinfo/users.activemath.org

signature.asc (205 bytes) Download Attachment

Fwd: How do you divide?

2009-11-12T01:09:30Z

Début du message réexpédié :

> Réenvoyé-De : www-math@...
> De : Stephen Watt <smwatt@...>
> Date : 11 novembre 2009 18:47:15 GMT+01:00
> À : www-math@...
> Objet : How do you divide?
> Archived-At: <http://www.w3.org/mid/a207b5150911110947t4c490bccra09cc6b1fb0c06e3@...
> >
>
> Dear www-math participants:
>
> Can I ask you to please spend about 30 seconds participating in a
> mathematical survey? It might even be interesting to you.
>
> The 3rd edition of MathML will include elements to portray certain
> elementary arithmetic operations. In this case we want to make sure
> that the main variants for long division are covered.
>
> Could you please visit the link
>
> http://www.doodle.com/ibre4qknzy2bbt99
>
> and select which method(s) you learned for long division? It would
> be helpful if you indicated which country you were living in at the
> time.
>
> Thank you,
>
> Stephen
>

_______________________________________________
Users@... mailing list
Users@...@activemath.org
http://www.dfki.de/mailman/cgi-bin/listinfo/users.activemath.org

smime.p7s (2K) Download Attachment

generalized metadata inheritance

2009-09-11T08:43:26Z

... is now up and running in ActiveMath and jEditOQMath:

http://eds.activemath.org/en/node/237

paul
_______________________________________________
Users@... mailing list
Users@...@activemath.org
http://www.dfki.de/mailman/cgi-bin/listinfo/users.activemath.org

Re: Metadata for ActiveMath

2009-08-29T09:06:36Z

Dear Ili Farhana Md Mahtar,

unfortunately I cannot answer your question for ActiveMath. I am one of the
developers of OMDoc, but our version of OMDoc and the version used by
ActiveMath have diverged years ago, also in terms of their metadata support.
Until recently, "our" OMDoc (version 1.2) did not have extensible metadata
either. However, we developed a solution that will be incorporated into the
upcoming version 1.3. Tool support in terms of editors is not yet there, but
in terms of extracting these extended metadata to an RDF representation for
easier querying.

See https://svn.omdoc.org/repos/omdoc/trunk/doc/blue/foaf/mkm09.pdf for
background. We basically integrated RDFa into OMDoc. RDFa allows for
adding arbitrary metadata to any XML language -- originally XHTML, but more
are coming; see http://rdfa.info.

Thus, if you already have an ontology for the metadata you need, you could
simply reference it from such RDFa markup. Otherwise, there is also the
possibility of writing your own ontology -- either in one of the ontology
languages known from the semantic web (RDFS or OWL), or in OMDoc itself.

If you are planning to process your additional metadata not within ActiveMath,
but with your own application/library, it might be possible to simply add them
to your OMDoc documents in an RDFa style, and hoping that ActiveMath will
gracefully ignore them.

Hope that helps,

Christoph

--
Christoph Lange, Jacobs Univ. Bremen, http://kwarc.info/clange, Skype duke4701
_______________________________________________
Users@... mailing list
Users@...@activemath.org
http://www.dfki.de/mailman/cgi-bin/listinfo/users.activemath.org

Metadata for ActiveMath

2009-08-25T05:50:55Z

Dear ActiveMath users,

I’ve read a paper titled 'Metadata for ActiveMath'. What I’m interested with is the specific metadata in the <extradata> tag. The paper described about difficulty, abstractness, learningcontext, field and verbosity elements. I think none of the elements above is suitable for my web, because I will develop a web that will adapt with the student learning style (whether the student prefers animated image+less text or plain image+more text). My question is, should I create my own metadata (learning style)? What tool do ActiveMath team used to create metadata?

Thanks

Share your memories online with anyone you want anyone you want.
_______________________________________________
Users@... mailing list
Users@...@activemath.org
http://www.dfki.de/mailman/cgi-bin/listinfo/users.activemath.org

Re: jEditOQMath

2008-08-19T23:09:19Z

Le 20-août-08 à 01:05, Salvatore Enrico Indiogine a écrit :
> I now have a directory called ActiveMath on my MacBook that mounts
> /usr/local/activemath-1.0.1
> I am now reading the jEdit help page "Creating a collection about
> free groups"

You could comment on the following pages:
http://eds.activemath.org/en/node/80

> I created the collection called free_groups
> I do not understand the tree of this collection. Where is it? It
> is not under /usr/local/activemath-1.0.1

Well, it will create it under <activemath-home>/content/ (with the
name of the directory be the name of the collection), maybe creating
content. are you sure you configured activemath-home to be the
directory containing conf? The directory content should be aside of
conf.

> there is no /usr/local/activemath-1.0.1/content directory. Maybe
> it is somewhere else?
> Do I need to create it manually?

You may need to do so.

paul

_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

smime.p7s (2K) Download Attachment

Re: jEditOQMath

2008-08-19T16:05:36Z

Paul:

I now have a directory called ActiveMath on my MacBook that mounts
/usr/local/activemath-1.0.1

I am now reading the jEdit help page "Creating a collection about free groups"

I created the collection called free_groups

I do not understand the tree of this collection. Where is it? It is
not under /usr/local/activemath-1.0.1

there is no /usr/local/activemath-1.0.1/content directory. Maybe it
is somewhere else? Do I need to create it manually?

Thanks,
Enrico

--
Enrico Indiogine

Mathematics Education
Texas A&M University
hindiogine@...
_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

Re: jEditOQMath

2008-08-19T15:44:31Z

Paul:

2008/8/19 Paul Libbrecht <paul@...>:
>> I have ActiveMath installed on a Linux (Ubuntu) server on my home LAN.
>> I have jEdit running from Java webstart on my MacBook OSX 10.4. So
>> far so good.
>
> That's a nice idea, it requires a bit of tweaks but has chances.

I was forced to so since I can not run jEditOQMath on my Linux
machine. Anyway, it makes sense, run ActiveMath on a Linux server and
jEditOQMath on a client computer where the coding is done.

> why /conf ? It should just be /usr/local/activemath-1.0.1/
>
>> However, when I check "Apply" I get the following message:
>> This directory cannot be an activemath-home directory!
>> I guess I am doing something wrong here.
>
> The test there just checks for a lib folder with a file called
> activemath.jar, so changing the path should set it.

Yes, that works.

Thanks,
Enrico

--
Enrico Indiogine

Mathematics Education
Texas A&M University
hindiogine@...
_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

Re: jEditOQMath

2008-08-19T15:15:03Z

Hello Salvatore,

Le 19-août-08 à 23:53, Salvatore Enrico Indiogine a écrit :
> I have ActiveMath installed on a Linux (Ubuntu) server on my home LAN.
> I have jEdit running from Java webstart on my MacBook OSX 10.4. So
> far so good.

That's a nice idea, it requires a bit of tweaks but has chances.

> In jEdit:
> Options: Plugins: OQMath JEdit
> - URL of your ActiveMath:
> http://192.168.1.1:8080/ActiveMath2/ <- this works over the network

(to the Linux box I suppose, yes, that's good).

> Check Activate LibreSource update on Build <- do I need to check
> this?

No.

> Choose the directory containing ActiveMath (local run).
>
> /Users/salvatoreenricoindiogine/AMconf
>
> AMconf is an NFS export of /usr/local/activemath-1.0.1/conf directory
> on the Linux server that is mounted on the MacBook. The NFS mount
> works.

why /conf ? It should just be /usr/local/activemath-1.0.1/

> However, when I check "Apply" I get the following message:
> This directory cannot be an activemath-home directory!
> I guess I am doing something wrong here.

The test there just checks for a lib folder with a file called
activemath.jar, so changing the path should set it.

Since you're working over a shared server, there's no other work to do.
But if you want a looser coupling, you can use rsync scripts at upload
time, this is done in LeAM_calculus (if you have a special property).

paul

_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

smime.p7s (2K) Download Attachment

Re: jEditOQMath

2008-08-19T14:53:15Z

Greetings!

I have ActiveMath installed on a Linux (Ubuntu) server on my home LAN.
I have jEdit running from Java webstart on my MacBook OSX 10.4. So
far so good.

In jEdit:

Options: Plugins: OQMath JEdit

URL of your ActiveMath:

http://192.168.1.1:8080/ActiveMath2/ <- this works over the network

Check Activate LibreSource update on Build <- do I need to check this?

Choose the directory containing ActiveMath (local run).

/Users/salvatoreenricoindiogine/AMconf

AMconf is an NFS export of /usr/local/activemath-1.0.1/conf directory
on the Linux server that is mounted on the MacBook. The NFS mount
works.

However, when I check "Apply" I get the following message:

This directory cannot be an activemath-home directory!

I guess I am doing something wrong here.

Thanks,
Enrico

--
Enrico Indiogine

Mathematics Education
Texas A&M University
hindiogine@...
_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

Re: jEditOQMath

2008-08-13T15:11:40Z

Enrico,

this seems to be, again, a bug in the java web-start installation
which does not seem to be the sun one:
at
net
.sourceforge
.jnlp.services.XFileContents.getInputStream(XFileContents.java:68)
can it be the javaws command in the path is still the old one?

thanks

paul

Le 13-août-08 à 23:45, Salvatore Enrico Indiogine a écrit :

> Paul:
>
> 2008/8/12 Paul Libbrecht <paul@...>:
>>
>> It would be interesting to debug that but I acknowledge that Java
>> web start
>> on Linux tends to be picky. Try
>> javaws
>> http://dave.activemath.org/~ilo/jnlpDev/jEditOQMath/jnlps/jEditOQMath-over-jnlp-0.4-SNAPSHOT.jnlp
>
> I replaced OpenJDK with Sun JDK and ran the above command. Here is
> the error message:
>
> henk@bahji:~$ javaws
> http://dave.activemath.org/~ilo/jnlpDev/jEditOQMath/jnlps/jEditOQMath-over-jnlp-0.4-SNAPSHOT.jnlp
> Now launching WebstartJEditLauncher
> Trouble at single-instance-service:
> javax.jnlp.UnavailableServiceException:
> javax.jnlp.SingleInstanceService
> org.activemath.author.webstart.simpleinstaller.JNLPUtil
> $JNLPUtilException:
> Can't read muffin of path
> http://dave.activemath.org/~ilo/jnlpDev/jedit-home-install-witness.
> at
> org
> .activemath
> .author
> .webstart.simpleinstaller.JNLPUtil.getMuffinValue(JNLPUtil.java:122)
> at
> org
> .activemath
> .author
> .webstart
> .simpleinstaller.JNLPUtil.getInstalledDirectory(JNLPUtil.java:167)
> at
> org
> .activemath
> .author
> .webstartjedit
> .WebstartJEditLauncher
> .findjEditHomeFromJNLPUtil(WebstartJEditLauncher.java:223)
> at
> org
> .activemath
> .author
> .webstartjedit
> .WebstartJEditLauncher.ensureJEditHome(WebstartJEditLauncher.java:207)
> at
> org
> .activemath
> .author
> .webstartjedit.WebstartJEditLauncher.main(WebstartJEditLauncher.java:
> 91)
> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> sun
> .reflect
> .NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
> at
> sun
> .reflect
> .DelegatingMethodAccessorImpl
> .invoke(DelegatingMethodAccessorImpl.java:43)
> at java.lang.reflect.Method.invoke(Method.java:616)
> at
> net.sourceforge.jnlp.Launcher.launchApplication(Launcher.java:328)
> at net.sourceforge.jnlp.Launcher$TgThread.run(Launcher.java:561)
> Caused by: java.lang.reflect.UndeclaredThrowableException
> at $Proxy8.getInputStream(Unknown Source)
> at
> org
> .activemath
> .author
> .webstart.simpleinstaller.JNLPUtil.getMuffinValue(JNLPUtil.java:109)
> ... 10 more
> Caused by: java.security.PrivilegedActionException:
> java.lang.reflect.InvocationTargetException
> at java.security.AccessController.doPrivileged(Native Method)
> at net.sourceforge.jnlp.services.ServiceUtil
> $PrivilegedHandler.invoke(ServiceUtil.java:162)
> ... 12 more
> Caused by: java.lang.reflect.InvocationTargetException
> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> sun
> .reflect
> .NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
> at
> sun
> .reflect
> .DelegatingMethodAccessorImpl
> .invoke(DelegatingMethodAccessorImpl.java:43)
> at java.lang.reflect.Method.invoke(Method.java:616)
> at net.sourceforge.jnlp.services.ServiceUtil$PrivilegedHandler
> $1.run(ServiceUtil.java:158)
> ... 14 more
> Caused by: java.io.FileNotFoundException:
> /tmp/pcache/http/dave.activemath.org/~ilo/jnlpDev/jedit-home-install-
> witness
> (No such file or directory)
> at java.io.FileInputStream.open(Native Method)
> at java.io.FileInputStream.<init>(FileInputStream.java:137)
> at
> net
> .sourceforge
> .jnlp.services.XFileContents.getInputStream(XFileContents.java:68)
> ... 19 more
> Can't find jEdit-home from JNLP.
> jEdit home found: "null"
> java.lang.IllegalStateException: The jEdit-home found doesn't seem to
> match. Found at "null".
> at
> org
> .activemath
> .author
> .webstartjedit
> .WebstartJEditLauncher.ensureJEditHome(WebstartJEditLauncher.java:211)
> at
> org
> .activemath
> .author
> .webstartjedit.WebstartJEditLauncher.main(WebstartJEditLauncher.java:
> 91)
> at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
> at
> sun
> .reflect
> .NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
> at
> sun
> .reflect
> .DelegatingMethodAccessorImpl
> .invoke(DelegatingMethodAccessorImpl.java:43)
> at java.lang.reflect.Method.invoke(Method.java:616)
> at net.sourceforge.jnlp.Launcher.launchApplication(Launcher.java:328)
> at net.sourceforge.jnlp.Launcher$TgThread.run(Launcher.java:561)
>
>
>
>
>
>
> --
> Enrico Indiogine
>
> Mathematics Education
> Texas A&M University
> hindiogine@...
> _______________________________________________
> Users@... mailing list
> Users@...@dfki.de
> http://www.dfki.de/mailman/listinfo/users.activemath.org

_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

smime.p7s (2K) Download Attachment

Re: jEditOQMath

2008-08-13T14:45:27Z

Paul:

2008/8/12 Paul Libbrecht <paul@...>:
>
> It would be interesting to debug that but I acknowledge that Java web start
> on Linux tends to be picky. Try
> javaws
> http://dave.activemath.org/~ilo/jnlpDev/jEditOQMath/jnlps/jEditOQMath-over-jnlp-0.4-SNAPSHOT.jnlp

I replaced OpenJDK with Sun JDK and ran the above command. Here is
the error message:

henk@bahji:~$ javaws
http://dave.activemath.org/~ilo/jnlpDev/jEditOQMath/jnlps/jEditOQMath-over-jnlp-0.4-SNAPSHOT.jnlp
Now launching WebstartJEditLauncher
Trouble at single-instance-service:
javax.jnlp.UnavailableServiceException:
javax.jnlp.SingleInstanceService
org.activemath.author.webstart.simpleinstaller.JNLPUtil$JNLPUtilException:
Can't read muffin of path
http://dave.activemath.org/~ilo/jnlpDev/jedit-home-install-witness.
at org.activemath.author.webstart.simpleinstaller.JNLPUtil.getMuffinValue(JNLPUtil.java:122)
at org.activemath.author.webstart.simpleinstaller.JNLPUtil.getInstalledDirectory(JNLPUtil.java:167)
at org.activemath.author.webstartjedit.WebstartJEditLauncher.findjEditHomeFromJNLPUtil(WebstartJEditLauncher.java:223)
at org.activemath.author.webstartjedit.WebstartJEditLauncher.ensureJEditHome(WebstartJEditLauncher.java:207)
at org.activemath.author.webstartjedit.WebstartJEditLauncher.main(WebstartJEditLauncher.java:91)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:616)
at net.sourceforge.jnlp.Launcher.launchApplication(Launcher.java:328)
at net.sourceforge.jnlp.Launcher$TgThread.run(Launcher.java:561)
Caused by: java.lang.reflect.UndeclaredThrowableException
at $Proxy8.getInputStream(Unknown Source)
at org.activemath.author.webstart.simpleinstaller.JNLPUtil.getMuffinValue(JNLPUtil.java:109)
... 10 more
Caused by: java.security.PrivilegedActionException:
java.lang.reflect.InvocationTargetException
at java.security.AccessController.doPrivileged(Native Method)
at net.sourceforge.jnlp.services.ServiceUtil$PrivilegedHandler.invoke(ServiceUtil.java:162)
... 12 more
Caused by: java.lang.reflect.InvocationTargetException
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:616)
at net.sourceforge.jnlp.services.ServiceUtil$PrivilegedHandler$1.run(ServiceUtil.java:158)
... 14 more
Caused by: java.io.FileNotFoundException:
/tmp/pcache/http/dave.activemath.org/~ilo/jnlpDev/jedit-home-install-witness
(No such file or directory)
at java.io.FileInputStream.open(Native Method)
at java.io.FileInputStream.<init>(FileInputStream.java:137)
at net.sourceforge.jnlp.services.XFileContents.getInputStream(XFileContents.java:68)
... 19 more
Can't find jEdit-home from JNLP.
jEdit home found: "null"
java.lang.IllegalStateException: The jEdit-home found doesn't seem to
match. Found at "null".
at org.activemath.author.webstartjedit.WebstartJEditLauncher.ensureJEditHome(WebstartJEditLauncher.java:211)
at org.activemath.author.webstartjedit.WebstartJEditLauncher.main(WebstartJEditLauncher.java:91)
at sun.reflect.NativeMethodAccessorImpl.invoke0(Native Method)
at sun.reflect.NativeMethodAccessorImpl.invoke(NativeMethodAccessorImpl.java:57)
at sun.reflect.DelegatingMethodAccessorImpl.invoke(DelegatingMethodAccessorImpl.java:43)
at java.lang.reflect.Method.invoke(Method.java:616)
at net.sourceforge.jnlp.Launcher.launchApplication(Launcher.java:328)
at net.sourceforge.jnlp.Launcher$TgThread.run(Launcher.java:561)

--
Enrico Indiogine

Mathematics Education
Texas A&M University
hindiogine@...
_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

Re: jEditOQMath

2008-08-13T13:09:10Z

Enrico,

believe me
http://www.activemath.org/projects/downloads/jEditOQMath/
(with the trailing slash) does respond a directory list with the
following content:
Index of /projects/downloads/jEditOQMath
• Parent Directory
• jEditOQMath-0.4-SNAPSHOT.tar.gz
• jEditOQMath-0.4-SNAPSHOT.zip
where the two links are:
http://www.activemath.org/projects/downloads/jEditOQMath/jEditOQMath-0.4-SNAPSHOT.tar.gz
http://www.activemath.org/projects/downloads/jEditOQMath/jEditOQMath-0.4-SNAPSHOT.zip

It's just the front proxy which does not relay to the right place if a
directory and not with slash.

paul

Le 13-août-08 à 18:22, Salvatore Enrico Indiogine a écrit :

> Paul:
>
> 2008/8/12 Paul Libbrecht <paul@...>:
>>
>> Sorry, it should go to:
>> http://www.activemath.org/projects/downloads/jEditOQMath/
>> (note the trailing slash)
>
> Sorry, but this does not work. http://www.activemath.org/projects
> exists, but not
> http://www.activemath.org/projects/downloads/jEditOQMath
>
> Enrico
>
>
>
>
> --
> Enrico Indiogine
>
> Mathematics Education
> Texas A&M University
> hindiogine@...
> _______________________________________________
> Users@... mailing list
> Users@...@dfki.de
> http://www.dfki.de/mailman/listinfo/users.activemath.org

_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

smime.p7s (2K) Download Attachment

Re: jEditOQMath

2008-08-13T09:22:39Z

Paul:

2008/8/12 Paul Libbrecht <paul@...>:
>
> Sorry, it should go to:
> http://www.activemath.org/projects/downloads/jEditOQMath/
> (note the trailing slash)

Sorry, but this does not work. http://www.activemath.org/projects
exists, but not
http://www.activemath.org/projects/downloads/jEditOQMath

Enrico

--
Enrico Indiogine

Mathematics Education
Texas A&M University
hindiogine@...
_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

Re: jEditOQMath

2008-08-13T06:32:18Z

Paul:

2008/8/12 Paul Libbrecht <paul@...>:
>> netx:
>> /tmp/cache/http/dave.activemath.org/~ilo/jnlpDev/jEditOQMath-over-jnlp-0.4-SNAPSHOT.jnkp
>> (No such file or directory)
>
> That tastes likes the worst javaws I've ever heard, is it java-web-start of
> sun?
> You can download the file then javaws <thei-file>.

It is actually the openjdk. I could install the Sun JDK and see it
that works better. I will let you know.

Danke,
Enrico

--
Enrico Indiogine

Mathematics Education
Texas A&M University
hindiogine@...
_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

Re: jEditOQMath

2008-08-12T12:36:42Z

Le 12-août-08 à 18:08, Salvatore Enrico Indiogine a écrit :

> Hi Paul!
>
> 2008/8/12 Paul Libbrecht <paul@...>:
>> It would be interesting to debug that but I acknowledge that Java
>> web start
>> on Linux tends to be picky. Try
>> javaws
>> http://dave.activemath.org/~ilo/jnlpDev/jEditOQMath/jnlps/jEditOQMath-over-jnlp-0.4-SNAPSHOT.jnlp
>
> Did not work. The error message is:
>
> netx: /tmp/cache/http/dave.activemath.org/~ilo/jnlpDev/jEditOQMath-
> over-jnlp-0.4-SNAPSHOT.jnkp
> (No such file or directory)

That tastes likes the worst javaws I've ever heard, is it java-web-
start of sun?
You can download the file then javaws <thei-file>.

>> On the jEditOQMath page, there's a download link which gives you a
>> zip with
>> all the jars. Please note that jEditOQMath (as also included there)
>> contains
>> a few more tweaks than just he plugins, in particular wrt copy-and-
>> paste and
>> the jE button.
>
> The dowload link does not work. Firefox times out.

Sorry, it should go to:
http://www.activemath.org/projects/downloads/jEditOQMath/
(note the trailing slash)
We had several such issues since the move to a CMS.

I'll fix this for next site-upload.

paul

_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

smime.p7s (2K) Download Attachment

Re: jEditOQMath

2008-08-12T09:08:07Z

Hi Paul!

2008/8/12 Paul Libbrecht <paul@...>:
> It would be interesting to debug that but I acknowledge that Java web start
> on Linux tends to be picky. Try
> javaws
> http://dave.activemath.org/~ilo/jnlpDev/jEditOQMath/jnlps/jEditOQMath-over-jnlp-0.4-SNAPSHOT.jnlp

Did not work. The error message is:

netx: /tmp/cache/http/dave.activemath.org/~ilo/jnlpDev/jEditOQMath-over-jnlp-0.4-SNAPSHOT.jnkp
(No such file or directory)

> On the jEditOQMath page, there's a download link which gives you a zip with
> all the jars. Please note that jEditOQMath (as also included there) contains
> a few more tweaks than just he plugins, in particular wrt copy-and-paste and
> the jE button.

The dowload link does not work. Firefox times out.

Gruessen und danke,

Enrico

--
Enrico Indiogine

Mathematics Education
Texas A&M University
hindiogine@...
_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

Re: jEditOQMath

2008-08-11T22:53:02Z

Hello Enrico,

beware that using a different email address triggers moderation... so
it took me till this morning to allow yout mail.

Le 12-août-08 à 05:53, Salvatore Enrico Indiogine a écrit :
> I have installed ActiveMath on a Ubuntu PC. I have jEdit-OQMath
> running on a Mac using Java Webstart.

> I am not able to run
> jEdit-OQMath from Java Webstart on he Ubuntu PC. I receive an
> UnavailableServiceException error when I start jEdit-OQMath using
> webstart on Ubuntu.

It would be interesting to debug that but I acknowledge that Java web
start on Linux tends to be picky. Try
javaws http://dave.activemath.org/~ilo/jnlpDev/jEditOQMath/jnlps/jEditOQMath-over-jnlp-0.4-SNAPSHOT.jnlp

> However, I am able to install jEdit on Ubuntu and have installed all
> the plugins that jEdit-OQMath has except the following:
> OQMath, Templates, Jelly Plugin, and OMPEditor.
> Is it possible to obtain these plugins and install them on my Ubuntu
> PC?

On the jEditOQMath page, there's a download link which gives you a zip
with all the jars. Please note that jEditOQMath (as also included
there) contains a few more tweaks than just he plugins, in particular
wrt copy-and-paste and the jE button.

hope it helps

paul

_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

smime.p7s (2K) Download Attachment

jEditOQMath

2008-08-11T20:53:04Z

Greetings!

I have installed ActiveMath on a Ubuntu PC. I have jEdit-OQMath
running on a Mac using Java Webstart. I am not able to run
jEdit-OQMath from Java Webstart on he Ubuntu PC. I receive an
UnavailableServiceException error when I start jEdit-OQMath using
webstart on Ubuntu.

However, I am able to install jEdit on Ubuntu and have installed all
the plugins that jEdit-OQMath has except the following:

OQMath, Templates, Jelly Plugin, and OMPEditor.

Is it possible to obtain these plugins and install them on my Ubuntu PC?

Thanks,
Enrico

--
Enrico Indiogine

Mathematics Education
Texas A&M University
hindiogine@...
_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

Re: Annoucement: LaTeX2OQMath

2008-06-27T06:24:30Z

Dear Eric,

this is nice to see such an announcement. Could you also post an OQMath
and OMDoc version of your test data on the web site?

From the code I could not really see how the system works. But there is
one way, where we can help you; we have a LaTeXML web service running at
http://tex2xml.kwarc.info (well actually, it seems to be down just now,
but we will restart it soon :-)). You could just integrate the curl
expression in the docu into your web script.

The background for this web service is the arXMLiv project, which
extends the set of LaTeXML bindings (and thus the packages you can use
in your LaTeX sources. On http://arXMLiv.kwarc.info, you can find the
project and over 200000 LaTeX papers and their xhtml+mathml form you
could test your conversion.

Michael

Eric Andres wrote:

> Dear ActiveMath users,
>
> We have implemented a working prototype of a conversion tool
> transforming LaTeX to OQMath.
> For more information and downloads, please visit
> http://www.activemath.org/~eandres/l2o.php
>
> Best,
>
> Eric Andrès
> _______________________________________________
> Users@... mailing list
> Users@...@dfki.de
> http://www.dfki.de/mailman/listinfo/users.activemath.org
>

--
----------------------------------------------------------------------
Prof. Dr. Michael Kohlhase, Office: Research 1, Room 62
Professor of Computer Science Campus Ring 12,
School of Engineering & Science D-28759 Bremen, Germany
Jacobs University Bremen* tel/fax: +49 421 200-3140/-493140
m.kohlhase@... http://kwarc.info/kohlhase
skype: m.kohlhase * International University Bremen until Feb. 2007
----------------------------------------------------------------------

_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

Annoucement: LaTeX2OQMath

2008-06-27T05:07:16Z

Dear ActiveMath users,

We have implemented a working prototype of a conversion tool
transforming LaTeX to OQMath.
For more information and downloads, please visit http://www.activemath.org/~eandres/l2o.php

Best,

Eric Andrès
_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

Re: enhanced access to users@activemath.org

2008-04-25T04:58:36Z

Oh, I should have invested a few more minutes. You can also access the forum
via Nabble's web site at:

http://www.nabble.com/ActiveMath-Users-f32974.html

It's advantage is that the forum uses the full window width.

Happy discussion!
MH

>
> Dear Members of the ActiveMath Users mailing-list,
>
> I am happy to announce a slightly enhanced version of our mailing-list
> archive thanks to the usage of the Nabble Forum service.
>
> On:
>
> http://www.activemath.org/projects/maillists/users
>
> you now find a browsable archive of our mailing-list.
>
> This mail is also to invite you to raise the usage of this list for
> any technical question, also authoring questions instead of direct
> mailing to individual developers of ActiveMath.
>
> thanks in advance
>
> paul
> ------------------------------------------------------------------------
>
> _______________________________________________
> Users@... mailing list
> Users@...@dfki.de
> http://www.dfki.de/mailman/listinfo/users.activemath.org
>

_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

Re: enhanced access to users@activemath.org

2008-04-25T04:50:38Z

Is there a link to the Nabbles Discussion Forum? Or is the forum completely
hosted at activemath.org?

MH

> Dear Members of the ActiveMath Users mailing-list,
>
> I am happy to announce a slightly enhanced version of our mailing-list
> archive thanks to the usage of the Nabble Forum service.
>
> On:
>
> http://www.activemath.org/projects/maillists/users
>
> you now find a browsable archive of our mailing-list.
>
> This mail is also to invite you to raise the usage of this list for
> any technical question, also authoring questions instead of direct
> mailing to individual developers of ActiveMath.
>
> thanks in advance
>
> paul
> ------------------------------------------------------------------------
>
> _______________________________________________
> Users@... mailing list
> Users@...@dfki.de
> http://www.dfki.de/mailman/listinfo/users.activemath.org
>

_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

enhanced access to users@activemath.org

2008-04-25T04:11:55Z

Dear Members of the ActiveMath Users mailing-list,

I am happy to announce a slightly enhanced version of our mailing-
list archive thanks to the usage of the Nabble Forum service.

On:

http://www.activemath.org/projects/maillists/users

you now find a browsable archive of our mailing-list.

This mail is also to invite you to raise the usage of this list for
any technical question, also authoring questions instead of direct
mailing to individual developers of ActiveMath.

thanks in advance

paul

_______________________________________________
Users@... mailing list
Users@...@dfki.de
http://www.dfki.de/mailman/listinfo/users.activemath.org

smime.p7s (2K) Download Attachment

ActiveMath Release 1.0 available

2008-02-01T05:21:55Z

We are pleased to announce that release 1.0 of ActiveMath is finally
available.

ActiveMath 1.0 is a stabilization of years of ActiveMath development and
a polish of most of its features. The highlights of this release are:

* New learner model SLM (Simple Learner Model)
* Slumb is now the default MBase
o supports content reloading
o improved error reporting for authors
* New languages for the user interface:
o Français (French)
o Magyar (Hungarian)
o Čeština (Czech)
o Nederlands (Dutch)
* Extended user help system
* Improved overall performance and stabilty
* Now licencsed under the d-fsl (German Free Software License)

ActiveMath is a web-based platform for creation of Mathematics learning
environments. It supports the user by integrated semantic mathematical
input and output, by interactive exercises, by learner-modelling and
course generation. More information about ActiveMath can be found at
<http://www.activemath.org>.

ActiveMath 1.0 is licensed under the d-fsl (German Free Software
License). You can download and install the server application from
<http://www.activemath.org/downloads>.

An installation of ActiveMath 1.0 is currently running under the
preliminary address <http://am-preview.activemath.org>, and will soon
be installed on our regular demo server <http://demo.activemath.org>.

Enjoy!

Paul, Stefan and the whole ActiveMath group

ActiveMath 1.0 release candidate 1

2007-12-21T02:26:52Z

ActiveMath 1.0 release candidate 1
ActiveMath 1.0 is nearing completion and we are happy to announce the availability of the release candidate 1. It can be seen at: http://am-preview.activemath.org/.
ActiveMath is a web-based platform for creation Mathematics learning environments. It supports the user by integrated semantic mathematical input and output, by interactive exercises, by learner-modelling and course generation. More information about ActiveMath can be read at http://www.activemath.org.
ActiveMath 1.0 is a stabilization of the ActiveMath development and a polish of most of its features. We would be looking forward to all interested parties to ActiveMath to actually give a spin at testing this installation and tracking any issue encountered.
What has changed ? Lots since the LeActiveMath releases… among others the addition of languages (now ActiveMath speaks 9 languages), the support for several computer-algebra-systems, and the maturation of most components. Detailed changes can be seen on the jira detailed release notes.
If all goes well, the release 1.0 should happen early in January.
Enjoy!
Stefan, Paul, and the whole ActiveMath Group

smime.p7s (2K) Download Attachment

Announce: ActiveMath Authoring Online Tutorial | eds.activemath.org

2007-03-30T14:12:46Z

Announce: First ActiveMath Authoring Online Tutorial

We are starting a series of online tutorials for authoring ActiveMath content. The first series will be on April 20th and May 4th, 11th, 18th, and 25th. The goal of these tutorials is to introduce practically to the current concepts and tools to create content for the ActiveMath learning environment using a computer that authors are familiar with.

Participants are expected to attend using a computer with at least 1Gb of RAM and 1GHz of CPU as well as a high-speed internet connection (at least 512kb/s).

A familiarity with web-browsing and source-authoring is exptected. Participants’ visions for the usage of the ActiveMath server will be requested to be presented.

Registrations should be done to Paul Libbrecht paul@... providing details of their machines’ environment. They will be requested to install all needed components as a first shot and we shall verify this installation the week before the first tutorial using remote computer control methods. Participants will be expected to experiment the authoring activities in the weeks between the tutorials.

We look forward to your participation.

Note that the number of participants is limited to 8.

paul (and the other trainers)

(this announce is on the web and more details will be posted there).

ActiveMath Authoring and Discovery Tutorial Jan 30th - Feb 1st 2007

2006-12-19T20:57:52Z

Dear ActiveMath interested,

our group will be organizing an ActiveMath authoring and discovery
tutorial on January 30th and Feb 1st 2007 at DFKI in Saarbrücken, Germany.

This workshop will feature a strong authoring tutorial but will be based
on an active dialogue: expectations of each participants about the
ActiveMath learning environment will be discussed and approaches to
realize them in ActiveMath will be introduced and realized concretely.

It will provide participants with strong insights on the
usage and capabilities of ActiveMath as well as the authoring tools that
exist for it.

It is aimed at persons comfortable with the usage of web-browsers and
text-editors.

The workshop will start at 10:00AM on January 30th (with preliminary
set-ups) and end on Feb 1st at 15:00 after a presentation of the results.

It will be lead by members of the ActiveMath group with an
active involvement of each participant. The participation is free of
charge, in exchange, we expect this participation to be logged and
analyzed for evaluation purposes.

The participants are expected to:

1. send us a mail of pre-registration until Jan 10th indicating their
names and relevant capabilities (please include experience and
names of relevant XML, text, html, TeX, Wikis, word-processing,
RDF, concept-mapping, adaptive hypermedia, LMS, CMS tools)
2. send us until Jan 15th about a page describing the content they
wish be realized in ActiveMath, with the most amount of references
and URLs, along with pointers to two e-learning environment demoes
they find good.
3. participate to the tutorial actively with concrete realizations
and critiques of others' projects
4. bring a laptop which is capable of running ActiveMath server (min.
1GHz, and 512 Mb of RAM) with wifi card, having Mozilla 1.5 or
later and a java development environment version 1.4 or later
pre-installed

Linux desktops, with KDE environment, can be arranged to accomodate for
expectation 4).

Registration will only be accepted after the expectations 1) and 2) are
met. Doing this enables us to make sure that the system is ready to
accomodate such wishes. It has been our training experience that caring
for such in advance guarantees that small glitches are ironned out,
permitting, e.g., a new file-type to be properly served.

More details about the tutorial will be posted on:
http://eds.activemath.org/?q=en/AuthTrain-Jan2007
Questions can be addressed to Paul Libbrecht, by email at
paul@....
Please contact our secretaries, Irmtraud Stein or Astrid Thönes, for
help in finding accomodations in Saabruecken: stein@... or
Sek-DMAS@....

paul

smime.p7s (4K) Download Attachment

announcing eds.activemath.org

2006-10-17T07:08:35Z

Hello,

I wanted to drop a line about the ongoing construction of
http://eds.activemath.org
meant to become the online community of ActiveMath authors, users, and
other enjoyers.

As a wiki, this content-management-system is open for everyone to
register and may be used with a plain-text input syntax (called textile).
This system does a bit more though and allows the input for complete
HTML if wished, or using a visual html editor, it supports various forms
and provides forum and chat communication channels.
Many actions will be, in the future, doable by mail even, thus making it
possible to have the comfort of a (well archived) mailing-list.
By doing this, I hope to bring to a more visible space the very many
fruitful exchanges I had with many authors and lead them faster to
tangible results.
Get there, have a look, play around, ask questions, make blog or a
profile...

paul

smime.p7s (4K) Download Attachment

Re: AM-use workshop on usages of ActiveMath

2006-04-05T05:50:22Z

Unfortunately, we shall not make this workshop happen this year.
Better chances for next year.

paul

Paul Libbrecht wrote:

> The ActiveMath Group is currently planning a workshop to discuss the
> many facets of using the ActiveMath environment.
> It would happen in Saarbrücken either the 6-7th of July or 24-25th of
> August of this year.
>
> Would you consider taking part ?
> Would you present a vision or experience ?
> Which dates would suit you ?
>
> thank you in advance
>
> Paul Libbrecht, George Goguadze, and the ActiveMath Group
>

Re: AM-use workshop on usages of ActiveMath

2006-03-08T00:11:20Z

Paul Libbrecht wrote:
> It would happen in Saarbrücken either the 6-7th of July or 24-25th of
> August of this year.

am 6-7 Juli soll auch das LeAM Projektmeeting stattfinden (wurde so in
Malaga beschlossen), oder?

Carsten

--
Carsten Ullrich
Deutsches Forschungszentrum für Künstliche Intelligenz/
German Research Center for Artificial Intelligence
Stuhlsatzenhausweg 3
D-66123 Saarbrücken
Germany
Tel: +49-(0)681-302-5370
Fax: +49-(0)681-302-2235
http://www.activemath.org/~cullrich
vcard: http://www.activemath.org/~cullrich/cullrich.vcf