Chronological Thread 

< Chronological >      < Thread >

---

On 6/08/2004, at 11:28 PM, Alan Garny wrote:

> Hi Matt,
>
> Looking at it from a pure end-user / programmer point of view, if that
> was
> to be an option, then that could potentially result in some "errors"
> to be
> reported to the end-user.
>
> Say that a CellML compliant software (ours for instance) knows,
> internally,
> about all those standard units. In this case, to open a CellML 1.1
> file that
> imports a set of standard units would report some "errors" to the
> user. At
> least, our software would, since it would see those new standard units
> as
> being a redefinition (valid, I grant you that) of the "internal"
> standard
> units. To avoid those errors (and therefore be able to use the CellML
> file),
> the end-user would have to remove the import...

We will from time to time want to consider changes to the specification
that may break software that are compatible with earlier versions. The
version of a model is clearly indicated in the namespace defined in the
model element, so differentiating them will not be a problem for
software. Keeping CellML API implementations backwards compatible is
certainly something we take seriously, so that upgrading models to new
structures always remains possible for application developers.

>
> Now, from a programmer point of view, one could report any valid
> redefinition of an internally known standard unit as a warning, and
> not an
> error... or maybe simply ignore it...
>
> I can't recall whether the specifications mention anything about
> redefining
> standard units (and I unfortunately don't have time to go through them
> right
> now), but if not maybe something along the lines of what I have just
> said
> might be added.

The 1.0 and 1.1 specification says that a units declaration cannot
redefine any of the SI units which I have talking about. I agree that
some kind of protocol is necessary for warning when the SI units are
being redefined. At the moment the syntax for naming each of the SI
units in a units structure is illegal according to the 1.0 and 1.1
specifications where we cannot name a units structure using any of the
SI terms from the table that was given. It is certainly more
appropriate to consider these unit declarations in future versions of
CellML rather than modify 1.0 or 1.1.

regards
Matt

>
> Cheers, Alan.
>
>> -----Original Message-----
>> From:
>> cellml-discussion-bounces AT cellml.org
>> [mailto:cellml-discussion-bounces AT cellml.org]
>> On Behalf Of
>> Matt Halstead
>> Sent: 06 August 2004 11:49
>> To: For those interested in contributing to the development of CellML.
>> Subject: Re: [cellml-discussion] cellml units
>>
>> Hi Alan,
>>
>> the recommendation for units being defined through an import
>> would only affect any system that currently implements or
>> reads CellML 1.1. There is nothing stopping this being
>> optional to make things tidier for those cases where it may
>> make sense. I certainly wouldn't want to specify that a
>> CellML compliant reader would implicitly define such an import.
>>
>> Something I'd like to push as we bring ontologies and typing
>> to CellML is standardization of various axes such as units
>> such that specializations are always traceable back to a
>> common set that everyone agrees on. The import construct
>> references an imported resource by URI; it should be a simple
>> matter to assert that a model imports a known and accepted
>> set of units.
>>
>> The freshness aspect of imported units is a question that
>> applies more generally to the import of both models and
>> units. I have a couple of concerns with the way import is
>> structured in CellML 1.1, I'll outline these in another email.
>>
>> cheers
>> Matt
>>
>>
>> On 6/08/2004, at 6:31 PM, Alan Garny wrote:
>>
>>> Dear Matt,
>>>
>>> Just a quick thought on your suggestion...
>>>
>>> I have yet to go through the CellML 1.1 specification. As
>> far as I can
>>> remember from CellML 1.0, each variable must have a unit
>> associated to
>>> it and, in our software, we re-inforce that rule by having all the
>>> standard units "hard coded". In other words, the user can use them
>>> straight away and there is no need to declare them in the
>> first place
>>> or, for instance, to import them (using CellML 1.1).
>>>
>>> Now, from your second example, I can see a reason why one
>> would do as
>>> you suggest. This said, any software that currently implements the
>>> standard units the way we do it will have to be amended, so
>> that the
>>> standard units don't appear as being re-declared. Another
>> concern is
>>> that someone may not have your latest version of the
>> standard units.
>>> How do you deal with that?
>>>
>>> In conclusion, though I appreciate your original concern, I would
>>> personally stick to the way we have done it so far. Our
>> software being
>>> CellML 1.0 conformant, the user knows that s/he has access to these
>>> standard units, no matter what, and doesn't need to think about
>>> importing them, or make sure that s/he has the latest
>> standard units
>>> file. Should that list of standard units get new additions, I would
>>> expect that to be done through the release of a new version of the
>>> CellML specification.
>>>
>>> Cheers, Alan.
>>>
>>>> -----Original Message-----
>>>> From:
>>>> cellml-discussion-bounces AT cellml.org
>>>> [mailto:cellml-discussion-bounces AT cellml.org]
>>>> On Behalf Of Matt
>>>> Halstead
>>>> Sent: 06 August 2004 00:50
>>>> To: For those interested in contributing to the
>> development of CellML.
>>>> Subject: [cellml-discussion] cellml units
>>>>
>>>> Section 5.2.1 of the cellml 1.1 specification
>>>> (http://www.cellml.org/public/specification/20030930/
>>>> cellml_specification.html#tab_units_cellml_units_dictionary)
>>>> describes the dictionary of standard units. Since we now have
>>>> imports, is there any reason not to create a SI base unit library
>>>> that declares each of these units in this table as a
>> base_unit, and
>>>> which modellers should probably import by default?
>>>>
>>>> A case example where this is useful in application
>> development is an
>>>> editor application that forces a user to select a units for each
>>>> variable. If these units in table 2 were represented as a
>>>> library and
>>>> imported, then we can simplify the method for building units
>>>> available for a variable.
>>>>
>>>> Another use-case is providing a mechanism to curate this table of
>>>> units, we would be free to add new definitions, annotate them with
>>>> RDF, and then argue about them.
>>>>
>>>> just a thought.
>>>>
>>>> cheers
>>>> Matt
>>>>
>>>> _______________________________________________
>>>> cellml-discussion mailing list
>>>> cellml-discussion AT cellml.org
>>>> http://www.cellml.org/mailman/listinfo/cellml-discussion
>>>
>>> _______________________________________________
>>> cellml-discussion mailing list
>>> cellml-discussion AT cellml.org
>>> http://www.cellml.org/mailman/listinfo/cellml-discussion
>>
>> _______________________________________________
>> cellml-discussion mailing list
>> cellml-discussion AT cellml.org
>> http://www.cellml.org/mailman/listinfo/cellml-discussion
>
> _______________________________________________
> cellml-discussion mailing list
> cellml-discussion AT cellml.org
> http://www.cellml.org/mailman/listinfo/cellml-discussion

---

• [cellml-discussion] cellml units, Alan Garny, 08/06/2004
  • [cellml-discussion] cellml units, Matt Halstead, 08/06/2004
    • [cellml-discussion] cellml units, Alan Garny, 08/06/2004
      • [cellml-discussion] cellml units, Matt Halstead, 08/07/2004
        • [cellml-discussion] cellml units, Alan Garny, 08/07/2004
  • <Possible follow-up(s)>
  • [cellml-discussion] cellml units, Warren Hedley, 08/07/2004

---

Archive powered by MHonArc 2.6.18.