Thread: [OpenFormula-discuss] FYI: Formulas not specified by Microsoft XML, either

FYI: I've learned that Microsoft's Brian Jones takes some shots at
OpenDocument because it doesn't specify spreadsheet formulas in detail; see:
 http://blogs.msdn.com/brian_jones/archive/2005/10/04/477127.aspx

Those in glass houses should not throw stones.  As far as I know,
Microsoft's XML format is just as unspecified as OpenDocument.
In both cases, formulas are a "string" with some examples.

Microsoft's XML has EXACTLY the same limitations as
OpenDocument, in terms of its formula specifications.
Thus Jones' claims that they uniquely support "full fidelity formats"
are complete nonsense.  Microsoft's formats do manage
to exchange data between two identical versions of Microsoft Office.
So what? All the spreadsheets can do that, there's nothing
interesting about that.  What is DESIRED is
a standard, public format, implementable by ANYONE (without
restriction on their licenses, etc.) so that ALL spreadsheets can
exchange data.  Microsoft's XML format does not even begin to do this.
Microsoft's licensing approach makes it illegal to have open competition,
or for customers to access their data in arbitrary ways. The failure
of Microsoft's XML formats to specify the formulas means
that their approach is not even TECHNICALLY an improvement
over OpenDocument in this area.

"Whatever Excel appears to do today" is NOT a specification.
It changes with each release.  And are the bugs in Excel REALLY
required? For example, Excel still does not correctly implement AND and OR
in an array formula, making these two fundamental functions useless
in an important circumstance.  A _real_ specification of functions would
make it clear what is intended... vs. what is a bug.

Thanks; I thought everyone on this list ought to know.

--- David A. Wheeler

David,

 I'm confused.

1. In the page on Brian Jones blog  that you link to (and is based on a
concern expressed by Tim Bray of Sun) there is this statement, near the end:

  "The Microsoft Office Open XML formats are specifically designed as an XML
representation of our full file formats. Everything you can do in our
default format is represented as XML. Our formats are primarily designed
around viewing, editing and integrating the files with data, formulas, and
other application behavior." 

2. Later, in a comment response to information about your work on
OpenFormula,
<http://blogs.msdn.com/brian_jones/archive/2005/10/04/477127.aspx#477463>
, Brian makes a further observation:

  "I'm actually curious about what peoples thoughts were around formulas.
One could make an argument that using strings for formulas is the right way
to go, but in order to have a shared document format I'm assuming they still
need to have everyone agree on a single type of syntax for those strings.
Otherwise the formats aren't interoperable."

3. I can't tell from these comments whether the Office "12" Open XML format
for Excel will use strings or XML structures for formulas.  Until I saw your
posting, I had the (unjustified) thought that they intended to use XML
structures, except I am sure the possible bloating of the resulting files
and parsing time would be a critical consideration.

4. In the provisional schemas that have been published, I notice that the
simple type, ST_Formula, is based on xs:string but I am not comfortable
enough with XML Schema and the sketchy form of the Office "12" preliminary
schema to know the import of that.  (For example, is this an entity name or
is it an attribute name, that kind of thing.  I haven't dug into their
schema enough to understand how it all holds together.)

5. If formulas are in strings, the XML schema cannot define the syntax.
That doesn't mean the syntax and semantics are not defined, it means they
aren't to be found in the schema.

6. Maybe we should ask Brian for clarification before we speculate so much?

Also, where are the examples you are looking at?  Are they using the current
Office 2003 Excel XML or are they done with the proposed Office 12 format,
which is supposed to be a complete round-trip fidelity-preserving solution
compared to the initial steps taken in Office 2003?

 - Dennis

-----Original Message-----
Thread
http://sourceforge.net/mailarchive/forum.php?thread_id=8911498&forum_id=4663
2 by David A. Wheeler
Sent: Monday, November 07, 2005 10:40
To: ope...@li...
Subject: [OpenFormula-discuss] FYI: Formulas not specified by Microsoft XML,
either

FYI: I've learned that Microsoft's Brian Jones takes some shots at
OpenDocument because it doesn't specify spreadsheet formulas in detail; see:
 http://blogs.msdn.com/brian_jones/archive/2005/10/04/477127.aspx

Those in glass houses should not throw stones.  As far as I know,
Microsoft's XML format is just as unspecified as OpenDocument.
In both cases, formulas are a "string" with some examples.

[ ... ]

A short after-thought:

1. However Microsoft matures the specifications for Office "12" Open XML and
details of the elements and attributes, I would think that has nothing to do
with the value of producing an OpenFormula specification that is widely
acceptable in conjunction with OpenDocument.

2. Our task would seem to be unchanged and maybe we should not allow
ourselves to be distracted by suspected Microsoft maneuvers.

 - Dennis

-----Original Message-----
http://sourceforge.net/mailarchive/message.php?msg_id=13781771 On Behalf Of
Dennis E. Hamilton
on thread
http://sourceforge.net/mailarchive/forum.php?thread_id=8911498&forum_id=4663
2
Sent: Monday, November 07, 2005 13:06
To: ope...@li...
Subject: RE: [OpenFormula-discuss] FYI: Formulas not specified by Microsoft
XML, either

David,

 I'm confused.

1. In the page on Brian Jones blog  that you link to (and is based on a
concern expressed by Tim Bray of Sun) there is this statement, near the end:

  "The Microsoft Office Open XML formats are specifically designed as an XML
representation of our full file formats. Everything you can do in our
default format is represented as XML. Our formats are primarily designed
around viewing, editing and integrating the files with data, formulas, and
other application behavior." 

2. Later, in a comment response to information about your work on
OpenFormula,
<http://blogs.msdn.com/brian_jones/archive/2005/10/04/477127.aspx#477463>
, Brian makes a further observation:

  "I'm actually curious about what peoples thoughts were around formulas.
One could make an argument that using strings for formulas is the right way
to go, but in order to have a shared document format I'm assuming they still
need to have everyone agree on a single type of syntax for those strings.
Otherwise the formats aren't interoperable."

3. I can't tell from these comments whether the Office "12" Open XML format
for Excel will use strings or XML structures for formulas.  Until I saw your
posting, I had the (unjustified) thought that they intended to use XML
structures, except I am sure the possible bloating of the resulting files
and parsing time would be a critical consideration.

4. In the provisional schemas that have been published, I notice that the
simple type, ST_Formula, is based on xs:string but I am not comfortable
enough with XML Schema and the sketchy form of the Office "12" preliminary
schema to know the import of that.  (For example, is this an entity name or
is it an attribute name, that kind of thing.  I haven't dug into their
schema enough to understand how it all holds together.)

5. If formulas are in strings, the XML schema cannot define the syntax.
That doesn't mean the syntax and semantics are not defined, it means they
aren't to be found in the schema.

6. Maybe we should ask Brian for clarification before we speculate so much?

Also, where are the examples you are looking at?  Are they using the current
Office 2003 Excel XML or are they done with the proposed Office 12 format,
which is supposed to be a complete round-trip fidelity-preserving solution
compared to the initial steps taken in Office 2003?

 - Dennis

-----Original Message-----
Thread
http://sourceforge.net/mailarchive/forum.php?thread_id=8911498&forum_id=4663
2 by David A. Wheeler
Sent: Monday, November 07, 2005 10:40
To: ope...@li...
Subject: [OpenFormula-discuss] FYI: Formulas not specified by Microsoft XML,
either

FYI: I've learned that Microsoft's Brian Jones takes some shots at
OpenDocument because it doesn't specify spreadsheet formulas in detail; see:
 http://blogs.msdn.com/brian_jones/archive/2005/10/04/477127.aspx

Those in glass houses should not throw stones.  As far as I know,
Microsoft's XML format is just as unspecified as OpenDocument.
In both cases, formulas are a "string" with some examples.

[ ... ]




-------------------------------------------------------
SF.Net email is sponsored by:
Tame your development challenges with Apache's Geronimo App Server. Download
it for free - -and be entered to win a 42" plasma tv or your very own
Sony(tm)PSP.  Click here to play: http://sourceforge.net/geronimo.php
_______________________________________________
Openformula-discuss mailing list
Ope...@li...
https://lists.sourceforge.net/lists/listinfo/openformula-discuss

Dennis E. Hamilton scripsit:

> 3. I can't tell from these comments whether the Office "12" Open XML format
> for Excel will use strings or XML structures for formulas.  

I think it's clear that they will be strings.  "Represented as XML"
just means that everything is, one way or another, embedded inside XML
elements and attributes.  It says nothing about how the attribute values
and element content may or may not be structured.

> 4. In the provisional schemas that have been published, I notice that the
> simple type, ST_Formula, is based on xs:string but I am not comfortable
> enough with XML Schema and the sketchy form of the Office "12" preliminary
> schema to know the import of that.  (For example, is this an entity name or
> is it an attribute name, that kind of thing.  I haven't dug into their
> schema enough to understand how it all holds together.)

Simple types are neither entity names nor attribute names; rather, they
are schema-internal labels for particular kinds of character content or
attribute values.  For example, the simple type xs:integer describes
unlimited-precision integers, and indicates that the corresponding
character content or attribute value consists solely of digits with a
possible leading sign and possible leading or trailing spaces.

In this case, ST_Formula is a subtype of xs:string, meaning that there
are no particular interpretive semantics within the framework of XML
Schema that can be placed on it.  It is neither a number nor a boolean
nor a language code nor a URI nor ....

Depending on the exact definition of ST_Formula, it may be limited in
size or constrained to match a regular expression, though I doubt it
in this particular case.  (I haven't looked at the schemas, as I don't
want to risk having knowledge of patented material, which exposes one
to triple damages.)

> 5. If formulas are in strings, the XML schema cannot define the syntax.
> That doesn't mean the syntax and semantics are not defined, it means they
> aren't to be found in the schema.

Correct.

-- 
A rose by any other name                                John Cowan
may smell as sweet,                                     http://www.ccil.org/~cowan
but if you called it an onion                           http://www.reutershealth.com
you'd get cooks very confused.          --RMS           jc...@re...

 Thanks John, that's very helpful.

I don't think the XML schemas (or any schemas) are patentable subject
matter.  The Microsoft royalty-free patent license applies to Microsoft
necessary/essential claims for software that processes data according to
those schemas.

However, the preview materials are under an EULA and that does appear to
limit use and public discussion of them.  That's why I haven't done more
than glance them over, notice how thin they seem to be so far, and move on.
The schema is not enough to know what to do and what the intended
interpretation of the XML is, of course, or the OASIS OpenDocument
specification would be about 40 pages instead of over 700 [;<).  

I think there may be a Microsoft newsgroup somewhere to ask further
questions, but I haven't sought it out. There is a blog where new features
for Office "12" Excel are being discussed:
http://blogs.msdn.com/excel/default.aspx and their are a few entries about
formula-related improvements too.

 - Dennis

-----Original Message-----
http://sourceforge.net/mailarchive/message.php?msg_id=13790750 On Behalf Of
John.Cowan
on thread
http://sourceforge.net/mailarchive/forum.php?thread_id=8911498&forum_id=4663
2
Sent: Tuesday, November 08, 2005 07:15
To: Dennis E. Hamilton
Cc: ope...@li...
Subject: Re: [OpenFormula-discuss] FYI: Formulas not specified by Microsoft
XML, either

[ ... useful observations on how ST_Formula is used as part of the Office
"12" Excel schema ...]

Depending on the exact definition of ST_Formula, it may be limited in size
or constrained to match a regular expression, though I doubt it in this
particular case.  (I haven't looked at the schemas, as I don't want to risk
having knowledge of patented material, which exposes one to triple damages.)

> 5. If formulas are in strings, the XML schema cannot define the syntax.
> That doesn't mean the syntax and semantics are not defined, it means 
> they aren't to be found in the schema.

Correct.

[ ... ]