W3C home > Mailing lists > Public > www-multimodal@w3.org > December 2012

Re: [emo] floating point number in <trace> element

From: Gerhard Fobe <gerhard.fobe@s2009.tu-chemnitz.de>
Date: Wed, 19 Dec 2012 16:21:35 +0100
Message-ID: <50D1DB7F.8010909@s2009.tu-chemnitz.de>
To: www-multimodal@w3.org
CC: Felix.Burkhardt@telekom.de
Hello Felix, all,

sorry for my late reply. Many to do at the end of the year.

However, I agree with your changes. A single blank is a good version. In
my mind the representation of floting point numbers only with \d+\.\d+
has a technical term, but I can not find it. The phrase is okay, perhaps

The value of this attribute MUST be a positive unsigned floating point
number without exponent (formatted e.g. "10" or "10.5") followed by
optional blank followed by "Hz"

makes it more clear, that both exponent and algebraic sign should be
banned from notation to simplify it.

Kind regards

Gerhard Fobe


------ Original Message ------
Date:    Fri, 7 Dec 2012 13:40:58 +0100
From:    Felix.Burkhardt@telekom.de
To:      www-multimodal@w3.org
Subject: Re: [emo] floating point number in <trace> element

> Hi Gerhard, all
> Sorry this took really long, we were busy with the Implementation reports.
> 
> With respect to your comments from 18th of August, we'd accept your suggestions in all three issues and will change the spec accordingly:
> 
> Re 1) - clarify that only blank is optionally allowed before "Hz"
> 
> Re 2)  - only allow for \d.\d+ format
> To replace the sentence in [1]
> The value of this attribute MUST be a positive floating point number followed by optional whitespace followed by "Hz".
> By
> The value of this attribute MUST be a positive floating point number, formatted e.g. "10" or "10.5", followed by optional blank followed by "Hz".
> Any better ideas to express this restriction?
> 
> Re 3)
> Replace
> \d+(\.\d*)?\s*Hz
> By
> \d+(\.\d+)? ?Hz
> Not as suggested, but also replaceing "\s" with " "
> 
> Please respond until the 14th December, if this is acceptable to you.
> Cheers,
> Felix
> 
> [1] http://www.w3.org/TR/2012/CR-emotionml-20120510/#s2.5.2
> 
> 
> 
> -----Urspr√ľngliche Nachricht-----
> Von: Gerhard Fobe [mailto:gerhard.fobe@s2009.tu-chemnitz.de]
> Gesendet: Samstag, 18. August 2012 21:44
> An: www-multimodal@w3.org
> Betreff: [emo] floating point number in <trace> element
> 
> Hello,
> 
> I found some lack of clarity in context of the freq attribute of the trace element. In EmotionML Candidate Recommendation of 10 May 2012 you can read following sentence: 'The value of this attribute MUST be a positive floating point number followed by optional whitespace followed by "Hz".' All in all I interpret here that I have to use e.g. "42.0 Hz".
> 
> 1. Interpretation of whitespace
> -------------------------------
> In my opinion it is not clear what whitespace means in detail. My first interpretation was that only one whitespace character, e.g. space or tab, is allowed. But in the emotionml-fragments.xsd you test of many whitespace characters. Perhaps it will be better to clarify it more?
> What is right?
> 
> 2. Different design of floating points
> --------------------------------------
> It's clear that floating points like 42 or 42.0 are allowed, but what is with other designs of positive floating point numbers such as +1234.456 or +1.2344e56 ? Should this be allowed too? In my opinion in the moment it is, but it makes it needless complicate.
> 
> 3. Wrong RegEx in emotionml-fragments.xsd
> -----------------------------------------
> In the XSD this area is validated via the pattern \d+(\.\d*)?\s*Hz In my opinion this is wrong, because of \d* you can set as many digits you want. "42." will be a valid value too but I think this is not a valid floating point number.
> 
> In the XSD I prefer the RegEx \d+(\.\d+)?\s?Hz
> 
> 
> Kind regards
> 
> Gerhard Fobe
> Chemnitz University of Technology, Germany
> 
Received on Wednesday, 19 December 2012 15:22:09 GMT

This archive was generated by hypermail 2.2.0+W3C-0.50 : Wednesday, 19 December 2012 15:22:09 GMT