XForms Users Community Group Teleconference -- 23 Mar 2016

<ebruchez> I thnk I need the new new password

No, I just used the old one

I needed the room number for the first time

The room number is in the agenda

<ebruchez> which is the old one then?

I mean the same one as last week

Are you having problems Alain?

Alain?

Call back to normal time next week

[No link]

Steven: Summer time starts in Europe this weekend, so back to normal times.

Whitespace MIP

https://lists.w3.org/Archives/Public/public-xformsusers/2016Mar/0012.html

Steven: So writing this text raised some questions
... we can't have AVTs in binds if I remember right. So should it be a computed property?

Erik: It causes problems with ordering of calculations; so it should be static like type is.
... In an ideal world everything should be computable, but you have problems with spotting and avoiding loops

Steven: Take integers for instance, whitespace is not relevant
... how do we specify when whitespace should be transformed
... ?

Erik: If you called a service that supplied let us say an integer, could spaces be allowed?
... XML datamodel has lexical value and type value

Steven: Lexical space and value space

Erik: XForms allows control to take account of the value, dates being a good example.
... So I think inputs may well already ignore spaces for some types

Steven: So what do we say. The default is preserve, or type-dependent?

Erik: Good question...
... Boolean, integer, date/time all have representations that should have no whitespace.
... not sure it makes sense to say collapse on a number
... if a user types 1<space>2, not sure what happens. Leading and trailing whitespace there is an argument for.

Steven: Is "<space>1" invalid?

Philip: Schema allows leading and trailing spaces.

<alain> Hello! Room number??

Steven: So 'preserve' is OK for numbers, because "1<space>2" would not validate

Erik: So surrounding spaces seem to be OK, from a schema point of view, date works (checking now)

Steven: Sounds good to me.
... I'm assuming that the property doesn't inherit.

Erik: I agree; I don't think we need inheritance.

Steven: Last question, what to do with "trim collapse" in the example. Make it a single value? What should we call it then?

Erik: Normalise space in XPath strips leading and trailing, and collapses.

Steven: Maybe we should just call it normalise then.

Erik: We could but in Schema 'normalise' is called 'collapse'

Steven: I propose we call "trim-collapse" just "normalize" and be done, unless someone can come up with a better name.

Erik: Not sure of my preference at this moment
... I think normalize would be an acceptable choice.
... Two additions - 1) what whitespace means.

Steven: I didn't realise we didn't.
... we could either use the XML definition, or the Unicode definition/

http://www.fileformat.info/info/unicode/category/Zs/list.htm

http://www.fileformat.info/info/unicode/category/index.htm

<ebruchez> Character.isWhitespace(c) || isNonBreakingSpace(c) || isZeroWidthChar(c) || Character.isISOControl(c))

Erik: We use one of the Java APIs to define whitespace
... note we include the control characters.

Steven: I'll take a look to see if we can make a consistent story on this.

<scribe> ACTION: Steven to research what whitespace means [recorded in http://www.w3.org/2016/03/23-forms-minutes.html#action01]

<trackbot> Created ACTION-2059 - Research what whitespace means [on Steven Pemberton - due 2016-03-30].

Erik: We process all whitespace before calculations
... and if we write to a value that has a whitespace property, then the whitespace gets dealt with too.
... And 2) was saying WHEN processing happens, which we do just before calculate.

AOB

Steven: I'm going to start posting updates to the spec starting next week.

[ADJOURN]

- DRAFT -

XForms Users Community Group Teleconference

23 Mar 2016

Attendees

Contents

Call back to normal time next week

Whitespace MIP

AOB

Summary of Action Items

Summary of Resolutions

Scribe.perl diagnostic output