Re: Time zones in base data schema

Hi Massimo,

Thanks for the reply, I do have some follow-up suggestions below.

Massimo Marchiori wrote:

> Dear Chris, thanks for your feedback on the P3P specification.
>
> > 1. You specify only "text" as the type for the date.timezone field,
> > whereas this could be specified more precisely (like country code is).
> This is not the intent of the specification, and I'm sorry if you were
> mislead to assume there's no type constraint. After the date. table
> there is the precisation
> > All the fields in the date. type must be in the same format as those
> > in the most informative profile of the time standard ISO8601.
> (btw, because of your feedback I've changed "are" from the previous
> version of the spec into "must", to make this clearer).
> So, in essence it's a matter of taste whether to introduce a new type
> (like for Country), or using a more general type and specify a
> constraint on it. The current choice reflects the fact that eight
> types in date. now are compactly referenced to ISO8601, while making
> the type explicit would mean you've to create eight subtypes
> corresponding to ISO8601. Anyway, no semantic difference.

Understood -- I had overlooked that constraint.  The ISO8601 specification for
timezone -- basically as an offset from UTC -- does indeed work adequately for
representing the current timezone offset of an individual date/time, as you're
representing here with the date. object.

But see below for my feedback on a more general representation of a user's
timezone.

> > 2. Lots of sites request the user's timezone, yet you don't have that
> > represented as part of your user. schema, other than perhaps
> > accidentally as a field in the birthday element.
> The timezone information can be inferred from user.home and/or
> user.business (from various fields, either city, or stateprov, or
> postalcode etc). In general, we have tried to keep the base data set
> to a minimum, avoiding fields that are not of common use (after all,
> new data schemas can be created too). However, if you feel that the
> ability to *directly* request a timezone can be of significant impact
> for many of today's sites, we'd appreciate some examples and use cases that
> could justify the introduction of a new timezone field.

I certainly understand your desire to avoid redundancy and to achieve
compactness, but I do strongly feel that user timezone is commonly requested
enough, and it's inference from other fields non-trivial enough, that it should
be added as a new field in the schema (probably as a new field in either
contact. or postal.).

Many sites ask explicitly for timezone when registering or setting preferences;
some examples include "My" pages (http://my.netscape.com/,
http://my.yahoo.com/, http://my.cnn.com), calendaring pages
(http://calendar.aol.com, http://calendar.yahoo.com), and various other sites
(e.g., tvguide.com, passport.com).  Certainly for these sites and all the
others than request timezone, it'd facilitate adoption of and conformance to
P3P if timezone were added to the base data schema.

Unfortunately, the ISO8601 representation is insufficient to represent what
timezone the user is in, as the instantaneous UTC offset is different at
different times during the year (due to daylight savings time observance,
etc.), according to various rules in different localities.  The best (only)
standard I've seen for representing timezone explicitly is the tz database,
described at http://www.twinsun.com/tz/tz-link.htm.  It is the timezone
representation standard used by Java, various Unix systems, and many websites.
Thus the legal timezone names defined by this database is what I think should
be used as the P3P type for the timezone field.

Thanks again for listening and hope to see this added to P3P!

Chris

>
>
> Again, thanks for your precious feedback, and feel free to reply with
> further questions and clarifications, until every doubt/problem you
> have is completely solved.
>
> All the best,
> -Massimo
>
> /---------------------------------------------------------------\
> | Massimo Marchiori                    Room NE43-350            |
> | The World Wide Web Consortium (W3C)  545 Technology Square    |
> | MIT Laboratory for Computer Science  Cambridge, MA 02139, USA |
> | WWW: http://w3.org/People/Massimo    Phone: +1 (617) 253 2442 |
> | Email: massimo@w3.org                Fax:   +1 (617) 258 5999 |
> \---------------------------------------------------------------/