* Alan Burlison <[EMAIL PROTECTED]> [2007-10-04 14:45]:
> The current OpenSolaris user registration screens collect timezone and 
> location information, and I'm considering what we should do with these 
> in the new auth webapp, as in their current form both have problems.
> 
> Timezone
> --------
> 
> Currently this is built from a hard-coded timezone list 
> http://cvs.opensolaris.org/source/xref/website/portal/trunk/src/java/org/opensolaris/base/util/TimeZone.java
> and then looking up information in the J2SE timezone class (also called 
> TimeZone) - see 
> http://cvs.opensolaris.org/source/xref/website/portal/trunk/src/java/org/opensolaris/tonic/action/user/ListInfoAction.java#40.
>  
>   I have no idea of the sanity of the resulting list, and I can see at 
> least one problem - the initialisation code is only run once, and it 
> uses the current date/time in its calculations of the GMT offset, so if 
> the appserver runs over a DST change, the offsets will change but the 
> list won't.  Of course, the appserver has to be restarted every night 
> anyway for other reasons, so I guess we just got lucky ;-)  The 
> aggregation of timezones by GMT offset in the current version also means 
> that we get less info that we otherwise might, for example Amsterdam, 
> Berlin, Bern, Rome, Stockholm and Vienna are all lumped together.
> 
> My suggestion is that we just use the standard list as provided by the 
> Java TimeZone.getAvailableIds() method 
> http://java.sun.com/javase/6/docs/api/java/util/TimeZone.html#getAvailableIDs()
>  
>   preprocessing the list to discard the Etc/ entries, which are all just 
> GMT offsets, and the SystemV timezones, which I didn't even realise existed.
 
  +1.

> Location
> --------
> 
> This is a freeform text field, and has a whole range of information in 
> it, from a full postal address, just a postal code, a geographic area, a 
> country, /dev/null (hilarious) and nothing at all.
> 
> My suggestion is that we replace this with an explicit country list. 
> Google likes http://schmidt.devlib.org/data/countries.txt, but I'm open 
> to other suggestions for a data source.

  +1, noting that the related page

        http://schmidt.devlib.org/data/geo-data.html

  suggests that we could spend weeks picking among standards.

  - Stephen

-- 
[EMAIL PROTECTED]  http://blogs.sun.com/sch/
_______________________________________________
website-discuss mailing list
[email protected]

Reply via email to