Wolfram Alpha seems to conflate timezone identifiers that refer to specific offsets and those which refer to a particular location/city from the TZ database.
For example, take query 1 and query 2. It appears to always parse PDT to America/Los_Angeles, but this is not correct. America/Los_Angeles is only on PDT for part of the year. PDT (by definition) should always be 7 hours behind GMT/UTC, but as you can see in these queries it varies depending on the entered timestamp.
Is there somewhere I can report this?
Looking at the definition, PDT is a zone (geographic region) - which is only in effect during part of the year. Consequently, I believe it is an error to calculate PDT times in that zone outside of the defined time period.
Thanks for the reply.
I would say PT is the geographic region that switches between PST and PDT, while PDT is a specific offset from UTC.
In fact, Wolfram|Alpha displays a widget in the results denoting the time offset from GMT, which is equivalently the winter timezone for the UK. GMT is not currently in effect there either (currently we're under BST), so I'm not sure your interpretation would be consistent with that aspect.
As I interpret it, regardless of whether they are geographical, PDT being defined relative to GMT means it should always be GMT-7. Even if we disagree about whether we should even bother calculating in PDT throughout the year, I would expect Wolfram|Alpha to display an exception rather than something confusing and/or wrong.
"I would say PT is the geographic region that switches between PST and PDT".
The actual definition has PDT as a zone to accommodate portions of PST that do/did/will not change during the year.
But I endorse the idea returning a value other than a Time. For example consider the return value when f[x]:=1/x is evaluated at x=0.