Re: created, issued, modified [Re: comments on Atom 0.2]

["Bill Kearney" <wkearney99@xxxxxxxxxxx>]

> > Pepys' Diary provides both a good example and a bad example simultaneously.

+1.  This isn't about archival publishing so drop the example.

> I only brought it up to better present the date issue.  If you're
> claiming it's troublesome because of dates, take a more simple and
> common case:
> I go to the mountains for a week but bring my laptop.  I write one entry
> each day and save them locally, and when I get home I upload them all
> simultaneously.
> There's definitely a difference between the timestamp that represents
> when I wrote those entries and the timestamp for when they go up on my
> website.  I think we all agree on that, and the discussion here is which
> dc attributes correspond to which of those dates.

The question is what do you expect the consuming application to DO with the
timestamps?  This is a rats nest many an effort gets tangled within.  Let's look
at what the point of interchanged data is for in the first place.

If you're pushing this data back/forth between systems (a local offline tool and
it's hosted delivery mechanism) then you've got quite a range of timestamps to
consider tracking.  The local creation and modification times (to say nothing of
a timestamp of when your started it and when you actually saved it; ie File->New
and then File->Save timestamps).  Then you've got a timestamp for when you
uploaded/sync'd it with the publishing environment.  Now, at that point you
might also have need to timestamps for when it was /intended/ to be shown as
published, created or modified.  These may also be /different/ than the actual
timestamps.  As in, I could write an article in May but not have it published
until July, maybe even in a different year.  Then there's the whole idea of it's
shelf-life for expiration.

> I used Pepys' Diary as an example just because the difference in these
> two dates is easier to see.  It's also a good example for the optional
> time zone issues because I doubt he included a time zone when dating his
> entries.

What has timezone got to do with it?  You're not thinking you can depend on
/that/ for anything like geographic info are you?  If so, disabuse yourself of
that notion *right now*.

You're not incorrect that the range of the timestamps is an illustrative
example.  But it's what one might call and edge case.  And, quite often, those
edge cases turn into a quagmire in which any further actual discussion would
just be a waste of time.

> > In terms of pepysdiary.com, my understanding would be that issued would be
> > the timestamp for the moment at which pepysdiary.com published an excerpt
> > from the diary. Created would be used to indicate the original creation date
> > for the content being posted, 1660/08/07, for example.
>
> Hmm, really?
>
> Ok, so going back to what I wrote before:
> >  - created is the "real" create time;
> >  - modified is the "real" modified time;
> >  - issued is whatever time the person wants to say the post came from.
> Would it be correct to invert these definitions of "created" and "issued"?
>
> In that case, what we've been agreeing upon has been wrong, including
> the validator; it is *issued* that must have a timezone, and it is
> *issued* and modified that are redundant when a post has never been
> modified.

That given tools get timestamps *wrong* is not a reason to believe them.  Let's
be honest here, there's a lot of confusion of what a timestamp expresses and
most of the time it's borne out of developer ignorance (innocently of course).

> Here's what Dublin Core says:
>  created  Date of creation of the resource.
>  issued   Date of formal issuance (e.g., publication) of the resource.
> (http://dublincore.org/documents/dcmi-terms/)
>
> By my reading of those definitions, I think we had it right.  But as you
> wrote above, the most common interpretation is the one we should follow,
> and I'm certainly open to being set straight.

> I guess my confusion here is:
>  - Is the "resource" the entry itself?  Then Pepys' entries were created
>    hundreds of years ago.  Or the resource the web-accessible resource
>    as you published it, putting their create date within this year?
>  - Similarly, is the formal publication of the resource the actual date
>    the entries went up on the web?  Or is it the date Pepys decided to
>    tag on his entries back as he was writing them?

If we're talking about the republishing of historical or archival data it's an
entirely other set of issues.  Ones well outside the scope of a
weblog/syndication spec.

> Armed with just those definitions, I just asked my coworker here at
> LiveJournal what he thought the right interpretation was and he got the
> opposite interpretation (the one you presented) from me.

Which is?  Compare and contrast them, clearly, otherwise we continue to talk
past each other.

-Bill Kearney