Loading ...
Sorry, an error occurred while loading the content.

Re: [RSS2-Support] Re: Re: pubDate time zone?

Expand Messages
  • Bill Kearney
    ... Uh, I ve made the strongest comments about it and tools like Radio /still/ get it wrong. ... No, that s completely false. If you know where you are then
    Message 1 of 34 , Mar 5, 2003
    • 0 Attachment
      > You're losing information about the origin timezone. And this can be a big
      > problem. In this thread we have more than once spoken about the effect that
      > daylight saving can make on time.

      Uh, I've made the strongest comments about it and tools like Radio /still/ get
      it wrong.

      > Various timezones change from daylight savings to normal time at different
      > times. If these overlap and you therefore lose information because you're
      > converting everything to GMT you're no longer able to determine the right
      > time because you don't know in which timezone the message was generated in.

      No, that's completely false. If you know where you are then you know the offset
      from GMT. EVERY operating system supports providing this information. So
      regardless of what machine or programming environment you use it's possible to
      ask the OS what timezone to use. This makes it a snap to do any and all
      manipulations of timestamps with absolutely no reliability issues whatsoever.
      You're just wrong, plain and simple.

      As Morten points out, when you try using non-numeric timezone data it's prone to
      a number of errors.

      > I consider this an important loss of information which more or less makes
      > the conversion to GMT useless. What do you think? Did I miss anything?

      Yes, the much bigger picture. This thread is a waste of time, you're more than
      welcome to cling to your belief but it's just not correct.

      -Bill Kearney
    • petite_lapin_blue
      ... much... ... Good. Any interesting papers or references you would like to share? Or better yet: concrete solutions. ... Ok. Not worth discussing such
      Message 34 of 34 , Mar 9, 2003
      • 0 Attachment
        --- In RSS2-Support@yahoogroups.com, "Bill Kearney" <ml_yahoo@i...>
        > > <ot>
        > > Regarding geographical location... a time zone doesn't help
        > > but combined with the host IP address... it helps.
        > > </ot>
        > Now you're in an area where I'm extremely familiar.

        Good. Any interesting papers or references you would like to share?
        Or better yet: concrete solutions.

        > And you're wrong again.

        Ok. Not worth discussing such dogmatic statement. Plus, what is this
        fixation with "right and wrong"? Why getting all religious all of the
        sudden? This is just software after all.

        > Geographic data from IP address is nearly useless. We track it on
        Syndic8 and
        > it's wrong FAR more often than it's right. Mainly because the
        tools used to do
        > IP lookups reference netblock or whois information. Most of the
        time this info
        > is not "where" the feed considers itself to reside.

        That's fine. No point in running in circle trying to find the
        unconditional truth. Approximation is just swell considering that
        there is no other alternative at the moment. But as always, I'm most
        likely wrong about that also.

        > Look, I understand the point you're trying to make.

        Compassion at last.

        > But you're just plain wrong.

        No arguing about that.

        > If you want to indicate location information, the Dublin Core
        format for
        > coverage and DCMI point allow you to express it exactly.

        Perhaps. But this requires an human being somewhere taking the time
        to generate such information in the first place. I'm most likely
        utterly wrong about that, but from the outset IP addresses and
        time-zones seems in more widespread use than semantically meaningful
        Dublin Core scribbling.

        > Trying to 'fake it' with a timezone is going to be wrong way more
        than it's ever right.

        Well... waiting for Godot and the "semantic web" sounds like an
        attractive alternative of course. But there is hope: as this is "an
        area where [you are] extremely familiar". Any pointers to concrete
        alternatives much appreciated therefore. Thanks.

        > Consider what happens when you run your laptop from a different
        timezone. Or if
        > the server is running (or serving) the feed from a timezone
        different than the
        > authoring timezone. There's no way to reliably extract positioning
        > here. It's not like the people that created these specs didn't
        take this into
        > account.

        This is fine. Some zealots are obsessively looking for the absolute
        truth. Others are just trying to get by with what they have today.


      Your message has been successfully submitted and would be delivered to recipients shortly.