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

Re: [json] Re: IMPORTANT: Remove this line from json2.js before deployment.

Expand Messages
  • Andrea Giammarchi
    If you go in YUI 3 Configurator the Result page provides a script to copy and paste to include remotely the library.
    Message 1 of 9 , Dec 13, 2009
    View Source
    • 0 Attachment
      If you go in YUI 3 Configurator the Result page provides a script to copy
      and paste to include remotely the library.
      http://developer.yahoo.com/yui/3/configurator/

      If you go in Google Ajax libraries you will find external URIs to use third
      part hosts as trusted safe and secure host with updated libraries.
      http://code.google.com/apis/ajaxlibs/documentation/#AjaxLibraries

      If you are using dojo library you probably know about AOL:
      http://dev.aol.com/dojo

      What am I saying is that we are not everybody under https and we trust, for
      whatever reason, some external domain.
      Not everybody could have followed the "alert story", I can already imagine
      developers called 6am in the morning about an alert in the website that does
      not use alerts at all.
      These devs could have quickly solved the problem nullifying the alert
      without caring about why the alert was there and, in the worst case
      scenario, blaming you to have forgot an alert inside your library and
      feeling cool to have solved an unexpected alert problem forever (so try with
      prompt or confirm ...)

      Since the message as is could sound more like you were testing something and
      you forgot an alert, I would rather change the alert message with a link
      that points WHY there is an alert.

      I totally agree with you and it could often be about developers laziness (in
      YUI case they did not use the php loader, etc etc) but at the same time:

      1 - every website could benefit about common external resources thanks to
      distributed cache for common libaries
      2 - this message is not perfectly clear since somebody, YUI! itself, is
      suggesting external resources while you, a Yahoo! engineer, are saying that
      this is so bad that anybody should avoid this technique

      Do you see what I mean?

      Regards


      On Sun, Dec 13, 2009 at 6:38 PM, Douglas Crockford <douglas@...>wrote:

      >
      >
      > --- In json@yahoogroups.com <json%40yahoogroups.com>, Andrea Giammarchi
      > <andrea.giammarchi@...> wrote:
      > >
      > > Douglas, I think this move was brilliant, but as I have twitted, I
      > > wonder how many devs wrote a:
      > >
      > > window.alert = function(){};
      > >
      > > before including external resources, rather than get the real/original
      > > message
      >
      > The point I was making was that if you care about reliability, security,
      > and performance, then you shouldn't load scripts directly from third party
      > servers. Are you suggesting that it is ok if you stub out alert first?
      >
      >
      >


      [Non-text portions of this message have been removed]
    • Andrea Giammarchi
      Just to underline another thing: ... reliability ... they are including the de facto official JSON library for JavaScript security ... they trust your
      Message 2 of 9 , Dec 13, 2009
      View Source
      • 0 Attachment
        Just to underline another thing:

        On Sun, Dec 13, 2009 at 6:38 PM, Douglas Crockford <douglas@...>wrote:

        >
        > The point I was making was that if you care about reliability, security,
        > and performance
        >
        >
        reliability ... they are including the de facto official JSON library for
        JavaScript
        security ... they trust your implementation and they trust the fact you keep
        updating it
        performances ... they are using a potentially "common used external
        resource" so if the browser cached already that version performances will be
        better for every website that includes it plus they are saving bandwidth.

        As you can see somebody could think that your points ARE the reason they
        included JSON via the direct source, rather than their local copy
        potentially non updated and served even if almost every browser has stored
        somewhere exactly the same library.

        So, finally, I would think about a proper specific server or an official
        repository Github style so that people than use the raw minified and gzipped
        version with the 304 response, but if you think nobody should ever include
        external scripts, you should tell us why we all have YUI configurator
        scripts, google adsense/analytic files, etc etc.

        Regards


        [Non-text portions of this message have been removed]
      • Leen Besselink
        ... Hi Andrea, I think you are forgetting that the (current) json.org website is probably just a shared-hosting account. So that probably means it s not as
        Message 3 of 9 , Dec 14, 2009
        View Source
        • 0 Attachment
          On 12/13/2009 09:34 PM, Andrea Giammarchi wrote:
          >
          > Just to underline another thing:
          >
          > On Sun, Dec 13, 2009 at 6:38 PM, Douglas Crockford
          > <douglas@... <mailto:douglas%40crockford.com>>wrote:
          >
          > >
          > > The point I was making was that if you care about reliability, security,
          > > and performance
          > >
          > >
          > reliability ... they are including the de facto official JSON library for
          > JavaScript
          > security ... they trust your implementation and they trust the fact
          > you keep
          > updating it
          > performances ... they are using a potentially "common used external
          > resource" so if the browser cached already that version performances
          > will be
          > better for every website that includes it plus they are saving bandwidth.
          >
          > As you can see somebody could think that your points ARE the reason they
          > included JSON via the direct source, rather than their local copy
          > potentially non updated and served even if almost every browser has stored
          > somewhere exactly the same library.
          >

          Hi Andrea,

          I think you are forgetting that the (current) json.org website is
          probably just a shared-hosting
          account.

          So that probably means it's not as reliable as something Google or Yahoo
          might do for some of the js-libraries.

          Performance-wise it would be really bad if everyone started hotlinking
          to just that one (or maybe 2 or 3) server(s) as well.

          Security-wise, something like the CDN-like setup Google and Yahoo are
          doing have a lot of save-gaurds,
          like monitoring tools and employees for file-changes. Seperate dedicated
          datacenters or atleast 'cages' of
          dedicated 19"-racks of servers. And not to forget procedures.

          While I do think getting automatic updates of json[2].js would be really
          interresting, because it's a very
          security-sensitive library.

          So in the current situation, it's a really bad idea.

          >
          > So, finally, I would think about a proper specific server or an official
          > repository Github style so that people than use the raw minified and
          > gzipped
          > version with the 304 response, but if you think nobody should ever include
          > external scripts, you should tell us why we all have YUI configurator
          > scripts, google adsense/analytic files, etc etc.
          >

          Yes, I think some people would love to see Yahoo add json[2].js to their
          list of js-libraries
          they are already hosting on their own CDN (I think Google has a whole
          list of libraries).

          But maybe Mr. Crockford does not want his personal project to be tied to
          his (current)
          employer or Google. I don't know their, could be many reasons.

          > Regards
          >
          > [Non-text portions of this message have been removed]
          >
        Your message has been successfully submitted and would be delivered to recipients shortly.