<html><head></head><body style="word-wrap: break-word; -webkit-nbsp-mode: space; -webkit-line-break: after-white-space; ">Interesting.<div><br></div><div>Besides... size is largely overrated.<div><br></div><div>People love to impress other people with numbers that usually say nothing about real value. Surprisingly, sometimes even money is not the best metric of value. Hello twitter, positive example, hello derivatives, economically tragic example.</div><div><br></div><div>I wouldn't get crazy for people that doesn't get that and is insensible enough to try to impress you in the wrong way anyway.</div><div><br></div><div>Again: if you focus on costs, you will make them bigger, if you focus on value you will make it bigger. What things are you qualifying as valuable are the things that will inspire us (or produce the opposite effect.)</div><div><br></div><div>Nevin I'm glad to hear you made it work like that and thank you for being generous enough to share it with us</div><div><br></div><div><div><a href="http://twitter.com/#!/sebastianconcpt">sebastian</a></div><div><br></div><div>o/</div></div><div><br></div><div><br></div><div><div><br></div><div><br><div><div>On Apr 16, 2011, at 8:37 PM, Nevin Pratt wrote:</div><br class="Apple-interchange-newline"><blockquote type="cite">
<div bgcolor="#ffffff" text="#000000">
    Define "handful".&nbsp; You'd be surprised.<br>
    <br>
    We are a large enough company with enough orders, to support more
    than a dozen employees, and a half a dozen contractors.&nbsp; Our
    warehouse/office is now about 26,000 square feet.&nbsp; And, we're
    growing, and profitable, with 2011 looking to be our best year ever.<br>
    <br>
    The exact numbers are confidential, but I am confident that we are
    the largest company on the planet in our field (i.e., "Reborn Doll
    Supplies").<br>
    <br>
    Besides, the orders aren't exactly "hand reentered".&nbsp; I just didn't
    give enough detail, because I didn't feel it was relevant.&nbsp; We have
    an admin page that can easily slurp the data in rapidly, if needed.&nbsp;
    It's not hard to do.<br>
    <br>
    And Squeak/Seaside has served us well.&nbsp; And we just use the image
    for persistence, as previously mentioned.&nbsp; And we are very happy
    with it.<br>
    <br>
    Nevin<br>
    <br>
    <br>
    <blockquote cite="mid:B25FA454-F5E9-4305-9C59-F3D3454A60E2@deepcovelabs.com" type="cite">
      <div>Of course that only works when you have a handful of orders
        to reenter, which isn't the case for most systems.<br>
        <br>
        Sent from my iPhone</div>
      <div><br>
        On 2011-04-16, at 12:41, "Sebastian Sastre" &lt;<a moz-do-not-send="true" href="mailto:sebastian@flowingconcept.com">sebastian@flowingconcept.com</a>&gt;
        wrote:<br>
        <br>
      </div>
      <blockquote type="cite">
        <div>
          <div>So, 15 minutes times 4 makes a whole hour per year.</div>
          <div><br>
          </div>
          <div>
            <div>Beautiful :)</div>
          </div>
          <div><br>
          </div>
          <div>Doesn't even justify to implement automation of the
            manual rollback (that sounds feasible)</div>
          <div><br>
          </div>
          <div><br>
          </div>
          <div><br>
          </div>
          <br>
          <div>
            <div>On Apr 16, 2011, at 1:29 PM, Nevin Pratt wrote:</div>
            <br class="Apple-interchange-newline">
            <blockquote type="cite">
              <div bgcolor="#ffffff" text="#000000"> Sebastion, that is
                *exactly* what I initially did at Bountiful Baby.&nbsp;
                However, it has been several *years* since I've had to
                do your #4, so things have changed slightly since.&nbsp; <br>
                <br>
                Bountiful Baby is an eCommerce site, and the critical
                "object graph" (your #2, below) information consists of
                inventory data (the website keeps track of our
                inventory), and gift certificate data (for gift
                certificates that have been issued).&nbsp; Also, whenever
                either of those datums change, the website sends an
                email-- for example, emails are (obviously) sent for
                each order accepted, and, it sends emails when it issues
                a gift certificate.&nbsp; So, it is easy to discover (via the
                emails) what data was lost since the last image crash.<br>
                <br>
                Consequently, currently this is what happens:<br>
                <br>
                1. image is saved from time to time (usually daily), and
                copied to a separate "backup" machine.<br>
                2. if anything bad happens, the last image is grabbed,
                and the orders and/or gift certificates that were issued
                since the last image save are simply re-entered.<br>
                <br>
                And, #2 has been *very* rarely done-- maybe a two or
                three times a year, and then it turns out it is usually
                because I did something stupid.<br>
                <br>
                For us, it's a whole lot easier to do persistence this
                way than bothering with any persistence mechanism.&nbsp; And,
                it turns out, it is *very* reliable, with exactly one
                easily fixable glitch:<br>
                <br>
                The glitch is: occasionally the production image UI will
                freeze, for no known reason.&nbsp; It doesn't effect Seaside,
                though, so the website keeps going just fine.&nbsp; And, if
                we run the "screenshot" app (in the configuration
                screen), there is a link at the top for "Suspend UI
                Process", and "Resume UI Process".&nbsp; Just suspend and
                resume, and the UI becomes unstuck.<br>
                <br>
                We've been doing persistence this way for years now, and
                I've been *extremely* impressed with the reliability.<br>
                <br>
                Before that, we used PostgreSQL and GLORP for
                persistence.&nbsp; But I yanked that code out years ago.&nbsp; It
                wasn't worth the bother maintaining it.&nbsp; <br>
                <br>
                If you have a daily image save, then on average there
                will be 12 hours of lost data on a "random" crash.&nbsp;
                Re-entering 12 hours of orders and/or gift certificates
                (discovered from the emails, as mentioned above) might
                take 10 to 15 minutes.&nbsp; Not a big deal at all for us.<br>
                <br>
                Ten years ago I wouldn't have even considered doing
                persistence this way, but I've changed.&nbsp; Squeak has
                changed.&nbsp; Seaside has changed.&nbsp; And all for the better.<br>
                <br>
                Nevin<br>
                <br>
                <br>
                <br>
                <blockquote cite="mid:DCBD59A6-5593-4214-A21D-33E9AA6D47C9@flowingconcept.com" type="cite">
                  <div>Given that you don't need transactions, to make
                    that style work, I suggest this:</div>
                  <div><br>
                  </div>
                  <div>1. save the image from time to time (as best
                    suits your needs) like you're doing now</div>
                  <div>2. use a secondary way to dump the object graph</div>
                  <div>3. use the normal image for as long as things are
                    good</div>
                  <div>4. when shit happens you "transplant" the object
                    graph into a new "reincarnation" of your app in a
                    fresh image</div>
                  <div>5. repeat</div>
                  <div><br>
                  </div>
                  <div>For dumping the ODB you have options: image
                    segments, SIXX comes to mind now</div>
                  <div><br>
                  </div>
                  <div>If you make for yourself some "rescue" kit
                    (script, tools, preloaded code in fresh image), you
                    can make 4 quite painless (or, why not, monitored
                    and automated)</div>
                  <div><br>
                  </div>
                  <div>
                    <div><a moz-do-not-send="true" href="http://twitter.com/#%21/sebastianconcpt">sebastian</a></div>
                    <div><br>
                    </div>
                    <div>o/</div>
                  </div>
                  <div><br>
                  </div>
                  <div><br>
                  </div>
                  <br>
                  <div>
                    <div>On Apr 16, 2011, at 12:23 PM, Michal wrote:</div>
                    <br class="Apple-interchange-newline">
                    <blockquote type="cite">
                      <div><br>
                        hi - <br>
                        <br>
                        Despite the warnings, I am really interested in
                        sticking to the<br>
                        simplest way of saving my seaside application
                        data, ie periodically<br>
                        saving and backuping the image. The seaside book
                        states that <br>
                        <br>
                        &nbsp;"saving [the image] while processing http
                        requests is a risk you<br>
                        &nbsp;want to avoid."<br>
                        <br>
                        What is the status on that? Is that something we
                        can fix? I have been<br>
                        running an image in this mode for a few weeks,
                        with no ill effect so<br>
                        far, but I have had major problems with old
                        image/vm combinations. So<br>
                        is this something that might be fixed already?<br>
                        <br>
                        Also, I recall that Avi had made a number of
                        attempts at having an<br>
                        image saved in a forked background process, eg<br>
                        <br>
                        <a moz-do-not-send="true" href="http://lists.squeakfoundation.org/pipermail/squeak-dev/2005-October/095547.html">http://lists.squeakfoundation.org/pipermail/squeak-dev/2005-October/095547.html</a><br>
                        <br>
                        did anybody pick up on this, or did anything
                        come out of it?<br>
                        <br>
                        thanks, <br>
                        Michal<br>
                        _______________________________________________<br>
                        seaside mailing list<br>
                        <a moz-do-not-send="true" href="mailto:seaside@lists.squeakfoundation.org">seaside@lists.squeakfoundation.org</a><br>
                        <a moz-do-not-send="true" href="http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside">http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside</a><br>
                      </div>
                    </blockquote>
                  </div>
                  <br>
                  <pre wrap=""><fieldset class="mimeAttachmentHeader"></fieldset>
_______________________________________________
seaside mailing list
<a moz-do-not-send="true" href="mailto:seaside@lists.squeakfoundation.org">seaside@lists.squeakfoundation.org</a>
<a moz-do-not-send="true" href="http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside">http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside</a>
</pre>
                </blockquote>
                <br>
              </div>
              _______________________________________________<br>
              seaside mailing list<br>
              <a moz-do-not-send="true" href="mailto:seaside@lists.squeakfoundation.org">seaside@lists.squeakfoundation.org</a><br>
              <a moz-do-not-send="true" href="http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside">http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside</a><br>
            </blockquote>
          </div>
          <br>
        </div>
      </blockquote>
      <blockquote type="cite">
        <div><span>_______________________________________________</span><br>
          <span>seaside mailing list</span><br>
          <span><a moz-do-not-send="true" href="mailto:seaside@lists.squeakfoundation.org">seaside@lists.squeakfoundation.org</a></span><br>
          <span><a moz-do-not-send="true" href="http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside">http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside</a></span><br>
        </div>
      </blockquote>
      <pre wrap=""><fieldset class="mimeAttachmentHeader"></fieldset>
_______________________________________________
seaside mailing list
<a class="moz-txt-link-abbreviated" href="mailto:seaside@lists.squeakfoundation.org">seaside@lists.squeakfoundation.org</a>
<a class="moz-txt-link-freetext" href="http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside">http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside</a>
</pre>
    </blockquote>
    <br>
  </div>

_______________________________________________<br>seaside mailing list<br><a href="mailto:seaside@lists.squeakfoundation.org">seaside@lists.squeakfoundation.org</a><br>http://lists.squeakfoundation.org/cgi-bin/mailman/listinfo/seaside<br></blockquote></div><br></div></div></div></body></html>