<html>
<head>
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
</head>
<body bgcolor="#FFFFFF" text="#000000">
On 19/07/10 23:00, Andrew Sutherland wrote:
<blockquote cite="mid:4C44BCFA.6040008@asutherland.org" type="cite">
<meta content="text/html; charset=ISO-8859-1"
http-equiv="Content-Type">
On 07/19/2010 12:23 PM, Robert Kaiser wrote:<br>
<blockquote cite="mid:4C44A64E.7060308@KaiRo.at" type="cite">Anything
that can't be recreated from other data we have stored. If we
start doing that, we break the assumption that was laid on gloda
until this day, esp. if it means that other things than search
stop working with gloda disabled. And I'm not a friend of that,
as IMHO that would mean us basically creating another equivalent
of the Outlook .pst format, which is the largest failure in any
messaging app I ever saw and one of the worst data formats I
ever have heard of as it doesn't allow separation of data
belonging to different things. <br>
</blockquote>
[snip]<br>
</blockquote>
<br>
<blockquote cite="mid:4C44BCFA.6040008@asutherland.org" type="cite">
Export is a trickier problem. While we can always serve up data
as it was first served to us, additional user-created meta-data
may not map into that format at all or in a standard fashion. For
example, Thunderbird e-mail tags may or may not be persisted into
the mbox storage as a custom header line. While other software
may understand how to interpret this header line, it is not
remotely a standard.<br>
</blockquote>
True, but as long as we properly document what we do and try to have
extension authors do the same, we'll be playing nicely on that game.<br>
<pre class="moz-signature" cols="72">--
<a class="moz-txt-link-freetext" href="http://perso.hirlimann.net/~ludo/blog/">http://perso.hirlimann.net/~ludo/blog/</a>
<a class="moz-txt-link-freetext" href="http://flickr.com/photos/lhirlimann">http://flickr.com/photos/lhirlimann</a></pre>
</body>
</html>