>
>
> From:
> Ian Clarke <ian at freenetproject.org>
> Date:
> Sat, 19 Oct 2002 13:12:28 -0700
>
>
>That is a rediculously conservative approach for any software, let-alone
>a beta product. It would virtually guarantee that we don't get a
>release for perhaps two months, while actively *discouraging* people
>from fixing minor bugs since that would further delay the release.
>
>Even Mozilla, with their 1.0 release did not have such a rediculously
>conservative approach.
>
It is my understanding that the mozilla project does *exactly* what
Oskar suggests (with the exception that their check-in freeze periods
are a *month* now). However, since the code tree is branched at the
beta time, feature addition work can go on that makes it into the
nightlies but not into the beta.
You mention in another post that we have had 5 release candidates
already. I wouldn't consider the 0.5prex the builds as rc unless Mat
is managing a separate branch from the nightlies since 2002-10-19
snapshot won't even run for me (due to an error the Oskar pointed
out in an earlier message). That doesn't sound like the results of a
managed chekin process.
You also mention in another post that the windows installer is
largely irrelevent. While I am no fan of Windows and don't run it
myself, that is a very foolish thought process. The installation
process is a major part of any release. While a windows user with
a little bit of knowledge can run Freenet from the tarballs; most
cannot and will not touch an app that doesn't have a friendly, easy
to use, GUI installer. You would be cutting out a huge portion of
potential freenet userbase by not having some *tested* automagic
installers (including rpms and debs) ready for this release. Since
the 0.5pres seem to be shaping up quite nicely, this is the time to
put out a call to package managers (do we have any of those???)
to build up their installers. Giving them a full week and the target
snapshot to work on would be a minimum.
Talking to people that used to use Netscape, Netscape's premature
release of 6.0 has done more damage then good. It doesn't take
much for people to turn away from a product. Netscape learned
their lesson (that they had forgotten). So should Freenet learn
from other's mistakes ... we don't have to make them all ourselves.
Mike
> Be realistic. 0.5 doesn't have to be perfect,
>it merely has to be more stable that 0.3 (since technically 0.3 is the
>code that we are recommending people use until 0.5 is released). It met
>that criteria a LONG time ago.
>
>Ian.
>
>On Sat, Oct 19, 2002 at 09:20:10PM +0200, Oskar Sandberg wrote:
>
>
>>There is a simple algorithm for this:
>>
>>do {
>> freeze(code);
>> time = currentTime();
>> while (NOW < time + ONE_WEEK)
>> if (bugfound()) {
>> unfreeze(code);
>> fix(code);
>> continue;
>> }
>> }
>> release = code;
>> unfreeze(code);
>> return release;
>>} while (true);
>>
>>
>>--
>>
>>Oskar Sandberg
>>oskar at freenetproject.org
>>
>>_______________________________________________
>>devl mailing list
>>devl at freenetproject.org
>>http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
>>
>>
>
>
>
>------------------------------------------------------------------------
>
>_______________________________________________
>devl mailing list
>devl at freenetproject.org
>http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
>
>
>
_______________________________________________
devl mailing list
devl at freenetproject.org
http://hawk.freenetproject.org/cgi-bin/mailman/listinfo/devl
