'Community Beta' has lost its way [part 1]
Bernard Devlin
revolution at knowledgeworks.plus.com
Tue May 15 05:21:01 EDT 2007
Hi Ken,
(I'm going to address this to you, but obviously most of this is just
aimed at a public debate).
> Mark Waddingham has personally responded to virtually all the
> questions and
> bug-related posts on the Improve list
I put it to you that the love-fest to be found in the improve list
might not be an experience shared by those who are not members of
that group. Kudos to Mark for being so responsive to you guys.
> But to say that
> the overall effort of the Community Beta has 'lost its way' primarily
> because a specific bug that most likely affects a very small
> percentage
> of the Rev community is overstating things IMHO.
The point isn't that bug 3196 has not been fixed. It's about a
situation where a long-standing, reported, well-proven bug was being
allowed to remain, 6 months into a Beta test program that was
specifically supposed to remove these long-standing bugs. Trying to
get the attention of the bug-squashing process met with a resounding
silence -- exactly the situation that Bill Marriott was complaining
about so loudly last October.
In the past few weeks I've written to the list about this bug,
written to the relevant forum, written to the Bug-Meister himself,
and updated the bug report in Bugzilla. Until I 'overstated it', I
got zero response. That's a rather different experience to your
experience.
It's great that Bill is collecting information about satisfaction
levels from participants of the Community Beta. However, if someone
has already lost interest in the Beta and in Rev, are they really
going to bother filling in Bill's survey? In fact, I let that survey
sit in my inbox for days before I bothered to complete it, because to
me the whole thing has seemed like a futile, empty, time-wasting
exercise. And I'm one of the people who in the past has loudly
challenged those who proclaimed that Rev was buggy but who were not
going to take an active part in trying to make it better.
Tonight I had a look at who was the original poster of bug 3196, to
see why they have been so quiet in all this. In October 2006, a
whole year after logging this bug, he wrote to the list to say he was
going to stop using Rev because he was so dissatisfied. And there
have been no posts from him since, so I guess that means he followed
through with that decision. Don't you think that there might be
quite a few others who would give up on Rev before they got so far as
to 'overstate' things and get some attention? Most dissatisfied
customers will walk away rather than make a fuss
As to the "small percentage of the Rev community" affected by this
bug: it affects any OS X user who needs to programmatically interact
with the many Unix userland tools and applications found on their
OS. Searching my Rev mail list archive for the last 4 years brings
up 99 posts related to "open process" on OS X, and the vast majority
of these posts are from luminaries of this list. Furthermore, I put
it to you that there would have been even more posts concerning this,
if people realised what it meant to have this working i.e. once they
saw what people were doing with this technique. (When I first
brought up this long-standing bug on the list a few weeks ago, one of
the best x-talkers wrote to me privately to ask for more information
about what is possible using "write to process". That makes me think
it probable that there has not been much interest in this particular
bug because most of the Mac Rev users have a history from Classic
where "write to process" was quite alien.) There might also have
been more interest in this issue if the Dictionary didn't explicitly
tell us that we shouldn't use "write to process" on OS X, and that we
should use shell() instead. In simple scenarios, shell() will work,
but it is by no means an adequate substitute beyond the most simple
cases. And unlike many other bugs, there is no workaround for this.
It _is_ a blocker (and I'm _not_ the one who categorized it as such
in Bugzilla - it's had that status for at least 18 months).
[part 2 to follow]
More information about the use-livecode
mailing list