Re: Release/Tree closing

Subject: Re: Release/Tree closing
From: Paul Rohr (
Date: Wed Jul 11 2001 - 13:56:17 CDT

At 01:41 PM 7/11/01 -0400, Dom Lachowicz wrote:
>I'd like to close the tree in about 24-36 hours for the 0.9.0 release. Almost
>everything from my and Martin's TODO list has been addressed successfully. If
>anyone would like to object or protest, please holler now and holler loudly -
>this release has been public knowledge and looming for a while now...


Thanks for setting a clear deadline.[1] I'm sure everyone's excited about
getting 0.9.0 out there, and I assume that a lot of behind-the-scenes work
has been going on to make sure we're ready to make a huge stink about how
much cooler 0.9.0 is than all the 0.7.x-series releases.

This is a major, major milestone for AbiWord, and we should expect a lot of
attention -- both positive and negative -- from people who last looked at
AbiWord a long time ago. Based on past experience, the surge in downloads
for a properly-launched 0.x.0 release is *huge*. I wouldn't be surprised to
see *triple* the interest in this release (when compared to most other

Thus, I'd encourage every AbiWord developer to do the following two-hour
exercise *today*:

  - grab a fresh copy of the sources,
  - do a clean release build,
  - launch the product, and
  - imagine that you're your own worst Slashdot-trolling nightmare.

In short, take a fresh look at the product with a critical eye. We've all
gotten excited -- and rightly so -- about the tons of cool work we've added
in recent months, but that's not what people are going to see. A more
likely evaluation process is as follows:

  - Jump right in and start banging away.
  - How polished is the net effect?
  - What's obviously missing?
  - Can I break it, especially the new stuff?

In short, the only criteria most people care about is, as I say all too

  Does it Just Work?

If so, they'll love us. If we come fairly close, they'll start trying to
file bugs in Bugzilla (assuming there's enough disk space) and/or post to
the mailing list(s). If we miss by a mile, then, well ... let's not do
that. :-)

I'm sure that Dom and Martin have already made this kind of assessment, or
else they wouldn't be calling for a release now, but it behooves the rest of
us to take time, now, to see for ourselves.

>Win32 guys - two things:
>1) Does the List dialog work yet?
>2) Is the Style dialog working?
>If not, I will #ifdef WIN32 around these things, but in an ideal world, I
>wouldn't have to do that.

Ouch. I hope I won't be the only person looking at the Win32 build today,
because questions like this scare me. Not as much as Mike's post earlier
today, though. :-(

To be blunt -- it's scary to see us allowing Win32 to fall behind, because
that's where a *ton* of our existing (and potential) users are. According
to our primary source of download counts (SourceForge), we have more Win32
downloads than all other platforms combined!

Of course, these numbers undercount our userbase, perhaps drastically,
because unknown numbers of Linux users may be getting AbiWord from their
distros (none of whom share download numbers with us).

Even so, the SourceForge numbers are important, because anyone interested in
getting their hands on 0.9.0 binaries immediately is likely to come to us,
rather than wait for their distro to package it.

The question I'll be asking myself when I finish building a fresh, clean
view on Win32 is:

  Am I ready for somewhere between 50,000 and 200,000 people to
  judge us by the quality of *this* binary?

Scary thought, huh? Time to go look...

still deciding whether to holler

[1] That's pretty short notice, but at least it's an explicit deadline.

This archive was generated by hypermail 2b25 : Wed Jul 11 2001 - 13:48:53 CDT