Subject: feature matrix (was Re: AbiWord 0.7.10)
From: Paul Rohr (paul@abisource.com)
Date: Fri Jun 16 2000 - 11:48:35 CDT
At 08:46 AM 6/16/00 -0400, Thomas Fletcher wrote:
>Absolutely ... the biggest problem I find is actually trying to
>decide what things have changed on which platforms (so I can
>compare what I'm aiming for), not actually getting the work done.
>
>This is mainly a communication issue. For example when you changed
>the behaviour of the dialogs to go Modeless (a good thing ... too
>bad we didn't do it the first time around =;-) you sent a message
>to the list. The time between me reading that message and actually
>getting down and working on the dialog was a couple of weeks. In
>the meantime other things changed as well. What happens is that
>occasionaly things get lost in the noise. Myself I think that this
>really needs to be tracked as part of the feature matrix, combined
>with consistantly up to date screenshots.
I heartily agree. It'd be really cool if developers know which cells they
should keep an eye on and make green. For example:
- a new BeOS developer zeroes in on that column
- the RTF maintainer owns another column
- whoever's implementing a new dialog coordinates that row
- etc.
Those matrixes were originally created for exactly this purpose, and
originally Bob wanted to keep control of them as a QA check. However, the
usefulness of these documents really plummets if they're not kept
scrupulously up-to-date. IMHO, it's just not fair to expect any one person
to do all that work.
Does anyone have any suggestions of easier ways to ensure that those
documents are both current and accurate?
>Just as an example as
>I was going through comparing the unix code to the qnx/photon code,
>I discovered that there was some pseudo code in the unix side to
>disable/enable the rulers! I'm sure that this functionality was
>announced on the list, but I totally missed it!
Yep. This is a perfect example.
Paul
This archive was generated by hypermail 2b25 : Fri Jun 16 2000 - 11:42:56 CDT