bug-grub
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

my feelings


From: Yoshinori K. Okuji
Subject: my feelings
Date: Tue, 21 Aug 2001 04:31:24 +0900

What I've seen in GRUB development is that sometimes a problem happens
on a specific system and only on it. As you may know, software
development very often requires that you reproduce the same
phenomenon, at least in the front of someone who is able to
investigate what is the cause.

As GRUB is a boot loader, it uses BIOS extensively, so it is sensitive
to BIOS implementations, because it is rare that BIOS is written in a
robust way. This makes it difficult for the maintainer to solve a
problem which happened in a user's environment. That's because the
maintainer may not reproduce the same problem in his own
environment. Thus the user sometimes has to investigate the problem
herself. The maintainer can give her an advise, but she must have
technical knowledge and spare time enough to debug code, otherwise she
cannot even understand what the maintainer said.

If we were a business vendor, we might have bought a number of PCs so
that we could have almost all kinds of BIOS implementations GRUB users
own. But, in reality, we mainly consist of volunteers. So the
maintainer has a few machines actually, and, frequently, users have
much newer and more powerful machines than the ones the maintainer has.

For now, I know only four ways for that. This order is based on my
preference:

1. You fix your problem yourself, and send a patch to this list. Then
   the maintainer will apply your patch, if you give a good
   explanation.

2. You contribute to someone who may investigate your problem instead
   of you. For example, you can donate your computer, money, or
   something valuable.

3. You wish that your problem would be general and the maintainer
   could have time to investigate your problem.

4. You give up using GRUB.

Sometimes I chose another way, that is, I wrote code, a user tested it
and reported the result to me, I wrote code, the user tested it and
reported the result to me, (like an endless loop). But in my
experiments, that cycle consumes too much time, and that isn't
efficient. I'm not sure what we should do from now on, but those are
my honest feelings.

Okuji



reply via email to

[Prev in Thread] Current Thread [Next in Thread]