[6275] in Release_7.7_team
Re: The bash transition
daemon@ATHENA.MIT.EDU (Mitchell E Berger)
Mon Mar 9 22:02:56 2009
Message-Id: <200903100202.n2A22DNs029102@byte-me.mit.edu>
To: Jonathan Reed <jdreed@MIT.EDU>
cc: release-team@MIT.EDU
In-Reply-To: Your message of "Mon, 09 Mar 2009 14:30:39 EDT."
<A98E5B29-4840-4CE9-B8C8-905653184B58@mit.edu>
Date: Mon, 09 Mar 2009 22:02:13 -0400
From: Mitchell E Berger <mitchb@MIT.EDU>
X-Spam-Flag: NO
X-Spam-Score: 0.00
On the one hand, if you push people to transition, there will be a
decent number of them who won't want to, and I'm not sure you'll
solve the support load issue because instead of simply supporting
two shells both in popular use at once, you'll end up with tcsh
questions that are more rare and consultants who are out of practice
knowing the answers over the next four years. But I think the
degree to which that will be a real problem is pretty small.
On the other hand, documentation is ~always a good thing, and I
suspect that the number of users keeping their accounts for over
four years (they MEng, they work at MIT, they get sponsored, etc.)
is pretty high, so as time permits, I think what you propose is
a worthwhile project, because even if we just wait four years,
there is still going to be a quite nontrivial number of users with
tcsh, and far fewer consultants who know how to debug it. I'd
suggest thinking of this document both as benefitting users who
want to transition themselves, and as a stock-answer-like reference
for future consultants who need to answer these questions despite
never having fought with the pains of tcsh themselves.
Mitch
> By the end of the month, I'll be asking Garry to default new accounts
> to bash. We've been updating documentation to deal with this, and in
> doing that, I've noticed that we've removed a lot of customizations to
> our bash config between Athena 9 and Debathena. The stated goal of
> this was to bring things in line more with a default Ubuntu
> environment. However, we now have skew between the default tcsh and
> bash configurations. I realize that in 4 years, many tcsh users will
> have vanished due to attrition, but that's still 4 years away.
>
> I think we've decided that we don't want to change people's shells out
> from under them. However, I'm beginning to wonder if we want to
> publicize the bash transition and actively encourage existing tcsh
> users to migrate. I'm not sure about the best way to target the
> audience in question, but I think we could easily create a document
> about migrating to bash, and its implications, and also a "checklist"
> for the migration. I'm not too concerned about the documentation
> side, but I'm wondering if people think this would be a worthwhile
> effort, or if we should simply wait four years. It seems that the
> world hates tcsh at this point, and I'm concerned that if we continue
> to try and support tcsh, we'll end up where we are with lprng vs cups
> issue. I'm not suggesting we abandon existing Athena tcsh users, but
> I wonder if drastically reducing their number will benefit us in the
> long run, or simply make more work for us now. I'd be interested in
> hearing people's thoughts on this.
>
> Additionally, I wonder if we should bump ~/.generation, so we can
> identify tcsh users who actively chose it as opposed to those who have
> it because it was the old default. I'm not yet sure when such a case
> would be useful, but this is a significant change, and the overhead
> involved in bumping ~/.generation is minimal, so unless there's a
> really good reason not to, I'd like to do that.
>
> -Jon
