[nSLUG] Possible Berwick Migration

D G Teed donald.teed at gmail.com
Fri Mar 2 07:36:01 AST 2007


On 3/2/07, Robert Ashley <rb.ashley at gmail.com> wrote:
>
>
> Well actually both. I was considering a small redundant server and two
> or three desktops to start. A zero risk proposition, to experiment.


As the Barenaked's have sung "It's all been done before".  Maybe not all,
but you should be able to learn from other experiences.  Largo, Florida
was in the news a few years ago for making their employees use
Linux Desktops and utilizing open source widely.
http://www.largo.com/index.cfm?action=dept&drill=it
I'd imagine that invite is for internal use, but I'd think there might be
someone able to provide some info for free, and perhaps more for a fee.

In my opinion the hurdle which is bigger than "why is he making
us change everything", is "what happens if this guy is hit by
a truck".  That conservative thinking will make them want something
they know can be supported by a new hire if it is ever needed.

While you are at it, you might as well equip yourself for
the political part of the work.  Google "CIO 10 myths of open source"
and you should find a few takes on that.  The O'Reilly one
is probably more for kernel heads, while this one:
http://www.cio.com/archive/030104/open.html
speaks more to the Administrative level dudes.


> Yes. I can tell you that the needs occupy two poles. One is very
> basic, mainly email, word processing and a little else. The other pole
> is tricky, utility and tax billing and accounting. I don't even want
> to go there yet. But eventually, you bet. I'd love to find an OSS
> municipal billing or accounting application. Conceivably, we could try
> a completely redundant system in that area, make errors on purpose to
> see what happens.
>
>
> Outlook for email. Mail files on a Windows server. I think most are
> using the same versions, but I'm probably wrong. We can move around
> from one desk to another, but not all desktops, just some. Some of
> share Outlook calendars, tasks.  I think retraining would be minimal
> because most staff are not advanced users, so the bar is very low in
> that respect.


If they are aware they are giving up groupware, they will want it replaced
by groupware.  The landscape is varied for that in Linux right now.
I suggest something with rich features to make it feel worthwhile
adopting.  I think you will find the evaluation of this is something
you'll want to take the care and time to do right.  No one will
be impressed with a quick install of squirrelmail.  Horde might pass,
but something using ajax technology would be impressive.
Making good web interfaces isn't something people are likely to
get done without corporate backing.  The best ones I've seen
in my searching involve a dedicated mail server, but remain open
source.  Don't forget that open source can include stuff that is
mainly developed by companies.  Doesn't postfix come from IBM?
Ubuntu from Canonical?

Actually, Largo went for a package called Bynari groupware, which
is commercial.  http://www.bynari.net/  But a lot of bridges have
gone under water since they made that decision and there may
be other free options that perform well today.  Citadel looks
interesting - there are others.

Now here is the weird part.  If you deploy purely
open source software and don't spend a nickel, no one in the
world will know the thousands of dollars you have saved.
But if you spend a little money and still save, joint press releases
on your innovation go out, and you've brought some light to what
you are up to.  Possibly a positive thing.

This is why Redhat, Novell and others are in the news, while Debian
and udpcast and other massive successes are not household names.

With regards to the users, I wouldn't assume too much.  That quiet
user who seems to use their software very lightly is supported by a
handful of crutches.  They have one way to do everything, and
have developed a singular mental mapping between the visual
cues and what it means.  Changes may not be welcomed.
On the other hand, a power user has figured out many ways to do
something and is used to exploring their options.  They may have
a longer list of needs, but they will be more flexible in the
longer run.  For example, if the email client doesn't highlight
spelling and grammar issues the way they are used to,
the gentle user might be lacking their normal context in
a mere email draft.

In general, I think you have to avoid implementing this for the wrong
reasons.  Are you doing this as an open source zealot,
to purge your square of the earth of evil and waste, or are you
bringing in changes which have sound business reasons to
make happen.  In other words, if you explain why you are doing
this in one sentence and have to use the term "open source",
then you are not going to be talking a language the
accountants and politicians understand.  That doesn't mean
you can't explain what open source is, but only
choosing it for purist reasons isn't a reason enough to adopt it
in a business sense.

If I were in your shoes, I would not make this a sweeping venture,
but take it one exploratory step at a time.  I'd call it "exploring
alternatives to Microsoft" rather than "converting to OSS".

Someone suggested starting with server stuff they should not
notice.  I think that is a good idea.  Prove it works.  Prove
that you know what you are doing and can argue the benefits.
Once you are past that point, look at additional changes,
always keeping to the sense of what your business needs are,
and you will get the required respect for the migration efforts
(really, efforts on the part of everyone).

--Donald


!DSPAM:45e80c2614043167720179!
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://nslug.ns.ca/mailman/private/nslug/attachments/20070302/1b5e9652/attachment-0001.html>


More information about the nSLUG mailing list