[nSLUG] Possible Berwick Migration

Robert Ashley rb.ashley at gmail.com
Sat Mar 3 00:35:48 AST 2007

Thanks for the links, Donald.

On 3/2/07, D G Teed <donald.teed at gmail.com> wrote:

> 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.

It wouldn't be that dependent on me. I'd just be the traffic cop. The
scenario you sketch is too stark. I wouldn't go without reasonable
assurance I had "better" support than I get now with MS. The business
case, the life-cycle costing would have to be a slam dunk or close to
it. That's why I'd start small and dabble, build on minor successes,
or cut my losses if it looked like a loser...early.

> 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.

I'm a little cautious of self-professed debunkers, setting up myths
for the rhetorical purpose of dismantling them. That said, neither
should they be dismissed wholesale. The straw man, afterall, is still
a man. :-)

I swim in the political pool everyday. My bosses are elected
politicians. You're right to intone that the political wrangle is a
big deal. Of course, there's also the small 'p' office politics (read
'culture') which I think poses the greater challenge, because the
staff are the actual users. The real politicians, however, can make
political hay from a $5 savings, they can hitch themselves to the
morality of open source, and the allure of advertising a leadership
position is almost irresistable.

All in all, though, I think the 'soft' challenges can be surmounted.
It's the relationship with the corporate supporters 'out there', and
there apparent sparse community that scares me. Realistically, I think
I need to hitch to a name, a Novell, IBM, HP, a brand-recognition
thing to build a comfort bridge for any migration.

> 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?

Make sense. Squirrel mail would kill it in one swift stab.
> 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.

Exactly. Political hay. It should never be underrated.

> 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.

Good point not to assume to much about users. But neither we let that
swamp the boat. The same thing might happen by switching to Apple,
although with the built-in "kewl" factor, I suspect the tolerance for
change would expand. Zoom, zoom, zoom.

> 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.

I suppose one person's 'zealot' is another's 'proponent'. So like my
friend is persistent but his enemies call him obstinate. Same guy.

I agree with you on the language trap, that OSS-speak can sound just
so unintelligible geek-jibberish to politicians and accountants. On
the other hand, politicians can understand the monstrosity of real or
virtual monopoly, and pestering licensing fees and frivolous law suits
or threat of them are an affront. It's child's play to make an enemy
of Microsoft, which is why so many jump on that bandwagon. By
contrast, ideologically speaking, the open source vocabulary nestles
into the discourse of public service or civic responsibility without

We can take an authentic political position with OSS, unavailable to
any elected official who might want to champion MS, apart from radical
objectivism of the type retailed by the zealots of pure Capitalism.
It's an interesting debate, to be sure, chocked with twists and turns.
As a public administrator, however, my professional duty is to back
away from partisan influence, to present both merits and demerits of
this or that choice. That said, I'm in charge of the administration
and IT stuff isn't even supposed be in the political sphere anyhoo.
You can't built an election platform on it, unless smitten with a
penchant for the theatre of the absurd.

> 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".

Well, I think  enthusiasm in this thread here has inadvertantly
escalated my toe-dipping into a sweeping tide. My end vision is the
size of Saturn's rings, sure, but my actual action plan at this
embyonic stage is a wee hamster wheel. Zoom, zoom zoom.

> 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).

Yeah, I like this server plan because it's organic, starting small and
invisible, but growable with care and feeding.

Thanks for writing up a storm, Donald. You're a vivid thinker.



More information about the nSLUG mailing list