Internetworkers: http://www.ibiblio.org/internetworkers/

[Paul Cory <paulwcory AT yahoo.com>]

Michael,

Thanks for well-thought out response. I agree with you that any 
changeover of school systems and government agencies to Linux is likely 
to be a long drawn out process.  Too many in the Linux community seem to 
think that a large organization can simply switch operating systems and 
applications with the same ease that technically astute Linux users dual 
boot between Windows for games and Linux for work. :-) I t simply ain't 
so.

In the interest of giving the Linux promoters a clearer idea of what 
they are up against, I would like to clarify some of the points you made.

> The cost assessment here is comparing paying MS licenses v.s. the
> alternatives. Given that the MS licensing costs for the larger school
> districts is in the millions of dollars, you can pay for Linux 
> training, and
> still save a lot of money.
The money at stake here is not as much as you think. At least, so long 
as M$ doesn't try to puts the broadsword to our throat on the annual 
licensing of the OS, like they have to those school districts in the NW.

First off, a lot of our licensing costs are currently sunk - for 
example, the cost of all the Windows licenses for all the machines we 
already own constitutes an unrecoverable cost at this point.

Also, M$ has been making some amazing deals to school systems on Office 
licenses.  While the expense is still considerable, it's not as big as 
you might think, and compared to the size of our overall budget, tiny. 
To be more specific, our site licenses for Office 2k and M$ Works weigh 
in at about $750,000 this year and $900,000 next year (we are growing 
fast, and an 11% price hike from MS on the Office licenses didn't help).

That's real money, all right, but let me put that into perspective. This 
year, Wake County Schools are requesting $211,000,000 in local (county) 
funding. So, we're talking less than 1/2 of 1% of the local budget. In 
the context of our total operating budget, $750 million, it's almost 
insignificant.

In addition, I can't tell exactly where the money to pay for the M$ 
license is coming from. School System budgets come from three sources: 
local funding (county and enterprise ops), state funding and federal 
funding. Money from the latter two sources often comes with strings 
attached, meaning savings can't necessarily be rerouted to teacher pay, 
for example. The point: the savings might not impact county resident's 
tax bill at all.

Even if all the funds spent on the M$ license were coming from local 
sources, they would only cover about 35 teachers. While that is 
certainly better than nothing, we're discussing 35 teachers in a system 
with a projected enrollment of more than 105,000 students next year. It 
wouldn't be a significant difference.

All of this changes if M$ comes after us like they did the schools in 
the NW. Such a move would about double our annual licensing expense, 
which is bad enough, but what would be really scary would be the 
potential fines and legal costs if we were found to have made a mistake 
and decided not to go for the Windows licensing program. That potential 
liability would, not too mention a near future of escalating costs as 
our system continues to grow by 3,000 students a year, would certainly 
force consideration of alternatives.
> >
> Instead of paying MS licensing fees, you can pay for consulting help to
> implement the change quickly. Alternatively, just about every Linux 
> user's
> group in the country would jump at the chance to help a school district
> switch to Linux.
Since saving aren't as big as you would suppose, they won't go as far as 
you think. We're still talking incremental rollout for larger school 
districts.

Also, hiring a large number of consultants brings its own challenges in 
the management area. The understaffed IT department will still be 
heavily burdened, as it will have to manage the work of the consultants 
(coordination, QA, and so on).   There's simply not enough of them to 
manage the changeover and keep everything running and keep up with state 
mandates if the switchover schedule is not drawn out over many years. 
You'll have the same problem working with volunteers, along with having 
to compete with their real lives for their time and attention.

Also, for large school districts, I doubt the local LUG has enough folks 
to effect a quick changeover. As an example, the Wake County  Schools 
(27th largest school district in the country, 2nd largest in NC) has 
something like 16,500 desktop/laptop computers now. How many computers 
per TriLUG member is that? :-)



> >
> Good point. Its my understanding that many hardware vendors will already
> supply hardware "stripped" without an OS. I purchased 9 Dell computers 
> this
> way 14 months ago, and saved $4,500. (v.s. the cost of paying for 
> Windows
> 2000 server edition.)
The problem isn't so much the vendors being able or willing to supply 
stripped versions. It's that the contracts that we have with them have 
to be rewritten to include the stripped option. As I said, it's a time 
and effort hurdle.

> > 4. It's not what we use at home/work: l.
> You have a good point. These types of changes will take time. I expect 
> the
> easy targets will be looked at first, such as replacing the back end 
> where no
> one cares what OS is running.
Already happening. :-)

> The actual parent reaction will be interesting to see, especially if 
> they are
> faced with a choice between reducing class size, reducing taxes, or 
> paying
> for MS licenses.

See my comments about the actual savings above. The choice for large 
school systems is not so simple and clear-cut, because the savings are 
comparatively tiny.
> > 5. Lack of software:
>
> > I look at this transition as a process that will take years, just as 
> > it took
> years for MS to take market share in the educational space away from 
> Apple.
> In places where emulators won't work, there is software that will allow
> windows apps to be run remotely from Linux computers.
True, but then you still have to pay M$, reducing the savings. :-)

> > 6. Regulations:
> Don Rosenberg, or someone more familiar with licensing may be able to 
> clarify
> this, but I would be very surprised if any procurement regulations were 
> even
> affected by open source. The reason is 99.999% of commercial licenses 
> are
> designed to protect the software manufacturer, not the purchaser.
The problem is procurement policies and other regulations that state 
that M$ software is required for compliance. In those cases, no other 
solution, even Mac, can get it's foot in the door until the policy or 
regulation is changed. For example, part of our MS site license is 3,000 
M$ Works licenses. Why? Because the state of NC requires M$ Works for 
the mandatory 8th grade computer test.

The obvious answer is to require that the software used produce files of 
the appropriate specification - but that's not how the rule is currently 
written. Getting it changed will take a lot of organized, coordinated 
effort - remember, it's not just a change in the wording, it's a 
potential change in the process of administering and grading these tests 
that opens up the possibility for more errors and means DPI will have 
work harder to make sure everything works as intended (someone will have 
to certify that the Linux option is producing acceptable files). Worse, 
from the Linux advocates' perspective, it's a looooow priority for NC 
DPI, which means even getting their attention on the matter will be 
difficult.

This is the regulatory/procurement  hurdle that Linux faces in the 
government arena. The above is just a small example: such regulations 
and procurement policies suffuse the structure of government and even 
make it into the structure of grants.

One more example: the state has hardware and software standards for 
certain Workforce Development Education programs. Local schools offering 
those programs have to meet the state standard. Linux is not mentioned 
in those standards, and thus can't be used.
> > 7. Politics:

> [... Good points deleted for brevity.]
>
> I believe that faced with draconian MS licensing policies, government
> institutions can now reduce immediate licensing fees, future software 
> costs,
> and all future hardware purchases.
Yes they can. The question is: will the pain and agony of the switch be 
worth the money saved? The more M$ pulls stunts like t has in t he NW, 
the more the answer will be yes. The more they get a clue and stop 
shooting themselves in the foot, the more the answer will be no.
> While there are significant hurdles to
> this switch, I believe that governmental organizations will be hard 
> pressed
> to justify the massive cost difference.
Just remember how small the relative numbers are. And how large the 
conversion costs are and how long it will be before savings are realized.

> . I predict we will
> see a number of groups come to together to affect this change:
>
> 1. Companies spring up that will specialize in Windows to Linux 
> migration.
> 2. Grass roots open source advocacy groups.
> 3. The large IT vendors such as IBM are investing billions to port 
> Linux to
> their hardware, and will have a vested interest in selling this 
> hardware and
> software services to the government.
I agree that if these groups coordinate, they can get the ball rolling 
for change.
> One of the reasons for this pressure, as I'm sure you're aware is local 
> and
> state governments are going through an unprecedented budget crises 
> right now.
> I sure wouldn't be the person who had to run for re-election and 
> explain why
> I had to fire teachers or increase class sizes by 10% just to pay 
> Microsoft.
Again, there's the rub - it's not really about firing a lot of teachers 
or increasing classes by more than a tiny fraction of a student. It is 
about making the best use of the public's money - but, especially in 
education, the relative numbers are so small that they are easy to 
overlook. A school system will spend something like 75%-85% of its 
budget on salaries and benefits. Faced with having to cut the overall 
budget by 2% this year, where do you think people are going to look for 
savings? Especially since it appears that we agree that any conversion 
program, because of changeover and training expense, won't generate any 
savings for at least a year or more.

Politics, alas, rears it's ugly head here again. The folks who oppose 
the public education and those who reflexively oppose higher taxes are 
the ones who lead the charge for cutting government costs in the 
political arena. Unfortunately, they rarely come up with specific 
suggestions for trimming government expense or question why we're not 
experimenting with lower cost alternatives to M$ to see if they are 
appropriate for our situation. They prefer to attack the personnel side 
of the budget, claiming that government and public schools are loaded 
with too many high paid administrators, or simplifying further to claim 
that "a bloated bureaucracy" is the source of all the budget woes. If we 
just force feed the government and schools some budgetary Slim Fast by 
eliminating people, everything will come out right in the end - no 
matter how many personnel reductions or hiring freezes have preceded the 
current budget problems, and regardless of any increase in demand for 
services.

This means that M$ and its licensing policies and costs never make it 
into the light of public debate. Which is a shame.

Paul



_________________________________________________________

Do You Yahoo!?

Get your free @yahoo.com address at http://mail.yahoo.com