[Author Prev][Author Next][Thread Prev][Thread Next][Author Index][Thread Index]

[school-discuss] Re: Typical school / local gov't employee retention rates?

on Thu, Dec 16, 2004 at 12:40:10PM -0800, Michael Dean (michaelldean@xxxxxxxxxxxxx) wrote:
> Karsten M. Self wrote:[1]
>> One of the costs associated with Linux tools is training.  Which if you
>> think about it, is really an investment in staff.  Which then amortizes
>> in a manner strongly dependent on employee retention.
>> While I've worked in tech environments with 20%, 50%, or even 200%
>> (yes!) turnover, I suspect that most educational settings have a far
>> more stable workforce.  Say, 5-10% a year, and likely the lower end (one
>> reference cites 6% for Tennessee[1].
>> Which means that if you have to engage in retraining, ten years down the
>> road, you've still got 50% of the staff around[2].
>> Contrast with PC hardware, assuming a service lifetime of five years
>> (probably typical for educational environments).  If you're rotating out
>> HW annually, that's a 20% rate of age-out on your investment.  Given the
>> proprietary industry's typical "HW/SW" upgrade lock, and product
>> lifecycles of 3-4 years, that also means you're probably looking at
>> maintaining and supporting at least two major versions (and likely 3-4)
>> on your network.
>> Message:  training costs are an enduring investment in staff.  Software
>> and hardware investments age out far more rapidly, particularly in the
>> public sector.

> Well Karsten, I don't agree, 

...with what?

> except perhaps at the network admin level.  

The thing I'm looking at here is the chimera of Win=>Lin migration costs
and training.  In educational and government sector work, you're looking
at a cost (however large or small) which has a long-term benefit.

The idea actually came up in a conversation with a local technical
services provider who pointed out that legacy MS Windows or GNU/Linux,
your costs are _pretty_ much the same.  I take most "TCO" studies with a
large grain of salt, but let's assume that what you win on licensing
costs and admin costs, you lose on training[1]

> At that level, the judicious use of best practices software (Apache, 
> etc.) is not a great step.

Sorry, you're not being clear.  What's not a great step?  Why not?  For

> At the desktop level, there are already many studies indicating that
> student switchover goes very smooth with minimal training.  If a
> teacher knows word and excel, which is problematic, they can easily
> take to openoffice, this is actually more straightforward.  

My experience with a desktop that paired OOo and MS Office (both were
available, OOo had desktop icons) was that kids were about equiïgnorant
in both (ages 6-18, bulk were 12-16), but had no greater level of
difficulty with OOo than MS Office, with the exception of OOo's
autocomplete feature (grossly confusing).

> Our school is moving totally to open source, and retiring NT software,
> not the computers.  And the life cycle on computers is more like 7
> years.

I arrive at the 4-5 year HW lifecycle as what's typical under legacy MS
Windows, not Linux.  Microsoft ages out support, and even where
available, doesn't provide sufficient levels (e.g.:  XP SP2 features not
supported in Win2K).  Which means that even should the HW itself be
fundamentally sound, a compatible MS Windows OS won't be tenable.  And
with the HW/OS lockstep, if you move up to a more current OS product,
you need to upgrade the HW.  E.g.:  WinXP pretty much wants 384-512 MiB
RAM, and a PIII 750 MHz or better processor.   Contrast with a GNU/Linux
workstation at 128 MiB PII-233, or a GNU/Linux thin-client on pretty
much anything network-capable that passes smoke test[3].

So:  yes, you can extend life under GNU/Linux to 6, 7, or more years
(preferably by offloading older HW to low-end server functions).
However, in factoring costs, you're looking at factoring about a four
year replacement cycle for Windows, strictly on the basis of what MSFT
are willing to support.



1.  A: Because it messes up the order in which people normally read text.
    Q: Why is top-posting such a bad thing?
    A: Top-posting.
    Q: What is the most annoying thing on usenet and in e-mail?

    For clarity and to support conversational discussion style, please use
    bottom-posting format:  your reply goes below the material cited.  Trim
    your quotes appropriately and ensure your attributions are accurate.  



2.  Note that I *don't* think that the net is even, if you're paying
    fair market rates.  As I pointed out to my boss a few months back:
    the time put in one weekend on scrubbing crud off a half dozen or so
    PCs would have bought replacements for 'em, at market consulting
    rates ($50/15 min).  Problem in education is that Microsoft are
    willing to discount desktop/client software (OS + Office) to $50 (or
    less, approaching zero) per seat.  At fair market prices, though, I
    see a clear win for 'Nix.

3.  Though I'd probably draw the line at Pentium HW and a screen capable
    of 1024x768x16bpp or better.  Of all system features, display is the
    one that makes the biggest impact on me.  Availability of
    replacement parts (memory, drives, CPUs, fans) is another concern.

Karsten M. Self <kmself@xxxxxxxxxxxxx>        http://kmself.home.netcom.com/
 What Part of "Gestalt" don't you understand?
    They took twenty-seven 8 x 10 colored glossy photographs with
    circles and arrows and a paragraph on the back of each one
    explainin' what each one was, to be used as evidence against us.
    - A. Guthrie

Attachment: signature.asc
Description: Digital signature