*BSD News Article 41624


Return to BSD News archive

Xref: sserve comp.os.386bsd.misc:5091 comp.os.linux.misc:34103 comp.os.os2.advocacy:76192
Path: sserve!newshost.anu.edu.au!harbinger.cc.monash.edu.au!msunews!agate!nntp-ucb.barrnet.net!well!henryh
From: henryh@well.sf.ca.us (Henry Hwong)
Newsgroups: comp.os.386bsd.misc,comp.os.linux.misc,comp.os.os2.advocacy
Subject: Re: Linux thoroughly insulted by Infoworld!
Followup-To: comp.os.386bsd.misc,comp.os.linux.misc,comp.os.os2.advocacy
Date: 27 Jan 1995 04:03:18 GMT
Organization: The Whole Earth 'Lectronic Link, Sausalito, CA
Lines: 95
Message-ID: <3g9ra6$a2o@nkosi.well.com>
References: <950116203411@lambada> <3g0mnn$ccp@nkosi.well.com> <3g267v$1e3@blackice.winternet.com> <3g34fr$jri@nkosi.well.com> <3g573n$8np@bar.autpels.nl>
NNTP-Posting-Host: well.sf.ca.us
X-Newsreader: TIN [version 1.2 PL1]

Ruurd Pels (ruurd@autpels.maxine.wlink.nl) wrote:
: In article <3g34fr$jri@nkosi.well.com>, 
: Henry Hwong <henryh@well.sf.ca.us> wrote:

: >: Certainly.  Would it cost less to develop?  Certainly.  Can you justify
: >: $1000 to buy a SCO license for that platform?  Only with great dificulty.
: >: It seems to me that Linux is a perfect solution.  That Linux isn't a
: >: very good solution for other problems doesn't strike me as very relevent.
: >
: >Yes, I can cost-justify purchasing a UNIX (though, I wouldn't personally
: >choose SCO, so you're making it hard for me to defend my position :-)).

: Maybe you can, but you're not everybody...

Can I take that as a compliment? :-)

: >I think the biggest point that I'm trying to make (and apparently not
: >making clear) is that there are other costs associated with going to a
: >Linux (or *BSD). Sure, inital costs are cheap -- free FTP, cheap CD-ROMs.
: >
: >Sure, I can try to convince my clients to look at Linux. It's free!
: >However, now I have to look at how much time it would take to maintain or
: >setup. For a consultant, time is big bucks. Many times, it is cheaper
: >for me to recommend a whole solution (Sun, HP, IBM) that comes in a package
: >with everything installed instead of me spending the time to put together
: >a 486 or Pentium system running Linux.
: >
: >Then we have to do technology transfer to the client. It's easy to find
: >classes for Solaris, AIX, SCO, or HPUX. Linux classes? Uh. Hmm. Well, I
: >could spend the time ($$) to teach you everything about Linux. RTFM just
: >doesn't cut it with clients. And I can't tell clients to just fire all 
: >their mainframe people and just hire people who recently graduated from
: >college, either.

: I don't want to be rude, but basically you're telling them what they want
: to hear, right? 

Not really.... see below....

: >That's why consultants like me pick the packages. I have to find the best
: >solution for my clients. Long term, that $1000 I tell my client to spend
: >on SCO is cheaper than going with Linux. The hidden people costs are
: >what kill you.

: Here you're touching on the trouble spot. The hidden people cost. Which is
: certainly the thing your clients don't want to hear about. Why not be honest
: and tell your customers about it? Wether a client has a bunch of boxes 
: running whatever UNIX variant or Linux, they all will end up having to hire 
: or train people for the maintenance of such systems eventually.

Actually, you'd be quite surprised. Most IS groups for large corporations
are quite aware of the people costs. That's why they like consultants,
because, even at our rates, we are cheaper than hiring a full time 
employee. The value we add is that we transition our knowledge over to
the client. If we didn't, they we would simply be contractors.

The point I'm trying to make is that there is more than just initial cost.
We have to take into consideration how much it will cost the client to
transition the information over. People time is often more expensive than
anything else.

IMHO, retraining and retaining is cheaper than churning employees. People
enjoy working for a place that cares for them of which retraining them for
new technologies is part. All that money that you save in not training
your employees does not cover the cost of recruiting, enculturating them
into your organization, and the higher salary you must pay because people
don't expect any kindness from the company. When you start looking at the
big picture, you start becoming aware of the implications of such
decisions.

(I know you weren't commenting on that -- we seem to agree. I just had to
 throw that in and get it off my chest. :-) )

: Oh, 
: certainly, everything runs just right out of the box. (I doubt that. I would
: like to give into consideration the time needed to set things up for networ-
: king such a plug-and-play UNIX for example). But the tendency is that 
: computer systems don't stay that way and need human intervention now 
: and then to keep healthy. Availability of classes to learn a person how to
: work with a system not available? There are enough alternatives to classes, 
: provided your clients are willing to pick someone in their workforce who is
: eager to do the job and give him the means and time to venture into the land 
: of UNIX. 

Exactly. We always tell them that. I think classes offer a valuable service
in that the books and notes end up being great references. Sure, I can
try to mentor (which I do as much as I can) someone in the intricacies of
UNIX, but when you see how much you are charging them per hour, you 
quickly understand the cost of people time. IMHO, classes speed up the
learning process. We don't expect UNIX kernel hackers out of the classes,
but at least they are then in a position to understand a lot more in a
shorter period of time. Classes don't teach you everything -- if anything,
they teach you how to learn.

-Henry