Subject: [ILUG-BOM] ZDNet article
Date: Tue, 04 Jul 2000 18:50:42 +0530

terrence d'souza wrote:

> These Companies should be sacking their IT staff.
> The job of IT dept. is to keep cost down thru IT.

A path that most companies won't take. Engineers with brains are /very/
difficult to find, so most companies have to settle for anyone who can
push a mouse. And if you give them an OS where their mousing skills are
irrelevant, they'll go straight back to Windows. It does their job just
fine, thank you. Who really cares for Nasscom and their handcuffs?

> They already have IT staff so seems the point is
> irrelevant.

Again, most IT staff know little more than point-and-click, and base
their decisions on a contact list of advisers for what to install to do
whatever.

> If the applications were customised then they
> should have the source and could recompile. There
> would be issues. But the IT dept better start
> earning their money. As it were they are in a
> mess because of closed source and continuing to
> throw good money after bad is not exactly good
> managment.

If you've been following the open source movement closely, you'll know
that there are times when it simply does not work. The gaming industry
is one, highly customised software is another.

All those companies had their COBOL sources, but did that give them any
foresight on the Y2k problem?

(skip the issue of the mess they would have been in if they didn't have
any source)

> Very conveniently forget that Linux and OS is
> based on standards and that the biggest violater
> of standards is MS.

Says who? Linux is based on accepted POSIX standards, but is anyone
maintaining standards on new developments? None of the Linux distros
have even managed to agree on an /etc/rc.d convention.

> It is. Too many examples (read erlier post on
> hardware). Also you would not dare use a MS
> product in an embedded system. Imagine your smart
> card reader in the lift controlling the lift
> ascent to the  15th floor hanging. You would wind
> up on the 128th floor (or worse if you were going
> down).

Microsoft has produced software of exceptional quality when pressed into
doing it. Take Windows NT 3.51 as an example. Give credit where it is
due.

> 2) Mis information like the above:
> If Managment did a cost benefit analaysis for the
> IT dept the cost savings of replacing the
> existing software system would be so high that
> Managment would immediately do everything
> neccessary (including kicking out a few from the
> IT dept).

I'll take you on this anyday. Take my situation for an example: I
maintain two Linux mail servers in two offices in Bombay. My company
needs to setup mail servers in other cities to handle the growing user
base there. But there are choices to pick from:

1) Get a Linux server and a linux admin to maintain it. Experience tells
me that it is impossible to setup a server with no frequent management
needs. Even misbehaving modems need a tech to handle them. And getting
Linux techs is next to impossible. The best you can usually find is
someone with his experience record stating "installed Red Hat and
connected to the Net". Like that's the greatest achievement of their
lives. Besides there's the complete waste of paying him ~8k a month just
to ensure that everyone gets their mail.

2) Get a Linux server and train one of the staff. Train a marketing guy
to admin Linux? Give me a break. This is just the kind of guy we love to
describe as "suit" and "marketdroid".

3) Get a Linux server and find a local consultant who can come in when
there's an issue. Given that most of my local users start to complain
two minutes (yes!) after there's an issue with the server, waiting half
a day for the consultant to come in is unacceptable. And if he can't do
it, he's going to put the blame on the person who originally configured
it, but the problem remains unsolved.

4) Pay a one-time 25k license fee for a commercial solution on Windows.
The average user knows Windows well enough to manage it himself, Windows
support is so easily available, and if there's a crash, fixing it as
simple as a reboot. The 25k is justifiable in that there's no recurring
cost in hiring an admin or having me travel every time there's an issue.

Now what do you think I would choose?

> For category B a first time user would not be
> very concerned wether the os is WIN LIN or
> anything else as long as it could read DOC/xls
> files, browse the net and PLAY GAMES (that's what
> got computers into the house ;-)). With the
> migration of computers to schools and cybercafes
> (both cash strapped entities) Linux will be
> perfect choice.

The issue is, can Linux read DOC/xls files and play games? The answer is
no for at least a good year to come.

> An existing SOHO will be reluctant to switch
> until  the next (forced) hardware/software
> upgrade. Thin client at Rs.7,000-10,000 and the
> old box as server- made to order for Linux.

And if the old box goes down, so do /all/ your clients. Have you ever
managed to convince management into a thin client solution after
explaining this possibility?

They'd much rather spend more on individual boxen. When an average sales
guy brings in 2 lakhs in contracts in a week, saving on the 50k for a
box that will last him one year is a silly idea. No sane management will
do that.

> One major diff btwn USA and here is the very low
> penetration of computing in India. As such Linux
> stands an equal chance and presents a great
> opportunity to all Linuxers.

The issue, my friend, is with the existing base. Newbies depend upon
existing users for directions, and if the existing user has had bad
experience with Linux, you can bet the newbie will steer clear.

There are very good reasons why everyone hasn't already switched over to
Linux. Think about it.

-- 

Kiran Jonnalagadda
http://lunateks.com

baby.sh: while true; do echo "^G^G^G^G^G"; sed -e 's/food/poop/';
sync; sync; sleep 15; done