[LINK] Linux poised for desktop failure: Gartner

[Malcolm Miles]

On Sat, 16 Nov 2002 12:39:19 +1100, you wrote:

>> Some may well break in the future but to date they continue to work
>> and the business sticks with the philosopy of "if it ain't broke,
>> don't fix it". 
>
>the point is that it's already broke, they just don't realise it yet.

Many of our applications have been running happily for years with no
complaints from the business. They don't see them as broken therefore
they won't fund any changes to "fix" them. 

>> VB developers come cheap.

>yes.  and it shows in the quality of the software.

Perhaps so, but the applications work well enough for the business to
get the job done.

>> things seem OK as they are, so why re-write?
>
> - because eventually the code will break, i.e. to become "future-proof".

Then they will worry about that when they do break. Nobody is going to
spend money now if they can defer it until next financial year.
Short-sighted maybe, but that's how our business runs. 

Many apps have lived through Windows 3.1, Windows NT, Windows 2000 and
XP without breaking. Why deliberately break them by moving to a new
o/s platform? 

> - to avoid paying the ever-increasing microsoft tax.

We pay little Microsoft "tax". Our PCs come with an OEM copy of the
o/s (yes, the cost of the o/s is bundled in, but this is hidden in the
leasing cost) and we haven't upgraded Office for some years. We could
move to OpenOffice or StarOffice without having to change our o/s
platform.

> - to avoid the extra expense in licensing fees (ranging from several
>   hundred to several thousand dollars per "seat").  whenever extra
>   staff are hired...

Our staff are reducing, not increasing. 

> - to avoid the risk of having disgruntled ex-employees dob them in to
>   the MS stormtroopers at the BSA for the licensing irregularities
>   (accidental and/or deliberate) that inevitably occur.

Been there, done that. Microsoft and the BSA are happy with our
licensing.

> - because money that is currently being spent on software licenses is
>   better spent on other things - e.g. general business expenses or
>   software customisation and development or support staff or hardware
>   etc.

We don't spend much on software licensing, certainly not enough to
fund converting all our desktops to Linux, rewrite all our apps and
retrain all our support staff and users.

> - to get better performance out of their hardware.

No problems with performance. Our oldest PCs are Pentium IIIs, enough
grunt for most of our work.

> - to get better reliability and security.

Our business users don't seem to have any concerns with the existing
reliability or security.

> - to have options open up that are currently closed to them.

Like what? It would need to be something that would save them millions
of dollars to justify the rewrite of all our applications and
retraining of all our support people.

>> In the short term, and that is all business is focussed on these days,
>> there is no cost in being trappped by one vendor. 
>
>in the short term, there's no cost involved in ignoring the small leak
>in the head gasket of your car....but if you ignore it, the end result
>will be catastropic damage that costs many times more to repair than it
>would have to fix it before it became a serious problem.

Software doesn't work like that. A bug doesn't become more serious
over time. If the bug starts costing the business money then it will
be fixed in the cheapest way possible. 

>i'm not trying to say "this is what you must do".  i'm just pointing out
>that it's not as difficult as you think it is and that there are some
>significant benefits to be gained.
>
>with a will to convert, a way will be found (and it'll be easier and
>cheaper and faster than you think)...without that will, it is basically
>impossible.

You need more than a will to convert, you need a quantifiable business
benefit to convert, one that is realised in the very short term and
with no significant impact on the business during the conversion.

At this point in time I can't see how converting our thousands of
desktops to Linux and rewriting all our applications would meet that
criteria.

I would love to see some case studies of large corporates with 10,000
PCs or more that have recently converted over to Linux.

-- 
Best wishes,
Malcolm