Re: Mainframe -> PC -> Mainframe

Hey.  Just adding in a couple of more $0.02 pieces.
(Note: things not commented here on I generally agreed with and the message 
was well written.)

> 8 KB of memory might have been dandy for a mainframe by 1960, but by 1980
> it would be more like 1 MB, by which time microcomputers sported 8 KB.

   This makes sense due to the design of what the computers were for: the 
mainframe was designed to deal with lots of users, microcomputers at that 
time were designed for a single user in mind.
   While the mainframe practially needs it's own power station to run it, it 
actually also ends up being more power efficient than having individual PCs 
at people's desks.  This is similar when it comes to busses that are fuel 
inefficent but end up being more fuel efficient than those individuals each 
seperately having their own cars.

> > PCs and users
>
> The ability to control your own computer was an important benefit for the
> Personal Computer. The lower price made it affordable for menial tasks like
> word processsing.

   Yes, agreed on both.  It's hard to pull out a mainframe down to the lab.

> Mainframe software was and is expensive, but when amortised over a thousand
> users, a mainframe can be cheaper than a PC solution.

   I agree with this - but that solution also makes computers less accessible 
and less customizable.  If it weren't for the microcomputer movement I 
probably wouldn't be doing wireless internet browsing with a laptop.

> When IBM introduced its Model PC in 1981, it quickly became standard office
> equipment, and the commercial software market grew very fast.

   ... And the thought there at the time was an expectation of increased 
productivity and a "paperless office" paradigm, which never materialized...  
And we're certainly _not_ saving paper, either.  The studies that I have read 
show that productivity with the advent of computers has actually decreased 
rather than increased due to the maintenance costs in terms of time.
   However, this is probably not taking into account the complexity of certain 
types of design and simulation.  It would be difficult to draw a design for a 
modern video card with technical drafting with a pencil.

> Microsoft
> makes the bulk of its income from the commoditised Windows OS and Office
> application suite. As PCs became more capable and complex, they hardly
> got cheaper, and today's information professional prefers the notebook
> over the cheaper desktop.

   Desktop PC's in the 80's were ~$3,000 and sported a 286 processor with < 
640 KB of RAM and had no GUI - and, mind you, that's 1980's dollars.  Today's 
2.4 GHz machines go for ~$1,000 and have complex GUI OS's.  I still believe 
that PCs have gotten much cheaper, at least to purchase individual machines.  
Manufacturing processes have gotten cheaper and run on lower profit margins 
than they previously did.
   They might more expensive in other ways, and the movement towards laptops 
is plainly evident, but even those are cheaper (for individual purchase) 
compared to older microcomputers "back in the day".
   There are lots of expense concerns from start to finish of the life of a 
computer, which includes humanistic issues, environmental issues, etc.  The 
"real" cost/benefit ratio of microcomputers is a nontrivial study.

> Commercial software has a future in every place where free software does
> not meed the user's needs. Companies and governments continue to buy
> tailor-made software.

   ... And I believe this will continue to be true because there are niche 
markets that free software is not able to focus on right now [and possibly 
never will, but that remains to be seen].  Somehow I'm not sure there's 
enough OSS programmers needing to do 3D electromagnetic simulations for an 
OSS program to come out for that, etc.
   ... So commercial software isn't going away anytime soon.

> There is no reason why a piece of software could not
> be sold commercially under the GPL license.

   Yes - agreed.  I have examples of this occuring as we speak, but I don't 
want to be a secondary advertiser, either, so I'm not going to list any at 
the moment.

> The recession caused the software market to shrink, and helped to foster
> cheaper alternatives. The (earlier) dot-com boom left us with a large
> number of (web, e-commerce) servers, with a demand for managing them.
> Following 11 sept.'01 the demand for security grew, and the continuing
> computer virus epidemics foster them. These demands remain unfulfilled
> because companies cut back on IT budgets.

   Unfortunately this is all true.  IT budgets are often considered "overhead" 
for companies as it's considered under support, verses where production and 
sales departments mean business.  This leaves IT in the corner of the 
chopping block due to the way that it's typically accounted for.
   The last company I worked for laid off the only Unix admin that they had as 
well as dropping commercial support for all their Unix machines - 
simultaneously - after 11 sept.  I guess they expected that "we'll just get a 
consultant when we need one", which I'm sure was shortsighted.  They had Unix 
machines all over the place.

> Most of the commercial potential is probably in system and network
> administration and management. Linux is rather laborious to manage.

   It can be, but doesn't _have_ to be.  Depends on the system/distribution in 
question.  I will admit that it often _IS_ for one reason or another.
   [Linux is not the only OS in this category.]

   There are many linux systems that I have to administer that make me seethe 
- there are others that are a pleasure.  Unfortuantely getting a company to 
switch to systems that are a pleasure seems to be incredibly difficult 
because they want the good feeling they get from having commercial support 
that they never ever use...  And don't even consider calling in an emergency.

	- Chris

-- 
Chris Knadle
Chris.Knadle-15hjz6xD4c1Wk0Htik3J/[email protected]