Re: [MLUG] licenses and choices (moved from discussion)

[Jack Smith <EMAIL:PROTECTED>]

On Tue, 2007-09-25 at 14:44 -0500, Mike Miller wrote:
> On Tue, 25 Sep 2007, Rick wrote:
> 
> > Well, while I have the linux bigot's love of open source projects, I'm 
> > pragmatic enough to understand that companies don't do things for some 
> > crazy altruistic reason. They do things to make money.
> 
> Lots of people do lots of things for what amount to altruistic reasons, or 
> maybe to get credit for doing a good thing.  I wouldn't call "love of open 
> source projects" a form of bigotry:  There is a lot to like.

Another scenario I can think of is that somebody solved a specific
problem with a piece of software they wrote (for example, a tool to
automatically shrink movies on 9 GB DL-DVDs into 4.4 GB single-layer
DVDs for backup purposes). The problem they solved was enough of a PITA
and they wished there was a program to do it, but there wasn't and they
had to write it. The problem isn't all that complicated, so it's not
likely that a program would be a big seller, so it's open-sourced
because otherwise, it would more or less just sit unused. 

> I agree that it is not unethical to sell proprietary software.  It just 
> hasn't gone well for me when I've used it.  My experiences with free 
> software (open source) has been much more positive.

It largely depends. Some proprietary software (especially games) are
much better than what's available for free, just due to the complexity
of the software and the manpower required to make it isn't very doable
by a small group of people working in their spare time. I like to play
Sauerbraten and Nexuiz, but those are more comparable to late 1990s or
early 2000s games like Half-Life 1 than to any of the premier titles of
today. Ditto with CAD programs- QCAD is about the best free CAD out
there and it's roughly like AutoCAD from the early 1990s. But with the
exceptions of those huge or very specialized projects, OSS competes
pretty well with proprietary software. 

The biggest advantage in using OSS over proprietary is really in the
deployment of the software. You don't have to keep track of licenses.
You don't have to pay per-seat. Distribution is easy- just download a
copy and install. You are also not tied to legacy hardware or OSes,
which is an especially big factor today with the 32 -> 64-bit switch and
MS shaking up OS compatibility.

> One irony is that when I pay for proprietary software, and I make a 
> recommendation about how it can be improved, that recommendation is nearly 
> always ignored.  I payed, but I get ignored.  With open source projects, 
> my recommendations are very often responded to immediately, often with 
> patches to the source code within a day or so.

It can depend on what bugs or requests you submit and what proprietary
vendor you submit them to. I'll bet that if you submitted a feature
request to the Lotus office suite folks, they'd be *far* more likely to
incorporate it than if you submitted a request to Microsoft for Office.
The Lotus guys would be quicker simply because of the incentive to gain
a competitive advantage, which MS really does not care about. 

> Me too.  It happens every time I buy something!  The problem with software 
> is in the vendor lock-in.  I get stuck in a rut using a certain program. 
> What if the company that sells it decides to make my life difficult and 
> suck more money out of me?

First, you should be on the lookout for lock-in. There are enough public
standards out there that you should be able to find at least one vendor
that will support them. If that's not the case, then go with a vendor
who is not the biggest one in the field as the big guys pretty uniformly
try to stay that way by locking people in. A smaller competitor would be
less likely to risk teeing off and losing their customer base as their
former customers would probably go to their larger competitor and make
them all that much more susceptible to going belly-up. 

>   Depending on the type of software, they might 
> do that very effectively.  Example:  SAS statistical analysis software has 
> an annual license.  They change the price at will every year.  We used to 
> pay $50 per PC per year with academic discount, but now it is $150.  When 
> I got a Linux box they said they'd give me a discount at $3,800 for the 
> first year which is much less than their usual $19,000/yr for that kind of 
> Linux box (dual Xeon machine).  On the other hand, with R you pay $0 per 
> year, every year for a constantly improving system.  R is free software 
> distributed under the GPL.
> 

That is simply ridiculous. Anybody that charges that much for software
either is the only one making it or darn close to it, or else the CIOs
are complete morons. For that much money, I'd hire a dev to make R do
what SAS did and R doesn't yet do rather than pay for SAS. Sure, hiring
a good software dev can be expensive, but they are *your* dev and you
(should) get exactly what you want at the end. Plus, a dev's yearly
salary will be what, 3-5 years of SAS licenses? It would be one thing if
SAS were $500 a year, but at $20k, this makes a lot more sense. I can't
see much of a downside in this other than you having to manage and be
responsible for the dev if they screw up. This is probably the biggest
hurdle as people don't like to put their neck on the line if they are
not forced to- many stagnant companies breed that philosophy. So buying
an outrageously expensive software package but being able to point the
finger at SAS is pretty comforting to many PHBs and should make any
decent manager livid.

--jack


_______________________________________________
members mailing list
EMAIL:PROTECTED
http://mlug.missouri.edu/mailman/listinfo/members