Re: And so it continues... -- MIDRANGE-L

"I guess I wasn't being clear enough."

No. You were pretty clear, at least on the OSS comments. I wrote: "I know you weren't slighting OSS." So I got that part. Your clarification of the "mind set" of the those who might consider OSS was helpful. That would fit with some of those who seek OSS as a solution in the first place: they are looking (usually) to reduce costs, and doing a bare bones installation (read: Kludged) is symptomatic of what you outlined in your missive. However, that same mind set is getting to be pervasive on many IT fronts, again as you outlined well.

I wonder what the solution will be. Selling value is just as hard in OSS as it is with proprietary software. Different issues but same struggle. The thing that most mystifies me is are the folks that transition from a System i (too proprietary) to generic servers running L/U/W OSS and then spending increasing amounts of money to support a multiple hardware environment. I'd much rather run any OSS on a System i since that minimizes hardware and OS administration and lets an organization maximize their application value. I just don't get the "moving to Wintel" mentality for business critical applications. Seems like a step down, in almost all cases.

So, no problem here. Your points are very well taken.

Pete

Joe Pluta wrote:

I guess I wasn't being clear enough. I'm neither pro- nor anti-OSS. It's
kind of like being pro- or anti-gravity. Doesn't matter; it's there.

My problem is with the perception that OSS is free, and so the research,
acquisition, integration and support of it should be free as well. It is
this concept, along with cheap outsourcing (a topic which I am trying really
hard to stay away from) which continues to diminish the value of IT as an
asset to a corporation.

The things we're seeing these days -- downtime of major servers, huge
security holes, failed conversions, on and on -- are symptomatic of the
notion that software is indeed "PVC pipe" as Buck put it. It's all rooted
in the idea that programming is something akin to old-fashioned auto
mechanics; set the right gap, clean the plugs, and off it goes. The truth
of course is that for the most part the most stable software is designed
from the ground up by people who understand the computer at the fundamental
level and who also can understand business logic - a rare combination. The
next most stable software is software built on tested frameworks, but again
by people who know the fundamentals of the system and are able to debug the
frameworks - this by the way implies that the frameworks have source, which
is one of the big benefits of OSS. The next most stable environment is
proprietary (closed source), tested frameworks used by people with solid
programming skills and business knowledge. The least stable is proprietary
frameworks used by end users with no programming skills.

This is simple common sense, really. And it isn't to say that productive
end user tools don't exist, but for the most part those tools that really
work with end users are designed for end users and built with a specific
purpose; the actual business logic is already coded into the tool (e.g., a
BI system with preloaded schema information or an ERP package configured for
a specific company).

Now, originally those tools were hand-coded. These days programming is so
complex that even programmers need frameworks to help them develop end user
tools. Web frameworks, CMS systems, raw data query tools, 4GLs, all of
these can help programmers more quickly develop end user tools. However,
you still need good programmers to integrate these tools, and in fact these
days you have to be a really good programmer -- something of a renaissance
man -- to understand all the pieces well enough to integrate them properly.

Unfortunately, in the push to reduce IT "expenses", rather than pay highly
skilled (read expensive) experts to configure these frameworks and tools
specifically for your business, instead what has happened is that the
programming tools are being pushed out to the end users, or to "integrators"
who understand neither the systems they are integrating nor the business
problems they are supposed to be solving. You end up with systems that
fail, and integrators that blame the tools.

And in the end, you end up with Error 404, or its equivalent.

Okay, that's another goal: only one long speech per month. This is it.
Let's see if I can stick to it <grin>.

Joe