[svlug] Proposed series of Linux articles in DDJ

[Reginald B. Charney]

Dear Tom,

Thanks for the feedback. Your point about minor, but very annoying,
incompatibilities between distributions in Linux is very well taken.
Like its older brother, Unix, these differences may cripple Linux just
as it seems to be coming into its own. It may even become its greatest
weakness from a marketing point of view. That said, there will always be
legitimate differences and
incompatibilities between versions from the same vendor (see differences
between MS Windows versions).

There is also the issue of distributions needing to distinguish
themselves. When this extends to exclusive features, we, as a community,
start to suffer. There is an effort in the Linux community to create a
base standard distribution. Interestingly enough, there was a discussion
recently on this mail reflector on where system files are to be placed.
There is a document,
called the Filesystem Hierarchy Standard
(http://www.pathname.com/fhs/index.html), that tries to set this out.
However,
it is still evolving and is currently not precise enough to predefine
all configuration, web or system files. Perhaps it would be of interest
to you?

To summarize, Linux has a golden opportunity to be the reference
platform for all of Unix, but it must get its act under control quickly.
Standardizing on its file structure could go a long way to achieving
this.

Reg.

P.S. This is a resend of a message that I sent to Tom, but mistakenly
forgot to copy to the list. My apologies to all.

Tom Harincar wrote:

> On Thursday, November 26, 1998 7:21 PM Reginald B. Charney wrote:
>
> > Subject: [svlug] Proposed series of Linux articles in DDJ
> >
> > >I have proposed writing a series of articles on Linux for Dr.
Dobb's
> > >Journal and they are interested. DDJ is the world's largest
subscription
> > >publication for programmer's. My slant is to discuss Linux's
> > >infrastructure and how the Linux programming community handles its
> > >constant technical evolution. For example, I have proposed articles
on
> > >the following:
> > >
> > >* the quality issues that they face
> > >* the size and dispersion of the developers issues
> > >* the different distributions
> > >* dealing with rapid change and stability
> > >
> > >Here is my question:
> > >
> > >What other issues do you believe should be covered in such a
series?
> > >
>
> Hi:
>
> I believe your subject area is very timely and very important.  My
> little operation  develops and markets applications software for the
> traditional UNIX market and (hopefully) for the Linux community.
> However, I am not a programmer.  My interest is in Linux is as a
> business opportunity.
>
> Unfortunately, it seems that we are beginning to see signs that
> Linux, as a technology, is diverging into multiple flavors much the
> same as is the case with UNIX.  By this I mean that third party
> software compiled on a Redhat release, for example, may not run
> without modification on Debian or SuSe - or even worse, that it
> won't run on different releases from the same distributor.  Over the
> years, my company maintained separate releases for many versions of
> UNIX, but we were hoping that would not be the case with Linux.
> Maybe we're being naive.
>
> Even though we statically link, we still run into stupid little
> issues like when a support file is moved to a different directory
> and a sym link is overlooked which cause an inability to install or
> run our programs.  Unfortunately, we lack the resources to maintain
> a test system for every available distribution of Linux.  Note that
> these are not kernel incompatibilities - they just seem to be little
> differences among releases.
>
> I have a special interest in seeing Linux gain acceptance as the
> platform of choice for small to medium size businesses needing a
> better solution for such things as accounting, order processing and
> database applications.  I believe Linux has a chance to overtake SCO
> in this area because Linux is far superior to Open Server in just
> about every respect.  However, the vertical software vendors need to
> be motivated to offer their software on Linux.  Ongoing maintenance
> costs will be a primary conern because their decision will be based
> on whether they can make money with a Linux product offering.
>
> With regard to Microsoft, my view is that many of their problems
> stemmed from the fact that they needed to maintain compatibility
> with the entire spectrum of legacy applications software for many
> years.  We can all criticize Microsoft for a wide range of bonehead
> moves (dll hell for example) but I have to concede that they did a
> pretty good job of maintaining this compatibility.  Moreover, their
> beta programs seem to be a good faith effort to build decent
> software for the masses.
>
> I believe the Linux O/S distributors need to do a better job of
> insuring that ongoing releases maintain compatibility with existing
> applications or run the risk of slowing the growth in commercial
> acceptance of Linux.  They should view the issue as being in their
> own self interest in terms of building the market for their
> products.
>
> Some will accuse me of ignoring the fact that Linux is robust and
> stable because it is constantly evolving and becoming even better.
> I know this to be true but I also know that an operating system, in
> and of itself, has value only if can be used effectively to
> accompish useful work.  In my mind, this means the availability of
> stable applications software.
>
> I will be interested in other views on this subject and your article
> when it is published.
>
> Tom Harincar
> Quadratron Systems
> <tomh at quad.com>




--
echo "unsubscribe svlug" | mail majordomo at svlug.org
^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ to unsubscribe
see http://www.svlug.org/mdstuff/lists.shtml for posting guidelines.