On Tue, 29 Sep 2015 20:38:33 +0300
Reco <recovery...@gmail.com> wrote:
> On Tue, Sep 29, 2015 at 07:03:57PM +0300, Eliezer Croitoru wrote:
> > +1
> >
> > * And if indeed the basic requirement from a system operator would
> > be "a programmer" then it would suck to be a sysadmin for many.
>
> -1.
>
> Being a sysadmin requires task automation (among other things), and
> that inevitably boils down to some programming skills at least.
>
> Inability to grasp basics of programming can produce only a helpdesk
> drone, not a real specialist.
>
>
There's programming and programming. Altering a modern multi-tasking
GUI-based application is *not* the same thing as stringing together a
grep with a cut or two in a three-line text file.
Neither are we suggesting, I hope, that the only Linux users are
sysadmins. A typical Linux user might be an engineer or other small
businessman, or a child web surfing on Ubuntu.
I might be willing to have a go at making a small, well-defined change
in an application if the need were great enough, but I'm under no
illusion as to the hours or days it might take me to even locate the
correct source file to do it in. I'm not scared of it, almost the first
thing I did with Linux was to recompile a kernel for something like Red
Hat 5.2 (and I don't mean Fedora), but I have some idea of the
magnitude of the task. I occasionally need to revisit one of the PIC
projects I've done over the last seven or eight years, and I know how
long it takes me to get comfortable with *my* *own* *code* written more
than a year ago, for a few kilobytes of object code.
Altering part of the operating system ('if you don't like my systemd,
write your own...') would be orders of magnitude more time-consuming,
unless you already happen to be fairly familiar with the operation of
the OS, which is unlikely unless you're a professional in that area.
I like the *idea* that you can alter things, but I'm aware that in
practice it doesn't make much difference to the average user. Far more
important, it seems to me, is that it is free as in beer. I don't mean
the issue of actually paying for it, I mean that paid-for software needs
to be protected against theft, and this weighs heavily on both the
design and implementation of the software, and particularly on how well
it plays with other software. You are very conscious that the primary
objective of the programmer(s) is not to produce the best software, but
the most profitable software.
Naming no names, but I recently spent about eight hours trying to
install a database application on an operating system made by the same
company, on a business desktop linked to a corporate VPN, fitted with
unknown web filtering and proxying. Despite installing from DVD, this
wretched program needed to talk to its mothership during installation,
and was not at all happy with the networking facilities it had to use.
I eventually had to completely remove the anti-virus software, drop the
VPN, and install it using a local Internet connection, completely
contrary to corporate policy. It worked, but what a waste of time... I
could have downloaded and installed a free integrated LAMP system in
about fifteen minutes, but of course that wouldn't have done the job in
hand.
And the other really big difference between the free and non-free
worlds is that writers of free software are (so far, mostly) not under
the impression that they own your computer and are free to do as they
like with your Internet connection and your personal data...
--
Joe
