In comp.os.linux.advocacy, Roy Schestowitz
<newsgroups@xxxxxxxxxxxxxxx>
wrote
on Fri, 25 Apr 2008 07:28:17 +0100
<1729180.4xVQiU5RiM@xxxxxxxxxxxxxxx>:
> 10 ways to sell corporate on Linux
>
> ,----[ Quote ]
> | #1: TCO is bunk
> |
> | [...]
TCO is *not* bunk, but the way Microsoft calculates it
probably is. There's a lot of factors, not the least of
which is that sickening feeling one gets when a production
server's gone down. Again.
How does one put a price tag on that?
Then there was the individual who was running an important
rendering job which died when Windows decided it really
needed to upgrade and reboot itself. While one might
assign partial culpability to the system administrator
(was there a way to disable that reboot?), unscheduled
auto-reboots certainly do not help production schedules.
> |
> | #2: Linux is not just for servers
> |
> | [...]
Linux can be used for a basis of just about anything
nowadays, given sufficient engineering work. It can form
the kernel of mobile phones (Google's Android project)
all the way to hyperpowerful servers (IBM z-series,
various multinode supercomputer projects), with desktops,
phones and other such in between.
> |
> | #3: Security is the name of the game
> |
> | [...]
Security is but one objective of many, albeit a rather
important one. My worry, admittedly, is that NAT renders
many objections against *Windows* obsolete, by eliminating
one infection vector. Good for the consumer, especially
if he's using Windows; not necessarily so good for Linux.
Of course there are a few other vectors, such as upstream
DNS hijacking (Linux doesn't defend too well against that;
nothing really can!), localhost hijacking (/etc/hosts is
very well defended in Linux; Windows, not so much),
browser hijacking (Linux can't do a lot about that though
Mozilla runs in userspace at least, but so does IE if one
uses the ies4linux/wine combination), and other such niceties.
> |
> | #4: Support is everywhere
> |
> | [...]
Support is IMO a little iffy in Linux, mostly because the
codebase mutates so much. I'm still trying to figure out
Xorg 1.4, for example, which has a somewhat different approach
to the multiple-monitor problem (an issue for those of us
with laptops and notebooks and the occasional dual-head
video card owner). The problem is exacerbated by what
appears to be an ATI bug in the new driver, as well; the
external monitor simply won't fire up. (Under Xorg 1.3,
it works fine.)
ALSA is working on my laptop but another point of iffiness.
For the most part, it works fine, but Rosegarden is a more
sophisticated client than Netscape's Flash plugin.
For completeness:
#5 Applications are key.
I'll admit I'd have to research this, but a lot of custom
code's been written out there -- and of course Linux and
even more so Linux-based distros have a lot of freeware,
but that freeware may need tweaking. Not sure how much of
that custom code is standards-compliant; presumably a lot
of said code uses MFC, because that's what the writer was
comfortable with at the time of its initial development.
Yuck.
#6 Kernel is just for you.
The kernel is for the system administrator. Granted, for
my personal hardware and for some of my business systems
(the ones on my desktop!), I can certainly make appropriate
modifications.
#7 Virtualization is virtually everything.
Virtualization has been around since DOS days. The term
"VAX", after all, means "virtual address extension"; it
was one of the first computers to implement page-based
translation in what passed for its MMU. The 286 variant
of the PC was born virtual, though that conventional
memory thing still haunts us today. (Granted, in Linux
one doesn't notice it for very long, but GRUB probably
has to deal with a quirk or two the PC throws its way.)
Of course nowadays we can virtualize more than RAM, which
is a good thing; VmWare and QEMU allow for an environment
where the program within said environment thinks it's
talking to a raster display; the surrounding emulator
takes care of the rest. Ultimately, I for one hope that
the emulator can start emulating certain card registers
as well, allowing for a OpenGL-within-a-window effect.
(Or perhaps DirectX, should one really need That Other
Gaming Solution(tm). Bleh. At the video card level,
though, I doubt it makes all that much difference.)
#8 Updating is simple and fast.
Depends on the update. For most, it's a simple matter of
downloading a new package or tweaking a config file.
However, the Xfree86->Xorg transition was painful for me,
though not insurmountably so. Of course the XP->Vista
transition appears to be bothering many a user.
#9 Administration is world wide.
Non-issue in many respects. Windows, after all, can
often be remotely administered with additional hardware
(powering the machine down if absolutely necessary),
and either rdesktop or vnc. Of course, Linux can work
with less hardware; ssh and a network connection is more
than sufficient for all but aforementioned powerdown,
and of course Unix before Linux was mostly text-based,
which means less bandwidth is needed -- the word "hello"
takes only 40 bits in ASCII, but might require a minimum
of 320 pels, each pel requiring 3 bytes (1 each for red,
green, and blue), for a total of 960 bytes in a graphical
context -- and that's assuming an 8x8 font, which is a
little small nowadays, and non-transparency (transparency
can be implemented using nothing but a "color zero" convention,
or add a full extra byte per pel for alpha).
#10 Linux is constantly gaining traction.
It certainly is getting a fair amount of positive press
(as well as some negative press, courtesy of various
Windows-centric stuff), which is good to see.
> `----
>
> http://blogs.techrepublic.com.com/10things/?p=342
>
>
> Recent:
>
> PC deal could save public sector billions
>
> ,----[ Quote ]
> | Suffolk told Gartner, ?I think we have fundamentally
> | failed on a worldwide basis as an IT industry to
> | understand the cost of what we do. And I roundly
> | blame Gartner for this, because you guys are the ones
> | who come up with TCO [total cost of ownership] benchmarking.
> | It has become a self-fulfilling prophecy.
> |
> | ?So, I go out and I pick boring desktop infrastructure.
> | What price do you think the suppliers broadly pitch? You will
> | not be shocked to know that it is somewhere around the
> | Gartner TCO benchmark.?
> `----
>
> http://www.computerweekly.com/Articles/2008/03/31/230047/pc-deal-could-save-public-sector-billions.htm
We'll see. Certainly it's nice to see the UK government
hewing to internationally-recognized standards, which
will probably save them billions (in rework time) *after*
the initial conversion, and allow customers (citizenry)
to save some money as well, by not constraining their
options to a single vendor.
Ideally government data would be available on everything
from phones to kiosks.
--
#191, ewill3@xxxxxxxxxxxxx
Useless C++ Programming Idea #8830129:
std::set<...> v; for(..:iterator i = v.begin(); i != v.end(); i++)
if(*i == thing) {...}
--
Posted via a free Usenet account from http://www.teranews.com
|
|
