Paradigm shift going from Gnome2 to Gnome3

[夜神 岩男]

On Tue, 2011-06-21 at 08:28 -0400, Darryl L. Pierce wrote:
> On Tue, Jun 21, 2011 at 02:23:51AM +0930, Tim wrote:
> > Is a corporation going to want to spend $100 per graphics card per PC,
> > so that the default Gnome 3 actually works, or are they going to
> > continue to only want to put in the $20 graphics card?  (That just won't
> > work with the new all-singing, all-dancing, Gnome 3.)  So there's the
> > next RHEL with Gnome shot down in flames.
> 
> Corporate and home users are already being expected (by that company in
> Redmond) to upgrade to higher end hardware for their latest offerings as
> well. The difference is that Fedora has fallback functionality that
> works without the hardware acceleration, as well as other options for
> the desktop.

My customers (small and mid-sized businesses in Japan) are interested in
Linux specifically because they can cheaply clean and refurbish old
desktop hardware and save tons of money in the office. Telling them that
the next big thing on the Linux desktop is a cumbersome beast which
requires high end hardware the in a similar way Windows does tips the
balance away from cheap (read as "harmless" or "low overhead")
experimentation with Linux and back towards just sticking with Windows.
And let's face it, any smart company plays with a future platform before
they commit, so the cost of experimentation is a significant point to
consider.

The limitation on Linux desktop adoption is not technical and it is not
usability -- it is almost solely a lack of promotion and sales knowhow
with regard to the non-technical business world. Gnome3 looks beautiful
-- and this is overwhelmingly important to a customer who has no clue
what you're talking about if you get to technicals. They understand
pretty pixels and have come to expect them because that is their level
of comprehension of the system (especially in the course of a one-hour
executive brief). So Gnome3 = pretty, that's wonderful. On the other
hand, Gnome2 is functional and can be made beautiful as well and does
not require acceleration or hardware upgrades vs WindowsXP. After the
executive brief or sales pitch, when numbers and projections start
getting discussed in private between the boss and his accountants, the
strength of the Linux position is significantly weakened if Gnome3 is
being discussed as opposed to Gnome2 or XFCE based purely on hardware
cost. (As a side note, some frequent-use Gnome3 keyboard shortcut
functions are not possible with one hand on a Japanese keyboard which
changes the picture a bit there as well.)

I am generating a significant amount of interest locally primarily
because everyone is still using legacy WindowsXP applications that don't
run on Windows 7 or Vista. Most of these small and medium sized
companies have custom coded applications which will cost a huge amount
to port forward to the new Windows environments and they are facing
forced hardware upgrades associated with the upcoming WindowsXP EOL.
They must change. There is no choice because the platform they are
invested in dies very soon. The learning curve and investment required
to move from XP to 7 is roughly the same as moving to Gnome2, but the
safety inherent in the longer-term stability and self-support potential
of Linux makes them feel protected against the possibility of a vendor
arbitrarily forcing them to blow tons of money on new hardware, forward
porting code and retraining the way the XP EOL is (this is not nearly as
decisive a factor as the near-term effect of offering a low-cost
alternative, as discussed below).

A Gnome2/XFCE offering fits with their current situation of being in
possession of 5-10 year old hardware in the office that we can suddenly
breathe new life into with minor component upgrades and refurbishment.
With the leftover money in the budget they can splurge on enormous
desktop displays and nice peripherals for everyone. (...which they are
just realizing for the first time has an impact on productivity whereas
just having a faster computer does nothing for the average office worker
who has seen no truly revolutionary enhancements in office productivity
software since the days of OS/2 and Windows98 despite a string of OS and
software suite upgrades and associated licensing fee and hardware
expenses.)

The weightiest argument I can bring to bear, particularly in an
uncertain economic environment, value I can provide by leveraging old
hardware. Convincing them that Linux is a reasonable desktop solution
for the office does not require weird desktop features, virtualization
technologies and an Android-ish 3D interface (on a non-touch screen, no
less?). XFCE is plenty to get them excited, actually, because it lets
them do the things they need for work without spending anything more
than a few afternoons poking around the system to self-train on basic
tasks (our average customer adjusts to Gnome2 and XFCE faster than OS X,
and moving to OS X isn't that hard, either). More complex tasks like ERP
customization and application porting are naturally a bit expensive, but
the expenses would have been about the same moving to Windows7 or to
Linux. The two big short/mid-term selling points come down to license
fee expenses* and savings on hardware costs. Removing one of them
significantly reduces the strength of the argument in favor of adoptiong
of a new, unknown, scary operating system.
(*Windows tech support companies charge about what we do for Linux
support, so there isn't much difference here.)

The security improvement arguments and technical merits of Linux are
unarguable to people who spend their days fixing cars or taking pictures
or writing contracts or managing construction projects instead of
building software. These sort of benefits only become apparent after a
sale is made, never before. It is a pleasant surprise and almost
guarantees that you won't lose the account in the future, but it is not
something you can really argue during the sale because every software
vendor says they have good security and great features (we are the only
ones who admit to bugs, though, which actually seems to help the sale).
The business security and position of power the customer inherits with
open source is a long-term concern and one that becomes apparent very
long after the sale, and while worth mentioning during a pitch, it is
merely something which lends a bright atmosphere to the conversation and
does nothing to alter what is actually being discussed (the
short/mid-term economics of migration).

The financial argument is where we win big on office desktop sales to
small and medium sized businesses (not big business or backend, that's
RedHat's thing anyway and they do well at it) and upping the hardware
requirements could act to significantly reduce one of the two
irrisitible arguments we can bring to bear in sales efforts.

Obviously, as my company grows and more of our developer time winds up
finding its way into projects like Fedora and upstream components our
interest will, at least across the mid-term, be in either adding further
polish to lighter-weight DEs or reducing the weight of DEs such as
Gnome3. This is in our interest even at the expense of coolness --
though considering that there are so many layers to the Gnome3 cake that
describing them as "indirection" as opposed to "abstraction" is perhaps
more appropriate, there is likely a huge potential for speed enhancement
and refinement in the future.

-Iwao

PS to all: A careful re-reading of the above blather hints at many of
the real reasons why Linux hasn't made much headway on the desktop yet
-- it has very little to do with the user experience or technicals and
everything to do with the FOSS community's almost universal ineptitude
at selling to non-technical people. Internally we call this having an
"RMS問題" or "Stallman Problem".