Eric Johnson <eric@qnx.com> wrote:
“Bill Caroselli” <> qtps@earthlink.net> > wrote:
Adam Mallory <> amallory@qnx.com> > wrote:
It’s just that the price is totally unreasonable for an individual.
I completely disagree. The 4K US price tag appear steep, but if you compare it
to what you paid for QNX4, it’s right in line. The equivalent QNX4 development
system may even have been MORE expensive, I haven’t run the actual numbers, but
the ball park ones were:
QNX4 OS runtime $1000
QNX4 Photon runtime 200
QNX4 Voyager 2.02 200
Watcom C 1000
Watcom C++ expansion 200
QNX4 TCP/IP runtime 350
QNX4 TCP/IP development 800
QNX4 Photon development 1500
QNX4 X runtime 500
QNX4 X development 1000
Total 6750
So the costs for an equivalent QNX4 development system are significantly
higher than the costs for the same environment with QNX6.
Especially since all I really want is the mkifs utility.
Bill, I’m sorry that you feel our pricing is out of reach for your purposes.
Obviously you are aware of QNX Momentics NC, which you can use for personal
experimentation. As I know you are also aware, the NC version is missing
some key features (such as mkifs) that would be necessary for ‘real’
embedded development. We found that commercial developers were using the
noncommercial version for commercial development, so we elected to draw a
line and remove functionality such that you could still use the NC version
to develop applications, but would be challenged to do commercial embedded
work.
I think the removal of mkifs was complete bull-shit myself, but I can often
be a crufty old fart. I heard the story a bit differently, and this is why
I have the “BS” opinion. Here is the story I heard:
Several customers were using NC to develop projects, and did most if not all
of the development work with NC. They only contacted QSS once they were
ready to go into production and would typically only purchase ONE development
seat. If they had 20 developers, the QSS sales group saw this as a loss of
19 dev seats, instead of the potential for X runtimes. That aside, they were
also upset that they were “out of the loop” in terms of knowing that company
X was developing a QNX-based project, and they quite simply wanted to have
the visibility of that up front.
Those are valid concerns, but they fail to address the actual way that a
“cheater” company would try to cheat thier way out of buying 20 dev seats.
It is doubtful, that even after being forced (by the removal of mkifs) to
purchase a dev seat that they would ever buy more than one. They could
easily buy one seat, and then copy the missing binaries to the other NC
machines. They could use qnet to make the full dev seat available to all
users, they could run it as a build machine, coding most things on their
NC boxes, and then dong the full / daily builds on the one dev machine.
My point is that the removal of mkifs doesn’t truely address this, and it
never can. It creates more bitching from the community than it’s worth,
since there are many legitimate uses of mkifs that do not pertain to
commercial development (building firewall on floppy for instance).
If QSS thinks the the removal of mkifs REALLY prevents the cheaters from
cheating, they are fooling themselves.
I DO completely agree with the removal of the alternate platforms
from the NC distribution (it was probably a mistake to include it in
the 6.2 NC distribution in the first place). The inclusion of ARM and
the iPaq/eQip reference platform makes sense, since it handily shows the
ease with which cross-platform development can be done.
Personally, I think that mkifs should be in the NC distribution.
Cheers,
Camz.
