“Haris Blentic” <hblentic@qnx.com> wrote in message
news:9v5tgp$gsj$2@nntp.qnx.com…
QNeXperts,
I’ve begun putting plans together for the next years’ QNX Developers
Conference. To make sure we deliver the content most appropriate to your
needs, I’m going straight to the source and asking you to let me know what
you’d like to see covered. Plus, for those after-conference activities,
you
also get to tell me where you’d like to see the conference
held…(Montreal/San Diego/Las Vegas/?)
What is timeframe? I’ve never been to any of those places, so I would not
care much but season-wise considerations might apply. Las Vegas sounds like
fun and it might be cheaper to attend with all those deals targeted at
hard-core gamblers
Below are some general topics - beside each, please give me an idea of how
important you think the topic is by giving each a grade from 1-5, with 5
being the most important. Please feel free to add to the list as well and
expand on any of the topics…Keep in mind there will be a ‘new user’ and
an
‘expert user’ track and the conference would be 2.5 days in length - the
content will also be as hands-on as possible.
If you also would like to participate in a conference advisory group, send
me an email - > hblentic@qnx.com > and let me know. This way we can work
closely together to create a true developer conference…
some topics:
Embedding Issues - IPL, BSPs,startup
That should get 5. And have few PDA-oriented examples to get people started
with toys they bring with them
Code development/debug on QNX
Vague.
Architecting Applications Using QNX
It is 5 for newbies and 1 for seasoned people
Cross development
This is very important to some people and not important at all to others.
I’d mark is as 4, but again we might mean different things. I think to make
it most interesting it should be about creating custom cross-environment,
not using existing ones. Remote debugging is quite interesting subject too
which involves many implications such as visibility of source for debugging
(off version control systems).
PhAB and User Interfaces
GUI style guidelines would be nice. That has little to do with PhAB (imho).
I say 4.
Writing Portable POSIX Apps.
This is interesting subject if you really dig into it and rather dull if you
make it shallow. Covering real life examples/caveats of porting between QNX
and couple of most widely used Unixes would be much better than just
abstract sentiments which everyone can read in POSIX book himself. Solaris
and Linux are good choices I think. It is probably 3.
DDKs (network, audio,MOST)
This is 5 of course. I’d really appreciate if the focus was made on
approaches allowing to use open-source driver sources as ‘documentation’. Of
course DDKs were not designed exactly for that, but in real life people
using them often don’t have specs and often there’s a free driver source for
another OS.
To make it clear what I mean, I’d like to see an example of ‘mapping’ of a
driver in each category from Linux to QNX. Some of them map quite easily
(like audio) and I’ve done that myself. Some might be harder, but in any
case there’s more or less stable framework for each category. Such framework
includes general organisation of driver, principal data structures, etc. For
someone experienced in certain class of hardware and both Linux & QNX
drivers of that class, it would be relatively easy to show how to ‘map’ such
structures, low-level routines and other elements. If there was an example
of doing that with Xfree video driver, I’d write Photon driver for my laptop
(and share it with others, of course).
HAT
HAT does not exist. Well it does, but for very small number of people. And
it does not do much of what one would need (for me anyway). This subject
really requires something more like a brainstorming than a workshop. QNX
architecture is lacking severely in area of PCI dynamic configuration
support and high-availability storage. It would be nice to discuss that with
interested people. To be more specific:
- there is no way to discover a PCI device inserted into CompactPCI chassis,
if carrier card has PCI-to-PCI bridge (which most of them do), because
there’s no concept of ‘PCI chipset driver’ in QNX;
- there’s no support for hot-swapable hard drives;
- there’s no support for RAID;
This has rating 5 in my priorities. Not many OSes have satisfactory
solutions, it would be fun to make QNX the one. But if all they gonna tell
is how you can nicely monitor/respaws processes, that’s old news too. We had
such solution long before HAT.
SAT
That will probably be most exciting new feature, except as with HAT not many
people will have their hands on it. It is 4 for me.
Migration to QNX 6
Migration from what? If from QNX4, that’s old news, 1
Netcom development - IONet,Qnet,SMP
Cool, but out of my field, 1
Target HW Considerations for QNX porting
Vague.
I’d say that given plans with Eclipse, the list is sorely missing
Java-related stuff. Developiong Eclipse plugins for QNX would be very
interesting addition. Progress/status on realtime Java specs/implementation
is also very interesting.
Another thing is doing custom installations. OSes like WinNT and Solaris
have specially designed tools and procedures for doing that. For QNX we had
to develop our own and so far we see each new major release being
progressively harder to manage in terms of field installations. What is
needed is hands-free automated solutions with parameters/feature sets
selectable through floppy or network server.
regards,