.NET Framework - Micro Framework OEM Developing

Asked By RudyBarbier
05-Mar-07 08:03 AM
I would like to develop a new OEM using ARM7 with Microframework, where i can
find the documention for the creation of the connection between hardware and
HAL?

Thanks in advance
--
Rudy Barbieri
EmbeddedFusion
(1)
MS
(1)
MF
(1)
OEM
(1)
OpenNETCF
(1)
CE
(1)
BarbieriRE
(1)
DFDCBE
(1)
  Eric replied...
11-Mar-07 09:42 PM
Other threads here discuss the fact that it's not publicly available and
there's apparently a significant cost involved to obtain this info. This
reminds me of the Microsoft of the 80's and 90's. Just when it looked like
things were opening up another door gets slammed in your face.
  Steve Maillet \(eMVP\) replied...
15-Mar-07 12:33 PM
That's a rather unfair characterization of MS here. The key thing to
understand is " at this point" - that is things will change with time. CE
started out with a $60k buy-in price with headless license costs  of $15.
the Porting process is currently sub-optimal & requires  heavy support From
MS of Based on the work being done now with the early adopters the whole
process is being improved.  Future releases will have broader support.
However, if there are as many OEMs porting the MF as there are doing CE,
then MS has seriously failed on the major design goal of the
Micro -Framework. The biggest blocking point for that goal at this point is
a lack of support for  memory mapped devices. Once a managed code solution
exists  for that there's really nothing you can't do in managed code.

--
Steve Maillet
EmbeddedFusion
www.EmbeddedFusion.com
smaillet at EmbeddedFusion dot com
  Eric replied...
17-Mar-07 02:15 AM
The lack of managed user classes for USB and Ethernet is also a big deal. It
seems that most of the projects for Arm9 devices will need at least one of
these two.

But the overriding constraint is the high cost of getting into the game.
Many of us embedded software authors are contractors who work with hardware
makers. The companies we work with usually want to use their own hardware
designs. They wouldn't even consider buying ready-made boards from someone
else who already did the porting effort. Hardware is their business, after
all.

That would leave people like me with the high cost of porting the MF to
their unique hardware, and that cost would have to be included up front at
the beginning of the project. Since I normally don't get paid a commission on
the number of boards they make, but only get paid a negotiated amount for
writing the software, that extra MF porting cost would cause my bid to be
very high and I wouldn't be able to compete with other consultants.

If MS were to agree to make their money only on a production basis, that
would eliminate the up-front costs that stop me from playing the game. But
the royalty would have to be reasonable. My gut feeling is that the royalty
per production board would have to be under $1 in order to get it approved
without a lot of fanfare. Anything over $2 a board would be out of the
question, except for low quantity projects. You have to understand how much
effort they expend to shave production costs. These are the people who try to
do without an extra resister to save money!

Eric
 
18-Mar-07 01:53 PM
As Steve points out, this is version 1.  The support cost for the MF team if
they just allowed anyone to get and use the tools would probably sink the
entire product.  It's just not yet at a state that just anyone could get the
tool and port the platform to any random hardware.  For now the buy-in
really is a filter to incent serious people with an active production
project in the pipeline to move forward.

These initial customers and their experience will be what drives the
direction of the platform and the tools and I would expect to see future
versions become easier to use.

If you look at Windows CE as a model, they're on Version 6.0 now and it's
still not a super-simple thing for just anyone to get up and running.  It
didn't even get reasonably straightforward for experienced developers until
the 3.0/4.0 time frame.

I expect to see the MF to become something that can be moved to a set of
common cores and processors as time moves forward.  If your customer is
using one of those processors, it will likely be a fairly easy job to get
the device up.  How long it will take to get there one can only guess.

--
Chris Tacke - Embedded MVP
OpenNETCF Consulting
Managed Code in the Embedded World
www.opennetcf.com
--
Create New Account
help
http: / / vadmyst.blogspot.com You wrote on Thu, 1 Nov 2007 12:45:00 -0700: MS> Hi, MS> I'm trying to build a robust application framework that will be used MS> by other developers in the future. Their applications should be MS> able to recover themselves if they are crashing MS> So, I thought, the best way to handle this is to lauch their stuff MS> in a separate AppDomain and when that AppDomain crashes, unload it MS> and restart it. MS> This seems to work fine, unless an unhandled exception occurs in a MS> background thread
1$1$1$1$1$S, S, S, S, S, S, S, S, S, S, MS, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, MS, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S S, S, S, S, S, S, S, S, S, S, S, S, S, S, MS, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, MS, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S S, S, S, S, S, S, S, S, S, S, S, S, S, S, MS, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, MS, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S S, S, S, S, S, S, S, S, S, S, S, S, S, S, MS, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S$1, S, S, S, S, S, S, S, S, S, MS, S, S, S, S, S, S, S, S, S, S, S, S, S, S, S
zum Spam sogar mit ein Grund für die Verlagerung: Newsgroups sind mehr oder weniger werbefrei. MS will aber, denke ich, den Service, den es mit den Diskussionsgruppen bietet, in sein eigenes Konzept der Werbung für MS- und Partnerprodukte einbinden - schließlich ist eine Diskussionsgruppe ein idealer Ort dafür. Und mit den Links von der Hand zu weisen; aber auch die häufige und nicht gerade harmlose Kritik an MS scheint mir dabei eine erhebliche Rolle zu spielen - Bessere und erleichterte Ausschöpfung der Inhalte; schließlich stellen die Teilnehmer der Diskussiongruppen eine große Gemeinde an Betatestern dar. Die Teilnehmer sollen dabei MS auch noch die Arbeit erleichtern, indem sie beurteilen, wie hilfreich Beiträge sind. Außerdem habe ich den Verdacht, daß MS mit dieser Verlagerung auch das MVP-Programm reduzieren oder sogar einfrieren will - wie es schon erfolgen und bis Herbst 2010 abgeschlossen sein. Den Anmerkungen ?ber fremdsprachliche Newsgroups entnehme ich dass MS an die zuletzt gehen will. Mal sehen wann wir Deutschen wirklich dran sind. Ich habe comp.lang.misc microsoft.public.de.vb.datenbank - -> de.comp.datenbanken.misc - -> de.comp.datenbanken.ms-access - -> de.comp.datenbanken.mysql Wer bisher nur ueber msnews.microsoft.com seine Gruppen bezogen traffic' ist aber nicht verschwunden sondern hat sich aufgrund der Aufgabe von VB 'classic' durch MS nur in die VB.NET-Newsgroup verlagert. Aber 'mal davon abgesehen: Was soll das Argument
are you applying to determine that FPGA is the best choice? Steve Maillet http: / / www.EmbeddedFusion.com It is interesting. There was discussions about the integrated CPU+FPGA for Windows CE solutions are more for high end multiprocessing system, which might be over kill for .NET MF. The idea of having a progammable logic with the processor was designed into the iPac with the ARM Cortex M1 or M3 in it would be quite nice for the MF. (Assuming that a port to the Cortex Thumb2 instruction set is made available from the nice folks in Redmond, of course! [ To MS: Hint, Hint, nudge nudge! ;-) ]) Combine that with the continuous improvements in programmable analog arrays and but, hey who knows what time will bring on that front. . .) Steve Maillet http: / / www.EmbeddedFusion.com Hi! I just feel need to second the hint hint nudge nudge thing! ;-) I would be very interested in having MF on FPGA. Have a nice day, Jan There is a good reason to support FPGA elaborate on those statements a bit. (e.g. why does one need an FPGA or MF support processor cores in FPGAs to be able to decompose a system into a sensor the possibilities and will support requests to add the capabilities to the Micro framework. (certainly MS CAN do it, but whether they do it or not and when is a numbers
SPI Support in Net MF and how to build your own? .NET Framework Hi there, Unfortunately it's been a while since I last had the time to work on our new developments with net MF due to other company priorities, but as of now i'm picking up where i off. For our new development i'm still looking for SPI support in the Net MF ports. We still haven't decided yet with which platform to go (e.g. Digi or EmbeddedFusion etc.), simply because we still cannot oversee all the features and most important the support of these features in Net MF. SPI in particular. So we have this SPI chip we need to hookup to the microcontroller and write code for it in Net MF. Now is there already SPI support build-in in the different MF ports out there? And if there isn't, is there a way to build our this is actually still a bit unclear to me. Do i need to be a MS partner for example to write drivers? Do i need to have the porting kit then