Hey folks,

This is a great idea and I would love helping to support it.

I would just like to make a few comments regarding the feedback that I have
received along with some of the discoveries that I have made.

1.  Bound the operating systems you will support very strongly.  I'm saying
this because Windows 95/98 were very easy to work with, Windows NT/2000 were
more difficult, but still similar and Windows XP is different to the
previous ones and you cannot count on the APIs used previously to be
available in XP.  Unfortunately, you can bound your software, but Microsoft
will proceed as they desire without any regard for anyone else.  Ideally,
the programmer software would work on Linux and the Mac.

2.  The software should be written in C/C++ (I'll refrain from specifying
which brand) and should produce native code.  This will give the maximum
amount of flexibility for different operating systems and avoid the need for
different device drivers (like Visual Basic).  Personally, I would push for
C++ to allow dialog box control of the programmer.

3.  I would like to see the programmer software would interface directly to
the MPLAB IDE.

4.  I can't believe what kind of variability there is in parallel ports.  I
have really done a lot to support all the different ones out there and it is
an daunting undertaking.  The differences are based in logic I/O and
timings.  This can be mitigated by using the standard parallel port device
drivers built into Windows, but it is challenging to choose a driver (with a
simulated printer) that will allow comprehensive bi-directional
communications.  The basic drivers will just round up all the handshaking
bits and report the printer is out of paper.

Right now, after my experience with the "El Cheapo" and the hours I've put
in trying to find solutions that will work for *everyone* and not take away
from people that already have working programmers, I probably will not
design another parallel port programmer.

5.  Serial ports are better, but I just discovered that very few new laptops
have serial ports built into them.  There are USB to RS-232 converters and
these work very well, but I want to experiment with the USB to Palm Pilot
Serial converters as these are cheaper and would probably be ideal for an
RS-232 based programmer.

6.  New users are looking to use the PIC16F62x and PIC16F87x, not the
PIC16F84.  Yes, the PIC16F84 is simpler, but it doesn't offer the advantages
of the other parts, one of those advantages being LVP.  Supporting just
these parts with LVP will make the power supply issues for the programmer
simpler.

7.  Another advantage to the PIC16F87x is the In Circuit Debugger (ICD).  I
think it would be increadibly useful for new users (as well as experts) to
be able to see how their applications are working.


Looking back over these points, I would end up saying that Microchip's
MPLAB-ICD just about fits the bill perfectly.  The issues I have with it
are:
a) Cost
b) Speed (Single stepping is too painful to even be considered)
c) Lack of programming support for other products (ie the PIC16F62x)
d) Serial Port interface (see above), while not a *big* issue it can be a
problem for people that have PCs without serial ports.  A USB interface
would be preferable - the MPLAB-ICD 2 has a USB interface (and I believe can
be used with the PIC16F87x as well as the PIC18F4x2 - can anybody confirm)?
If you were to come up with a version of MPLAB-ICD that eliminated these
issues, then you'd have something that is perfect for just about everybody.


Looking over the list of questions I have gotten this year, programming is
not the number one problem for most new users.  The list of questions I have
gotten from new users are:

1.  Power supplies (Do I have to buy a bench supply, How do I use a 78(L)05,
can I use batteries).  I would recommend that the beginner's kit programmer
has enough current left over to power most applications.

2.  Oscillators (don't run.)  Ideally, I would like to see the PICmicro MCU
built into the programmer and the entire assembly being being inserted into
the application, with the clock built in and providing current to the
external devices.

3.  Programming.  Where do they find one, which one is best.

4.  How does the individual get {Insert project available on the Internet
here} to run.  Many projects available on the Internet or written up in
different books are designed for obsolete parts or low-level (PIC16F5x)
parts which are not programmable by ICSP programmers.


These are my opinions, based on the feedback I have been receiving and my
own experiences.  I'm curious to see what everyone else thinks.

I honestly do believe that a beginner's kit would be a useful thing and
something I would be happy to help on and support.  I just want to make sure
that the problems I experienced with the El Cheapo are avoided and people do
not get locked into a mode of perpetually supporting something they don't
make any money off of.

myke

--
http://www.piclist.com hint: To leave the PICList
mailto:piclist-unsubscribe-request@mitvma.mit.edu