Onestone <onestone@...>:

> and the crystal
> tolerance is very liberal for this part

I had this problem with my 24L01 experimental design some years ago. The
crystal-freq was off range at hot tmeperatures. The crystal "from stock" I
was using was not sufficient.

M.

We have a product that we sell in the US that we are expanding into

European markets.  To do this, we need MID approval.  One of the

requirements is that our software revision number be AUTOMATICALLY and

INEXTRICABLY linked to the software.  That is, if the software changes,

the version number must change, and if the version number changes, the

software must be different, and of course, identical software should

produce the same version ID.  They suggest some pairing of version

number and checksum.

Since we load the software via bootstrap loader, I am considering having

the loader calculate a checksum for the ti.txt file as it loads.  I

would trap for the "q" entry in the file marking the end, then have the

loader place the checksum in a pre-determined memory location that is

always cleared to FFFF, then finish it's loader routines.

We are using the MSP430F5438 chip, IAR full, MSP-FET430UIF, and a

modified TI bootloader.  Does this scheme seem feasible?  Has anyone

done it?  Is there a better way?
Thanks Guys and Girls,
Mike Raines

> We have a product that we sell in the US that we are expanding into
>
> European markets.  To do this, we need MID approval.  One of the
>
> requirements is that our software revision number be AUTOMATICALLY and
>
> INEXTRICABLY linked to the software.  That is, if the software changes,
>
> the version number must change, and if the version number changes, the
>
> software must be different, and of course, identical software should
>
> produce the same version ID.  They suggest some pairing of version
>
> number and checksum.
>
> Since we load the software via bootstrap loader, I am considering having
>
> the loader calculate a checksum for the ti.txt file as it loads.  I
>
> would trap for the "q" entry in the file marking the end, then have the
>
> loader place the checksum in a pre-determined memory location that is
>
> always cleared to FFFF, then finish it's loader routines.
>
> We are using the MSP430F5438 chip, IAR full, MSP-FET430UIF, and a
>
> modified TI bootloader.  Does this scheme seem feasible?  Has anyone
>
> done it?  Is there a better way?
> Thanks Guys and Girls,
> Mike Raines

We happen to use the Subversion version number of our repository and
integrate that into our build process.  Works for us as we use Subversion,
if you use a different version control system, well, maybe not so simple.

-- Paul.

make a build script that compiles everything, grab the current date and time
at the start of the build script, pass it in as a define.


>We have a product that we sell in the US that we are expanding into

MCU: MSP430F5437A
IDE: Crossworks
Debugger: MSP-FET430UIF

Has anyone seen the assembly code is completely different from C code?  Here the
main() module starts at address 0x0110C4 as shown in the call stack.  And the
line to disable the watchdog timer: WDTCTL = WDTPW + WDTHOLD.  However, if you
look at the assembly code at address 0x110CA, it does not correspond to
disabling the watchdog timer.
Regards,
Louis

Another comment I like to make is that my program runs correctly if I don't
attach it to the debugger.  That is after programming the flash and disconnect
from the debugger, the program runs fine (confirm visually via LEDs & LCD
display).  The program crashes when I attempt to run attaching to the debugger &
single step.  Any input is greatly appreciated.  Thank you for your time.

[Non-text portions of this message have been removed]

Hi ANOOP,

It is possible...you can write your own bootstrap loader and then have it
flashed to an area that is reserved for the code. TI has a very good document
for this SLAA450. There is also numerus threds in the forums also going over the
procedure.

onebusyoldman




________________________________
From: anoop gs <gs_anoops@...>
To: msp430@yahoogroups.com
Sent: Sun, May 29, 2011 1:04:22 PM
Subject: [msp430] BootStrapLoader

 
hi all,

             I am new in this group and working in a central govt
firm.I got a
new task of controlling an msp430f5438 microcontroller.In my proceed i  get
into
a new issue of booting
msp430.In DSP we can give GPIO confugaration to boot from  spi/i2c/mcbsp
etc.But
these are not found in msp430.Then i came to some other features like
BootStrapLoader as i am new i want to know some more details about it.My aim is
to load code/hex file from EEPROM in my i2c interface.Is it possible and how can
it be done.

regards,
ANOOP G.S.

 

[Non-text portions of this message have been removed]




[Non-text portions of this message have been removed]

http://cnx.org/content/m23503/latest/
BIS(.B or .W) src,dst   src.or.dst��dst   Set bits in destination

BIS = OR
But none of the status bits are set.
And that is no big deal.


On the parallax prop mcu an OR can have these result (but is optional)
Z is set if Result = 0
C = Parity of Result

I have used an IR LED and a Photo Transistor to make a slot sensor on the msp430

The trick is to make the photo transistor sink the signal to ground
when it's on.

Set the msp430 pin to input, you could use the programble pull-up resistor or
use an external 10k to v3.3

So the signal will be a 0 if photo transistor is detecting light.


          3.3v
           |
         [10k]
           |
430<----------[photo transistor]
                       |
                      GND

http://processors.wiki.ti.com/index.php/MSP430_Flasher_-_Command_Line_Programmer

create a BAT file, they show an example bat

How would I use __TIME__ to set my 48bit counter at time of compiling/flashing
the msp430

Keep in mind I'm a assembler programmer.

I'm reserving 3 registers or maybe 6bytes of RAM for my counter,
as 32bit could result in the Y2K38 problem.

somethink like this:
mov.w   (utime>>16),R15        ; Upper Unix time 16bits
mov.w   (utime&FFh),R14        ; Lower Unix time 16bits

The Launchpad comes with a built-in application that displays
a single character every second or so once S2 is pressed to
start it in that direction.  Pressing S2 again should display
two characters and then one every second.

While it works on EVERY SINGLE OTHER system in my house,
there is one which does not seem to work over the serial
port.  It's a brand new machine, an Intel 2600K based system,
running Windows 7 Ultimate, 64-bit.  I did all the installs
and hardware build and everything is up to date with
Microsoft.

I've tried using a variety of COM port monitoring software --
all of which works with Launchpad just fine on other
computers.  But on this new system, it displays "X0" every
time I press S2 (which is wrong, anyway) and displays nothing
else, otherwise.  I've tried all this on another Windows 7,
64-bit machine, which is running the Professional version and
not Ultimate, and it works just fine.  Same device, same
cables, nearly the same operating system.  And the serial
port driver is the same version on these two, 1.3.0.0, from
TI, dated in 2007.

I've tried even using the CCS compiler IDE's "Terminal" pane
with the same results.  Works great on every other machine.
Fails on the new one, in the sense that I see the same "X0"
when pressing S2 and nothing otherwise.

I've also tried downloading code into devices on this new
machine.  It works, intermittently.  I'll get an error from
downloading without debugging, then try downloading WITH
debugging and it works great.  Then another time it will fail
that way, and succeed the opposite way.  Sometimes, nothing
works until I unplug and then replug things in.

I've disabled User Access Controls, and enabled them.  Makes
no difference at all.  I've tried running as administrator,
or not.  Same results.  I've tried many different serial port
monitoring programs.  Except for some oddities in the
Terminal on CCS which probably have their own explanation,
everything works/fails in the same ways.

I've no idea what else to check.  The IAR installation I used
for downloading is 5.20.4 and installed on May 9th, if I
recall.  But even a variety of independent serial port
monitoring programs fail out on the new machine.

I've also tried different bit rate settings, though I am
supposed to use 2400.  Odd thing is, on the failing machine,
the response is the same regardless of how I set up the bit
rate: "X0".  Which seems odd and a clue to me, but I'm not
sure what kind of clue.

Anyway, open to any thoughts to try.  It's bugging me.

Jon

Hi everybody,

Few days ago I have interfaced cc430f5135 with tmp102. After that I started
interfacing MPL115A2T1 pressure sensor that is also i2c. I managed to
successfully do both but one peculiar thing is bothering me.

In both cases the first byte that is sent and the first byte that is received is
not correct . When sending, first byte after start condition doesn't get send (I
checked on the scope) and when receiving the first byte received is always 0x01
but on the scope I can see the right value.

I solved this by repeating the procedure because after the first read/write
operation on the bus everything acts normal.

Did anyone have similar problems, errata for cc430 doesn't say anything about
this!?

Hello, I need some help.

I'm having some problems regarding access to FLASH control register 3 (FCTL3)..

My intention is to erase INFO segment A.

I managed to clear bits LOCK and LOCKA with tihs set of instructions.

while (FCTL3 & BUSY){}
if(FCTL3 & LOCKA)
{
FCTL3 = FWKEY + LOCKA;
}
else
{
FCTL3 = FWKEY;
}

Both LOCKA and LOCK bits clear perfectly, but LOCK bit goes back to its previos
state (1) immediately after the instruction above is executed. What I mean is
that, the LOCK bit is cleared (along with LOCKA), and in the next instruction,
let's say when I try to set ERASE bit in FCTL1, it goes back to 1. This happens
even when I use a breakpoint (or a the __no_operation() intrinsic) after
clearing LOCK bit, in brief, it happens when beggining to execute ANY
instruction following the clearing of LOCK bit, independently of the instruction
itself.

This is confusing me a bit, cuase I didn't find any documentation describing
this situation.

If anyone can help me, I would be grateful.

Thanks...

Hi Jon

I've got a 64-bit Windows 7 Ultimate machine at home that I've not tried the
launchpad with yet. I'll take a look tonight, or over the weekend, to see if I
can replicate your experience.

Bob

--- In msp430@yahoogroups.com, Jon Kirwan <jonk@...> wrote:
>
> The Launchpad comes with a built-in application that displays
> a single character every second or so once S2 is pressed to
> start it in that direction.  Pressing S2 again should display
> two characters and then one every second.
>
> While it works on EVERY SINGLE OTHER system in my house,
> there is one which does not seem to work over the serial
> port.  It's a brand new machine, an Intel 2600K based system,
> running Windows 7 Ultimate, 64-bit.  I did all the installs
> and hardware build and everything is up to date with
> Microsoft.
>
> I've tried using a variety of COM port monitoring software --
> all of which works with Launchpad just fine on other
> computers.  But on this new system, it displays "X0" every
> time I press S2 (which is wrong, anyway) and displays nothing
> else, otherwise.  I've tried all this on another Windows 7,
> 64-bit machine, which is running the Professional version and
> not Ultimate, and it works just fine.  Same device, same
> cables, nearly the same operating system.  And the serial
> port driver is the same version on these two, 1.3.0.0, from
> TI, dated in 2007.
>
> I've tried even using the CCS compiler IDE's "Terminal" pane
> with the same results.  Works great on every other machine.
> Fails on the new one, in the sense that I see the same "X0"
> when pressing S2 and nothing otherwise.
>
> I've also tried downloading code into devices on this new
> machine.  It works, intermittently.  I'll get an error from
> downloading without debugging, then try downloading WITH
> debugging and it works great.  Then another time it will fail
> that way, and succeed the opposite way.  Sometimes, nothing
> works until I unplug and then replug things in.
>
> I've disabled User Access Controls, and enabled them.  Makes
> no difference at all.  I've tried running as administrator,
> or not.  Same results.  I've tried many different serial port
> monitoring programs.  Except for some oddities in the
> Terminal on CCS which probably have their own explanation,
> everything works/fails in the same ways.
>
> I've no idea what else to check.  The IAR installation I used
> for downloading is 5.20.4 and installed on May 9th, if I
> recall.  But even a variety of independent serial port
> monitoring programs fail out on the new machine.
>
> I've also tried different bit rate settings, though I am
> supposed to use 2400.  Odd thing is, on the failing machine,
> the response is the same regardless of how I set up the bit
> rate: "X0".  Which seems odd and a clue to me, but I'm not
> sure what kind of clue.
>
> Anyway, open to any thoughts to try.  It's bugging me.
>
> Jon
>

I had horrendous problems with Windows 7 Ultimate and Pro. The only
programs I was trying to install were IAR and the jtag and dongle
drivers (CP2102). First try I was asked to reboot. On reboot Win 7 had
lost all its own internal programs, even simple things like the calc
app. Try to run anythinginstalled prior to its brain fart and it would
claim that rundll32.exe had been removed and could not be installed.
same for just about any app. Of course they were all still there. Took
it back to the people I got it from, same issues. Stepped down from
ultimate to pro. same deal. Oddly if I removed IAR after it had stuffed
up IAR would install and run fine. Of course the shop blamed IAR, but
I'd had a similar issue previously when my son had tried installing
windows 7 on his machine. on reboot it basically ate itself. Yet the
only app he had was a win 7 app.

As far as I'm concerned Win 7 is a bug. Its lack of backward
compatibility makes it effectively useless to me. I will stay with XP
until it becomes no longer viable then probably switch to linux. I don't
need most of the memory hogging garbage in XP, and definitely don't want
all the extra baggage that win 7 brings with it.

Al

On 3/06/2011 1:30 PM, Jon Kirwan wrote:
> The Launchpad comes with a built-in application that displays
> a single character every second or so once S2 is pressed to
> start it in that direction.  Pressing S2 again should display
> two characters and then one every second.
>
> While it works on EVERY SINGLE OTHER system in my house,
> there is one which does not seem to work over the serial
> port.  It's a brand new machine, an Intel 2600K based system,
> running Windows 7 Ultimate, 64-bit.  I did all the installs
> and hardware build and everything is up to date with
> Microsoft.
>
> I've tried using a variety of COM port monitoring software --
> all of which works with Launchpad just fine on other
> computers.  But on this new system, it displays "X0" every
> time I press S2 (which is wrong, anyway) and displays nothing
> else, otherwise.  I've tried all this on another Windows 7,
> 64-bit machine, which is running the Professional version and
> not Ultimate, and it works just fine.  Same device, same
> cables, nearly the same operating system.  And the serial
> port driver is the same version on these two, 1.3.0.0, from
> TI, dated in 2007.
>
> I've tried even using the CCS compiler IDE's "Terminal" pane
> with the same results.  Works great on every other machine.
> Fails on the new one, in the sense that I see the same "X0"
> when pressing S2 and nothing otherwise.
>
> I've also tried downloading code into devices on this new
> machine.  It works, intermittently.  I'll get an error from
> downloading without debugging, then try downloading WITH
> debugging and it works great.  Then another time it will fail
> that way, and succeed the opposite way.  Sometimes, nothing
> works until I unplug and then replug things in.
>
> I've disabled User Access Controls, and enabled them.  Makes
> no difference at all.  I've tried running as administrator,
> or not.  Same results.  I've tried many different serial port
> monitoring programs.  Except for some oddities in the
> Terminal on CCS which probably have their own explanation,
> everything works/fails in the same ways.
>
> I've no idea what else to check.  The IAR installation I used
> for downloading is 5.20.4 and installed on May 9th, if I
> recall.  But even a variety of independent serial port
> monitoring programs fail out on the new machine.
>
> I've also tried different bit rate settings, though I am
> supposed to use 2400.  Odd thing is, on the failing machine,
> the response is the same regardless of how I set up the bit
> rate: "X0".  Which seems odd and a clue to me, but I'm not
> sure what kind of clue.
>
> Anyway, open to any thoughts to try.  It's bugging me.
>
> Jon
>
>
> ------------------------------------
>
> To unsubscribe from the msp430 group, send an email to:
> msp430-unsubscribe@egroups.com
>
> Yahoo! Groups Links
>
>
>
>

Hi,

> I had horrendous problems with Windows 7 Ultimate and Pro. The only
> programs I was trying to install were IAR and the jtag and dongle
> drivers (CP2102). First try I was asked to reboot. On reboot Win 7 had
> lost all its own internal programs, even simple things like the calc
> app. Try to run anythinginstalled prior to its brain fart and it would
> claim that rundll32.exe had been removed and could not be installed.
> same for just about any app. Of course they were all still there. Took
> it back to the people I got it from, same issues. Stepped down from
> ultimate to pro. same deal. Oddly if I removed IAR after it had stuffed
> up IAR would install and run fine. Of course the shop blamed IAR, but
> I'd had a similar issue previously when my son had tried installing
> windows 7 on his machine. on reboot it basically ate itself. Yet the
> only app he had was a win 7 app.
>
> As far as I'm concerned Win 7 is a bug. Its lack of backward
> compatibility makes it effectively useless to me. I will stay with XP
> until it becomes no longer viable then probably switch to linux. I
> don't need most of the memory hogging garbage in XP, and definitely
> don't want all the extra baggage that win 7 brings with it.

Win 7 works perfectly for me on all my machines.

Most of the problems Microsoft is blamed for don't happen to be of MS's
making.  Bad 3P drivers and applications are to blame, especially graphics
card drivers.

Move to OS X.  It's great.

--
Paul Curtis, Rowley Associates Ltd�� http://www.rowley.co.uk
SolderCore arriving Summer 2011!   http://www.soldercore.com

Never been a MAC fan, sorry, never been an Apple fan. I tried OS X
several times on my daughter in laws machine. I didn't like it
particularly, more the machine and the 'feel' of things, and, although I
can supposedly run widnows apps on it I find it isn't keen on legacy
stuff. MACs are also considerably more expensive than PC's, and I can't
afford to change my hardware unnecessarily. Unfortunately I need to keep
running legacy stuff to support existing designs. XP is stable enough
for me, and more trusted than Win 7. I made sure to download all of the
latest drivers at the time. The applications might be bad for WIN 7, but
that's part of my point. With WIN 7 MS decided to abandon their backward
compatibility, which, to me at least, was one of the better things
they'd done over the years.

Then I'm a bit of a Luddite, I would still rather be using a modern
version of the Model B than almost anything else.

Al


On 4/06/2011 1:36 AM, Paul Curtis wrote:
> Hi,
>
>> I had horrendous problems with Windows 7 Ultimate and Pro. The only
>> programs I was trying to install were IAR and the jtag and dongle
>> drivers (CP2102). First try I was asked to reboot. On reboot Win 7 had
>> lost all its own internal programs, even simple things like the calc
>> app. Try to run anythinginstalled prior to its brain fart and it would
>> claim that rundll32.exe had been removed and could not be installed.
>> same for just about any app. Of course they were all still there. Took
>> it back to the people I got it from, same issues. Stepped down from
>> ultimate to pro. same deal. Oddly if I removed IAR after it had stuffed
>> up IAR would install and run fine. Of course the shop blamed IAR, but
>> I'd had a similar issue previously when my son had tried installing
>> windows 7 on his machine. on reboot it basically ate itself. Yet the
>> only app he had was a win 7 app.
>>
>> As far as I'm concerned Win 7 is a bug. Its lack of backward
>> compatibility makes it effectively useless to me. I will stay with XP
>> until it becomes no longer viable then probably switch to linux. I
>> don't need most of the memory hogging garbage in XP, and definitely
>> don't want all the extra baggage that win 7 brings with it.
> Win 7 works perfectly for me on all my machines.
>
> Most of the problems Microsoft is blamed for don't happen to be of MS's
> making.  Bad 3P drivers and applications are to blame, especially graphics
> card drivers.
>
> Move to OS X.  It's great.
>
> --
> Paul Curtis, Rowley Associates Ltd   http://www.rowley.co.uk
> SolderCore arriving Summer 2011!   http://www.soldercore.com
>
>
>
>
>
>
> ------------------------------------
>
> To unsubscribe from the msp430 group, send an email to:
> msp430-unsubscribe@egroups.com
>
> Yahoo! Groups Links
>
>
>
>

On Fri, 03 Jun 2011 15:31:54 -0000, Bob wrote:

> I've got a 64-bit Windows 7 Ultimate machine at home that
> I've not tried the launchpad with yet. I'll take a look
> tonight, or over the weekend, to see if I can replicate your
> experience.

Thanks.  The odd thing for me is that I have two Lenovo
laptops, both T510's, running 64-bit Windows 7 Professional
and they work beautifully with IAR download and debug.  Not a
problem, ever.  They also work just fine with the standard
application that comes downloaded into the Lauchpad before it
is sent out, out of the box, with most any serial port
monitoring program.  The virtual COM port just works (if I've
installed IAR on the machine beforehand, of course, to get
the TI driver in there.)  I bought the laptops about 1 year
ago.

I did the fresh installation of the Windows 7 Ultimate
system, myself, using my MSDN subscription (US$1200) disks.
It hasn't had much done to it, as it is new and I'm pretty
cautious about adding software to my development machine.

One thing is that the USB ports on the new machine are USB
3.0 style, I think.  So that makes me wonder if it is related
to this.  I also now think that the serial port monitoring
programs' problems (including the CCS Terminal) is somehow
related to the difficulties I have with IAR's IDE doing the
downloads to the target.  Likely, they both are related to an
underlying problem.  Perhaps due to USB 3.0 hardware and
drivers.

Jon

I just built my new machine a few weeks ago, and if I remember correctly, the
Intel MB I put in has one USB 3.0 port, so I'll take a look at that.  FWIW, the
Intel docs noted that there could be problems if the USB 3.0 port were used for
mouse or keyboard...
I have noticed that with IAR the downloads and debugging have NOT been
completely stable.  At times IAR can't find the device, or locks-up.  However,
I've had a couple of FETs plugged in simultaneously and had assumed (always
dangerous) that the extra VCOM port might be "confusing."  I have found it to be
safer to specify the COM port rather than using automatic mode for connecting
the device...

--- In msp430@yahoogroups.com, Jon Kirwan <jonk@...> wrote:
>
> On Fri, 03 Jun 2011 15:31:54 -0000, Bob wrote:
>
> > I've got a 64-bit Windows 7 Ultimate machine at home that
> > I've not tried the launchpad with yet. I'll take a look
> > tonight, or over the weekend, to see if I can replicate your
> > experience.
>
> Thanks.  The odd thing for me is that I have two Lenovo
> laptops, both T510's, running 64-bit Windows 7 Professional
> and they work beautifully with IAR download and debug.  Not a
> problem, ever.  They also work just fine with the standard
> application that comes downloaded into the Lauchpad before it
> is sent out, out of the box, with most any serial port
> monitoring program.  The virtual COM port just works (if I've
> installed IAR on the machine beforehand, of course, to get
> the TI driver in there.)  I bought the laptops about 1 year
> ago.
>
> I did the fresh installation of the Windows 7 Ultimate
> system, myself, using my MSDN subscription (US$1200) disks.
> It hasn't had much done to it, as it is new and I'm pretty
> cautious about adding software to my development machine.
>
> One thing is that the USB ports on the new machine are USB
> 3.0 style, I think.  So that makes me wonder if it is related
> to this.  I also now think that the serial port monitoring
> programs' problems (including the CCS Terminal) is somehow
> related to the difficulties I have with IAR's IDE doing the
> downloads to the target.  Likely, they both are related to an
> underlying problem.  Perhaps due to USB 3.0 hardware and
> drivers.
>
> Jon
>

On Sat, Jun 4, 2011 at 6:00 AM, Jon Kirwan <jonk@...> wrote:
> One thing is that the USB ports on the new machine are USB
> 3.0 style, I think. �So that makes me wonder if it is related
> to this. �I also now think that the serial port monitoring
> programs' problems (including the CCS Terminal) is somehow
> related to the difficulties I have with IAR's IDE doing the
> downloads to the target. �Likely, they both are related to an
> underlying problem. �Perhaps due to USB 3.0 hardware and
> drivers.

Quite possible. There are lots of problem with the USB 3.0
host controller drivers since Microsoft does not support
USB 3.0 as yet so the drivers come from vendors who
do not really have the capability of writing compatible
drivers as Microsoft.

The situation for USB 3.0 is unfortunate, I do not really
blame Microsoft here but rather Intel. By delaying the
chipset support for USB 3.0 to gain time for ThunderBolt
(Light Peak), Intel has successfully delayed USB 3.0.
However, Thunderbolt will remain niche and USB 3.0
will be for the mass market no matter what Intel does.

--
Xiaofan

On Sat, 4 Jun 2011 10:23:46 +0800, you wrote:

>On Sat, Jun 4, 2011 at 6:00 AM, Jon Kirwan <jonk@...> wrote:
>> One thing is that the USB ports on the new machine are USB
>> 3.0 style, I think. �So that makes me wonder if it is related
>> to this. �I also now think that the serial port monitoring
>> programs' problems (including the CCS Terminal) is somehow
>> related to the difficulties I have with IAR's IDE doing the
>> downloads to the target. �Likely, they both are related to an
>> underlying problem. �Perhaps due to USB 3.0 hardware and
>> drivers.
>
>Quite possible. There are lots of problem with the USB 3.0
>host controller drivers since Microsoft does not support
>USB 3.0 as yet so the drivers come from vendors who
>do not really have the capability of writing compatible
>drivers as Microsoft.
>
>The situation for USB 3.0 is unfortunate, I do not really
>blame Microsoft here but rather Intel. By delaying the
>chipset support for USB 3.0 to gain time for ThunderBolt
>(Light Peak), Intel has successfully delayed USB 3.0.
>However, Thunderbolt will remain niche and USB 3.0
>will be for the mass market no matter what Intel does.

Thanks.  The TI USB driver that got installed by IAR (and, I
suppose, also by CCS) is four years old, now.  It probably
could NOT have been written for USB 3.0, back then.  And so
far as I am now aware, there is no newer version, either.  So
it may be the case that we are not only talking about the
situation you mention above, where Microsoft hasn't yet done
the 3.0 drivers itself, but also about a TI driver that
yields the virtual com port itself also being written when
USB 3.0 wasn't yet something to worry about.

I think TI should probably grapple with this problem and get
their drivers updated so that they work with whatever is out
there right now for USB 3.0, even if the situation is hard
and confusing.  It _has_ been four years, after all.

Jon

On 04/06/11 05:10, Jon Kirwan wrote:
> On Sat, 4 Jun 2011 10:23:46 +0800, you wrote:
>
>  >On Sat, Jun 4, 2011 at 6:00 AM, Jon Kirwan <jonk@...
> <mailto:jonk%40infinitefactors.org>> wrote:
>  >> One thing is that the USB ports on the new machine are USB
>  >> 3.0 style, I think.  So that makes me wonder if it is related
>  >> to this.  I also now think that the serial port monitoring
>  >> programs' problems (including the CCS Terminal) is somehow
>  >> related to the difficulties I have with IAR's IDE doing the
>  >> downloads to the target.  Likely, they both are related to an
>  >> underlying problem.  Perhaps due to USB 3.0 hardware and
>  >> drivers.
>  >
>  >Quite possible. There are lots of problem with the USB 3.0
>  >host controller drivers since Microsoft does not support
>  >USB 3.0 as yet so the drivers come from vendors who
>  >do not really have the capability of writing compatible
>  >drivers as Microsoft.
>  >
>  >The situation for USB 3.0 is unfortunate, I do not really
>  >blame Microsoft here but rather Intel. By delaying the
>  >chipset support for USB 3.0 to gain time for ThunderBolt
>  >(Light Peak), Intel has successfully delayed USB 3.0.
>  >However, Thunderbolt will remain niche and USB 3.0
>  >will be for the mass market no matter what Intel does.
>

I /do/ blame Microsoft - it is Microsoft's responsibility to provide
drivers for basic, standard hardware with their OS.  It is only because
they fail to do a proper job that you get third parties involved in
driver writing.

Intel certainly didn't handle USB 3.0 very well - it kept far too much
secret or "not yet defined" for far too long, and perhaps part of that
was to try to get more support for Thunderbolt.

However, Linux has supported USB 3.0 for two years - MS has no excuse
for not providing good support and drivers for the most common host
chipsets.

> Thanks. The TI USB driver that got installed by IAR (and, I
> suppose, also by CCS) is four years old, now. It probably
> could NOT have been written for USB 3.0, back then. And so
> far as I am now aware, there is no newer version, either. So
> it may be the case that we are not only talking about the
> situation you mention above, where Microsoft hasn't yet done
> the 3.0 drivers itself, but also about a TI driver that
> yields the virtual com port itself also being written when
> USB 3.0 wasn't yet something to worry about.
>

USB support is built in layers - there is the host chipset driver that
is dependent on the host hardware, there is the general USB host
functionality that is entirely software and needs to support the
different USB standards, and there is the device-specific drivers that
are dependent on the particular device.  At the device-specific layer,
you don't see anything that is USB3 dependent unless the hardware is
USB3 and is connected via a USB3 host.  It does not matter that the TI
USB drivers were written before USB3 was developed - they only see USB2
(or, more likely, USB 1.1) packets.

> I think TI should probably grapple with this problem and get
> their drivers updated so that they work with whatever is out
> there right now for USB 3.0, even if the situation is hard
> and confusing. It _has_ been four years, after all.
>

It could be the case that TI's device hardware doesn't work well with
USB3, or that your computer's USB3 host doesn't work well according to
the specifications, or that there are issues with the USB3 drivers on
the machine.  But TI's USB drivers are probably the least problem - you
have to be pretty imaginative to write these in a way that will work
with USB 1.1 and USB 2.0 but not work with USB 3.0.

One way to check the cause of the problems would be to use a USB2.0
socket on the same machine - almost all motherboards only have USB3.0 on
a few sockets.  You could also connect it via a USB2 hub.  And of course
you could try it using an OS that is known to have working USB3 support
- any modern Linux distribution will do.  Use Virtual Box, and install
your IAR or other software inside a Windows virtual machine.  Then if
you still have problems, you can be sure it's a hardware issue with
either TI's hardware, or your USB3 host chipset.

On 03/06/11 18:06, Paul Curtis wrote:
> Hi,
>
>  > I had horrendous problems with Windows 7 Ultimate and Pro. The only
>  > programs I was trying to install were IAR and the jtag and dongle
>  > drivers (CP2102). First try I was asked to reboot. On reboot Win 7 had
>  > lost all its own internal programs, even simple things like the calc
>  > app. Try to run anythinginstalled prior to its brain fart and it would
>  > claim that rundll32.exe had been removed and could not be installed.
>  > same for just about any app. Of course they were all still there. Took
>  > it back to the people I got it from, same issues. Stepped down from
>  > ultimate to pro. same deal. Oddly if I removed IAR after it had stuffed
>  > up IAR would install and run fine. Of course the shop blamed IAR, but
>  > I'd had a similar issue previously when my son had tried installing
>  > windows 7 on his machine. on reboot it basically ate itself. Yet the
>  > only app he had was a win 7 app.
>  >
>  > As far as I'm concerned Win 7 is a bug. Its lack of backward
>  > compatibility makes it effectively useless to me. I will stay with XP
>  > until it becomes no longer viable then probably switch to linux. I
>  > don't need most of the memory hogging garbage in XP, and definitely
>  > don't want all the extra baggage that win 7 brings with it.
>
> Win 7 works perfectly for me on all my machines.
>
> Most of the problems Microsoft is blamed for don't happen to be of MS's
> making. Bad 3P drivers and applications are to blame, especially graphics
> card drivers.
>

That's MS's favourite excuse.  It had some justification 10 years ago -
but not now.  The fact is that both MacOS and Linux come with drivers
for just about all hardware, and it is written by the MacOS and Linux
developers - not third parties - on a fraction of the budget MS has.  MS
charges a lot of money for their OS, and basic hardware drivers for
common hardware is something people can reasonably expect to get for
their money.  And for rarer hardware whose drivers are written by
third-parties, MS charges a great deal of money for their
"certification" programs.  If that certification doesn't come with a
guarantee of a solid level of testing and reliability, then it is worthless.

There will always be third-party software, and there will always be
upper layers of the drivers that are badly written and cause problems.
But if their failure influences anything else - such as crashing the
machine - then the fault lies with the OS, not the drivers.

It is particularly popular to blame graphics drivers for windows
problems - but it is many years since they were a big cause of crashes.
   Even on Linux (where graphics drivers are one of the few third-party
drivers commonly used), they are usually very solid.


My experience with Win 7 is that it mostly works - it's better than XP
for some things (in particular, you can use more memory), and worse for
other things (such as compatibility problems).  But mostly it's just a
waste of time - it doesn't add anything of significant interest to me,
and changes a lot of details for no particular reason.  Networking is
particularly difficult for anything remotely complicated.


All machines I set up these days are Linux (typically either Fedora or
Linux Mint Debian).  When I have to have Windows compatibility, I use a
Virtual Box virtual machine.  This lets me have a fast, efficient and
reliable system that can take advantage of the PC's hardware, and have
an easily controlled and compartmentalised Windows machine inside as needed.


These things are mostly a matter of taste, and individual needs and
experiences.  But don't think "Windows 7 works fine for me" has any
bearing on whether it works or not for other people.  It was a sad day
for professional users the world over when MS decided to cut XP and
force people over to Vista and Win7.  It will be even sadder when they
decide that since no one will buy Windows Phone for phones, we will all
have to buy it for our PC's instead.


> Move to OS X. It's great.
>

Mac's are great if you can live with their simplicity and limitations,
and don't mind bowing to the whims of a control-freak megalomaniac
company whose greed makes MS look like a Good Samaritan.  Sometimes a
pretty toy is all you need, and then Macs make a good choice compared to
the extra complexity, frustrations and unreliabilities of the PC world.

On Sat, 04 Jun 2011 14:56:02 +0200, you wrote:

>On 04/06/11 05:10, Jon Kirwan wrote:

><snip of points about Intel/Microsoft>

>> Thanks. The TI USB driver that got installed by IAR (and, I
>> suppose, also by CCS) is four years old, now. It probably
>> could NOT have been written for USB 3.0, back then. And so
>> far as I am now aware, there is no newer version, either. So
>> it may be the case that we are not only talking about the
>> situation you mention above, where Microsoft hasn't yet done
>> the 3.0 drivers itself, but also about a TI driver that
>> yields the virtual com port itself also being written when
>> USB 3.0 wasn't yet something to worry about.
>
>USB support is built in layers - there is the host chipset driver that
>is dependent on the host hardware, there is the general USB host
>functionality that is entirely software and needs to support the
>different USB standards, and there is the device-specific drivers that
>are dependent on the particular device.  At the device-specific layer,
>you don't see anything that is USB3 dependent unless the hardware is
>USB3 and is connected via a USB3 host.  It does not matter that the TI
>USB drivers were written before USB3 was developed - they only see USB2
>(or, more likely, USB 1.1) packets.

I am still completely in the dark about the problem.  So I
won't argue about any of this.  I'm looking for differences
and similarities right now that may suggest or hint at a
direction, is all.

>> I think TI should probably grapple with this problem and get
>> their drivers updated so that they work with whatever is out
>> there right now for USB 3.0, even if the situation is hard
>> and confusing. It _has_ been four years, after all.
>
>It could be the case that TI's device hardware doesn't work well with
>USB3, or that your computer's USB3 host doesn't work well according to
>the specifications, or that there are issues with the USB3 drivers on
>the machine.  But TI's USB drivers are probably the least problem - you
>have to be pretty imaginative to write these in a way that will work
>with USB 1.1 and USB 2.0 but not work with USB 3.0.

So many possibilities to consider.  I have used other USB
devices that work fine in this new machine.  But of course,
it's a limited set.  The wifi is USB, as is the mouse.  (I've
used three different mice on this machine.. all worked fine.)

I did recently try the eZ430-F2013 Development Tool, the
newer one that includes a virtual COM port, as well.  I did
NOT write and install serial port software on it, but I do
have some other workspaces that I've developed for that tool
and I loaded them up and compiled and downloaded code both
with 'download and debug' and without the debug part.
Multiple times.  Over and over again.  Not a single failure
and, as well, all of them without any delays.  Very fast.

Interesting thing....  When I decided to first insert it into
a USB port, I got a message about installing new drivers. But
never a failure of any kind.  I did both types of downloading
into the tool perhaps 30 times to be sure. Worked as smoothly
as I could have ever hoped.

(It took a LONG TIME on this machine between the first
'installing' message and the eventual 'successful' one. Maybe
2 minutes?  Before it said they were successfully installed.
This is a very fast machine and on my slower machines such
messages never took that long.  No idea why. But there it
is.)

So I then closed that workspace, loaded up the one using the
Launchpad again, took out the eZ430 tool and stuck back in
the Launchpad one, then attempted the downloads. Similar
errors per earlier -- sometimes it works, sometimes not. Same
stuff.

Suggests to me something about the target system's installed
USB software.  Now I wonder if I need to find a new driver to
load into it.  However, there is a tool on the web that
checks it and I'd run that tool: LaunchpadFirmwareUpdater2.0.
It says it is up to date.  But that program is dated in
March, I think.  So....

>One way to check the cause of the problems would be to use a USB2.0
>socket on the same machine - almost all motherboards only have USB3.0 on
>a few sockets.  You could also connect it via a USB2 hub.  And of course
>you could try it using an OS that is known to have working USB3 support
>- any modern Linux distribution will do.  Use Virtual Box, and install
>your IAR or other software inside a Windows virtual machine.  Then if
>you still have problems, you can be sure it's a hardware issue with
>either TI's hardware, or your USB3 host chipset.

It's a Gigabyte P67-UD7.  It has some USB 2.0 and USB 3.0.  I
just tried quite a few ports, front and back.  So probably
caught some of both types.  But it's worth being absolutely
sure, so I will check that out a little later on and see
after I dig out the manual, a flashlight, and open up the
unit to be dead certain.  Worth being sure about it.

Jon

There are lots of previous discussions on this board about how to do it. I did a
field upgrade bootloader for our msp430 product, took about 3 weeks to get the
whole process well automated.

Lloyd

--- In msp430@yahoogroups.com, Howard Hansen <hrhan@...> wrote:
>
> How do you design a MSP430 product so it is field upgradeable?
>
> Howard
>
>

On Sat, Jun 4, 2011 at 9:34 AM, David Brown <david@...> wrote:

> Mac's are great if you can live with their simplicity and limitations,
> and don't mind bowing to the whims of a control-freak megalomaniac
> company whose greed makes MS look like a Good Samaritan. �Sometimes a
> pretty toy is all you need, and then Macs make a good choice compared to
> the extra complexity, frustrations and unreliabilities of the PC world.

You should really lose the dogma here.   Seriously.  I have been
primarily a Unix user since the late 1980s, but I have used both Macs
and PCs for specific applications without Unix solutions.  I've been
using Unix as my primary home desktop since the early 1990s, starting
with an AT&T 3b1 and going through several SGIs before adopting Linux
in 1997.  I switched from Linux to OS/X about five years ago.  I still
use Linux (and mostly Solaris) on servers, and I'm up to date with
Linux on the desktop, but I *still* prefer the OS/X desktop to the
Linux desktop *despite* 20+ years of experience with X-Windows.

Unfortunately, OS/X is not a terribly good solution when working with
the MSP430.  The two most popular compilers (CCS and IAR) do not
support OS/X, and TI has stated explicitly that it will not support
OS/X with CCS 5.  I have heard good things about Crossworks, but lack
of a free code/optimization limited version is a bit of a show-stopper
for me personally.  mspgcc development has been somewhat lacking in
recent years, and support for the chip I want to work with (the G2553)
is only available in the development (pre-alpha) branch.  Finally, the
wonderfully convenient serial tunnel in the Launchpad is not supported
under OS/X.  I am not sure if this is a failure on Apple's part or as
a result of an ambiguity in the USB specification.

Despite these drawbacks, they are still not substantial enough to
force me to run inside of a VM or switch to Linux.  I have not yet had
any problems with the "Uniarch" version of mspgcc and I can cope with
an extra serial dongle when I need it.

I do have an old laptop which did get a fresh install of Linux which I
probably will start using once I start adding external voltage sources
to my circuits, but I'll still likely access it remotely from my OS/X
box.

-p.

Hi Jon,

I've been unable to replicate your problem.  I tried with both my Win7 machine
and one of my XP machines.  If I set the baud rate incorrectly, I saw the
results you described. However, I recall from your original post you'd tried
various baud rate settings to no avail.

My one interesting discovery:  While the launchpad worked fine when plugged into
the ports on the PC, if I plugged it into a USB 2.0 Hub (D-Link), it would cause
a reset of the hub, and trying to run the app from the hub usually had the same
result...

Just for the record -
Win7 desktop configuration:
     Motherboard:          Intel DX58S02 with 24GB ram
                           2- USB 3.0 ports
                           8- USB 2.0 ports
     CPU:                  Intel i7 X980 running at default 3.33GHz
     OS:                   Windows 7 Ultimate 64-bit, service pack 1
     Apps:                 current releases of IAR & CCS
                           Terminal Emulation - Hercules 3.2.4
                           TI Driver 1.3.0.0 12/2007
Typically, the launchpad showed-up as Comm6, although it would ocassionally
cycle through comms 7 and 8 when it was causing resets on the USB hub...


XP Laptop:
                           1GB ram, 4 - USB 2.0 ports
    CPU:                   Intel T5500 running at 1.66GHz
    OS:                    Windows XP professional, servcice pack 3
     Apps:                 Earlier release of IAR, current CCS
                           and Crossworks releases
                           Terminal Emulation - Hercules 3.2.4
                           TI Driver 1.3.0.0 12/2007
On the laptop the launchpad would show up as comm12.

Sorry I couldn't identify something more useful.

Bob


--- In msp430@yahoogroups.com, Jon Kirwan <jonk@...> wrote:
>
> On Sat, 04 Jun 2011 14:56:02 +0200, you wrote:
>
> >On 04/06/11 05:10, Jon Kirwan wrote:
>
> ><snip of points about Intel/Microsoft>
>
> >> Thanks. The TI USB driver that got installed by IAR (and, I
> >> suppose, also by CCS) is four years old, now. It probably
> >> could NOT have been written for USB 3.0, back then. And so
> >> far as I am now aware, there is no newer version, either. So
> >> it may be the case that we are not only talking about the
> >> situation you mention above, where Microsoft hasn't yet done
> >> the 3.0 drivers itself, but also about a TI driver that
> >> yields the virtual com port itself also being written when
> >> USB 3.0 wasn't yet something to worry about.
> >
> >USB support is built in layers - there is the host chipset driver that
> >is dependent on the host hardware, there is the general USB host
> >functionality that is entirely software and needs to support the
> >different USB standards, and there is the device-specific drivers that
> >are dependent on the particular device.  At the device-specific layer,
> >you don't see anything that is USB3 dependent unless the hardware is
> >USB3 and is connected via a USB3 host.  It does not matter that the TI
> >USB drivers were written before USB3 was developed - they only see USB2
> >(or, more likely, USB 1.1) packets.
>
> I am still completely in the dark about the problem.  So I
> won't argue about any of this.  I'm looking for differences
> and similarities right now that may suggest or hint at a
> direction, is all.
>
> >> I think TI should probably grapple with this problem and get
> >> their drivers updated so that they work with whatever is out
> >> there right now for USB 3.0, even if the situation is hard
> >> and confusing. It _has_ been four years, after all.
> >
> >It could be the case that TI's device hardware doesn't work well with
> >USB3, or that your computer's USB3 host doesn't work well according to
> >the specifications, or that there are issues with the USB3 drivers on
> >the machine.  But TI's USB drivers are probably the least problem - you
> >have to be pretty imaginative to write these in a way that will work
> >with USB 1.1 and USB 2.0 but not work with USB 3.0.
>
> So many possibilities to consider.  I have used other USB
> devices that work fine in this new machine.  But of course,
> it's a limited set.  The wifi is USB, as is the mouse.  (I've
> used three different mice on this machine.. all worked fine.)
>
> I did recently try the eZ430-F2013 Development Tool, the
> newer one that includes a virtual COM port, as well.  I did
> NOT write and install serial port software on it, but I do
> have some other workspaces that I've developed for that tool
> and I loaded them up and compiled and downloaded code both
> with 'download and debug' and without the debug part.
> Multiple times.  Over and over again.  Not a single failure
> and, as well, all of them without any delays.  Very fast.
>
> Interesting thing....  When I decided to first insert it into
> a USB port, I got a message about installing new drivers. But
> never a failure of any kind.  I did both types of downloading
> into the tool perhaps 30 times to be sure. Worked as smoothly
> as I could have ever hoped.
>
> (It took a LONG TIME on this machine between the first
> 'installing' message and the eventual 'successful' one. Maybe
> 2 minutes?  Before it said they were successfully installed.
> This is a very fast machine and on my slower machines such
> messages never took that long.  No idea why. But there it
> is.)
>
> So I then closed that workspace, loaded up the one using the
> Launchpad again, took out the eZ430 tool and stuck back in
> the Launchpad one, then attempted the downloads. Similar
> errors per earlier -- sometimes it works, sometimes not. Same
> stuff.
>
> Suggests to me something about the target system's installed
> USB software.  Now I wonder if I need to find a new driver to
> load into it.  However, there is a tool on the web that
> checks it and I'd run that tool: LaunchpadFirmwareUpdater2.0.
> It says it is up to date.  But that program is dated in
> March, I think.  So....
>
> >One way to check the cause of the problems would be to use a USB2.0
> >socket on the same machine - almost all motherboards only have USB3.0 on
> >a few sockets.  You could also connect it via a USB2 hub.  And of course
> >you could try it using an OS that is known to have working USB3 support
> >- any modern Linux distribution will do.  Use Virtual Box, and install
> >your IAR or other software inside a Windows virtual machine.  Then if
> >you still have problems, you can be sure it's a hardware issue with
> >either TI's hardware, or your USB3 host chipset.
>
> It's a Gigabyte P67-UD7.  It has some USB 2.0 and USB 3.0.  I
> just tried quite a few ports, front and back.  So probably
> caught some of both types.  But it's worth being absolutely
> sure, so I will check that out a little later on and see
> after I dig out the manual, a flashlight, and open up the
> unit to be dead certain.  Worth being sure about it.
>
> Jon
>

On 04/06/11 20:23, Peter Johansson wrote:
> On Sat, Jun 4, 2011 at 9:34 AM, David Brown <david@...
> <mailto:david%40westcontrol.com>> wrote:
>
>  > Mac's are great if you can live with their simplicity and limitations,
>  > and don't mind bowing to the whims of a control-freak megalomaniac
>  > company whose greed makes MS look like a Good Samaritan.  Sometimes a
>  > pretty toy is all you need, and then Macs make a good choice compared to
>  > the extra complexity, frustrations and unreliabilities of the PC world.
>
> You should really lose the dogma here. Seriously. I have been

Well, I'll go as far as admitting to dogma - but I'm not ready to lose
it entirely yet.

The Mac is fine if you can fit entirely within the Mac way of thinking -
everything should come from Apple, it should look stylish, but it should
not do anything that is not blessed by Apple.  It is rare that something
goes wrong - the closed hardware and software (and reseller and service)
environment has a big advantage there.  But if something /does/ go
wrong, you have a much bigger problem and self-help is far harder.

For some uses, the Mac is a great solution.  My mother has one, and I
think it was the right choice for her.  She can do her email, use the
web, and do the word-processing and spreadsheet work that she needs.  It
looks good, and apart from a couple of occasions, it has had no hardware
or software problems.  But it definitely has its limitations - she has
the Apple word processor that came with it, and can't swap documents
with anyone else.  When something did go wrong with the hardware, there
was no choice but to take it to the Apple shop and send it off for
service.  And she is paranoid about avoiding anything non-Apple blessed
on the machine - that may be just my mother, but I understand it is not
uncommon for Mac users.

Basically, Macs are a poor choice if you need to have something
(hardware or software) a little unusual, or not commonly used by Mac
users.  It is certainly /possible/ to dig into a Mac system and install
anything you like - it's FreeBSD at heart, and you can install an X
server and pretty much everything else you want.  But then you have gone
very far from the Mac philosophy, and you lose most of the benefits of
having a Mac - you can no longer be confident that everything will work
together and keep working together.  In that case, why bother with the
Mac?  Because you like paying extra for having a mouse that looks like
it has only one button, and a cool fruity logo on the screen?

> primarily a Unix user since the late 1980s, but I have used both Macs
> and PCs for specific applications without Unix solutions. I've been
> using Unix as my primary home desktop since the early 1990s, starting
> with an AT&T 3b1 and going through several SGIs before adopting Linux
> in 1997. I switched from Linux to OS/X about five years ago. I still
> use Linux (and mostly Solaris) on servers, and I'm up to date with
> Linux on the desktop, but I *still* prefer the OS/X desktop to the
> Linux desktop *despite* 20+ years of experience with X-Windows.
>
> Unfortunately, OS/X is not a terribly good solution when working with
> the MSP430. The two most popular compilers (CCS and IAR) do not
> support OS/X, and TI has stated explicitly that it will not support
> OS/X with CCS 5. I have heard good things about Crossworks, but lack
> of a free code/optimization limited version is a bit of a show-stopper
> for me personally. mspgcc development has been somewhat lacking in
> recent years, and support for the chip I want to work with (the G2553)
> is only available in the development (pre-alpha) branch. Finally, the
> wonderfully convenient serial tunnel in the Launchpad is not supported
> under OS/X. I am not sure if this is a failure on Apple's part or as
> a result of an ambiguity in the USB specification.
>
> Despite these drawbacks, they are still not substantial enough to
> force me to run inside of a VM or switch to Linux. I have not yet had
> any problems with the "Uniarch" version of mspgcc and I can cope with
> an extra serial dongle when I need it.
>
> I do have an old laptop which did get a fresh install of Linux which I
> probably will start using once I start adding external voltage sources
> to my circuits, but I'll still likely access it remotely from my OS/X
> box.
>

If you like it, and it works for you, then that's great - I'm all for
freedom of choice.  I think software should often be written
cross-platform, because it should be the user's choice which OS they
want (and also because it's a natural consequence of good software
development, and good forward planning on behalf of the developer).

But different OS's have different strengths, and there can be little
doubt that the Mac is /very/ weak for technical and engineering work.
You might prefer the Mac's desktop to other systems (I don't, but I
haven't used it much, and it's a matter of taste anyway) - but that
won't change reality.  Personally, I prefer to use Linux - but I know I
can't avoid Windows for a number of programs.  Using a Mac would make
life even harder - and without the advantages of Linux for technical
software.


And if it sounded earlier like I have a bee in my bonnet about Apple,
then it's because I do - but that's getting /really/ off-topic for this
group!

On Sat, Jun 4, 2011 at 6:43 PM, David Brown <david@...> wrote:

> The Mac is fine if you can fit entirely within the Mac way of thinking -
> everything should come from Apple, it should look stylish, but it should
> not do anything that is not blessed by Apple.

I'm not quite sure where you get that from.  Apple has provided a
first-class development environment and complete documentation for
quite some time now.  And until recently, it was completely free.

It also ships with gcc, libraries, and headers, allowing you to
compile, install, and use just about anything available under Linux.

> �It is rare that something
> goes wrong - the closed hardware and software (and reseller and service)
> environment has a big advantage there. �But if something /does/ go
> wrong, you have a much bigger problem and self-help is far harder.

Apple is remarkably good about simply replacing things when they fail,
even if due to the user's fault, and even when out of warranty.

Granted, a large portion of Apple's sales are unified devices
(laptops, iPads, and iMacs) using almost entirely proprietary
components, obviously making end-user repair difficult if not
impossible.  But this should be obvious to anyone with even minimal
technical knowledge.

Yes, I have built plenty of desktop PCs over the years.  I have run
into a ridiculous number of problems just getting things to work from
the start.

Case in point:  My most recent build, about three years ago, was an
OpenSolaris/ZFS file server.  I chose a mobo with 6 SATA ports as well
as IDE.  Because OpenSolaris cannot boot from ZFS, the plan was to
boot from a single IDE drive and use the six SATA ports in RAIDZ2
configuration.  Turns out that when you have drives on all six SATA
ports, the BIOS won't let you boot from IDE.  So I have to boot from a
USB attached drive.  I spent an entire day tracking that down before
just giving up and waiting for a BIOS update -- that never came.  I
have run into similar problems with just about every single PC I've
ever built.

I am one of the people who does appreciate the styling of Apple
products, and when I purchased my first MB Pro, I was actually
planning to wipe OS/X and dual-boot Linux and Windows.  Turns out that
not only did I not do that, but I never felt the need to re-install
the factory O/S myself -- something I always did to new machines
previously.

> For some uses, the Mac is a great solution. �My mother has one, and I
> think it was the right choice for her. �She can do her email, use the
> web, and do the word-processing and spreadsheet work that she needs. �It
> looks good, and apart from a couple of occasions, it has had no hardware
> or software problems. �But it definitely has its limitations - she has
> the Apple word processor that came with it, and can't swap documents
> with anyone else. �When something did go wrong with the hardware, there
> was no choice but to take it to the Apple shop and send it off for
> service. �And she is paranoid about avoiding anything non-Apple blessed
> on the machine - that may be just my mother, but I understand it is not
> uncommon for Mac users.
>
> Basically, Macs are a poor choice if you need to have something
> (hardware or software) a little unusual, or not commonly used by Mac
> users.

Yes, a lot more people know Windows.  It should be statistically
obvious that since the ratio of PC to Mac users is 10:1, you will know
to PC fix-it people for every one Mac person.  The reality is that
Macs "break" far less often, so far fewer Mac people ever need to
develop the skills required to "fix" them.

For a long time I was the go-to guy for PC repair (which is to say,
mostly Windows repair) among my friends.  While I wanted them to
switch to Linux, I knew that was not a viable option.  However,
shortly after "going mac" I developed a new policy.  "If you buy a
Mac, I'll help you with you Mac problems.  If you want PC help, I'll
do that at market rates."  Those who listened and bought Macs have
required almost zero assistance.

>�It is certainly /possible/ to dig into a Mac system and install
> anything you like - it's FreeBSD at heart, and you can install an X
> server and pretty much everything else you want. �But then you have gone
> very far from the Mac philosophy, and you lose most of the benefits of
> having a Mac - you can no longer be confident that everything will work
> together and keep working together. �In that case, why bother with the
> Mac? �Because you like paying extra for having a mouse that looks like
> it has only one button, and a cool fruity logo on the screen?

That's not how I look at it at all.  I look at the Mac as a very
capable Unix machine with a *commercial* operating system and
*professional* support that just happens to have a very nice GUI and
just happens to be supplied on aesthetically pleasing hardware.

One of my biggest complaints about Linux is lack of standardization on
the desktop, standardization among distributions, and standardization
between GUI and command-line configuration.  It just drives me
friggin' nuts.

And don't even get me started with the people who like to "theme"
their systems.  I'm glad there is an option for people who like that
stuff, but I don't want any part of it!

> If you like it, and it works for you, then that's great - I'm all for
> freedom of choice. �I think software should often be written
> cross-platform, because it should be the user's choice which OS they
> want (and also because it's a natural consequence of good software
> development, and good forward planning on behalf of the developer).

Me too.  My initial response was merely to point out the factual
inaccuracies in your initial statements.

> But different OS's have different strengths, and there can be little
> doubt that the Mac is /very/ weak for technical and engineering work.

It is only weak because corporations do not support it, and not for
any technical reason.  It is the classic chicken-and-egg problem.
Fortunately, that is beginning to change.  MicroChip, for example, has
made the commitment to go fully cross-platform with MPLAB-X.

> You might prefer the Mac's desktop to other systems (I don't, but I
> haven't used it much, and it's a matter of taste anyway) - but that
> won't change reality.

Um, are you even listening to yourself here?  How can you possibly
state that you do not like something that you have not even taken the
time learn?

Sure, there is a bit of a learning curve with OS/X as there is with
any product.  The big difference for *me* is that the more time I
spend working with OS/X, the more I like it.  This is in stark
contrast to many other systems, where the more I use them the more I
hate them.

>  �Personally, I prefer to use Linux - but I know I
> can't avoid Windows for a number of programs. �Using a Mac would make
> life even harder - and without the advantages of Linux for technical
> software.

I agree, it really sucks when vendors do not provide OS/X support, or
provide only half-hearted or token support.  But that's not a fault of
OS/X.  The reality is that I probably know *more* about the internals
of Linux than I do about OS/X, but only because there are so many
things I have had to go in and fix over the years.  The reality is
that I have learned so little about OS/X because I have not needed to.

Personally, I am a believer in choosing the best tool for the job.
Unfortunately, the development options for the MSP430 *do* rather suck
under OS/X at the moment.  But they don't suck appreciably more than
they do under Linux.

> And if it sounded earlier like I have a bee in my bonnet about Apple,
> then it's because I do - but that's getting /really/ off-topic for this
> group!

The problem is that as long as people continue to propagate complete
un-truths, companies like TI are going to shy away from supporting
OS/X.  This is what I have a problem with and this is why the
discussion *is* appropriate for the list.

As a professional software developer of some 20+ years now, and as
someone who has both used and developed for most of the major
platforms, I can state unequivocally that OS/X is a first-class
development environment.  I am not going to state that it is better
than "x" or "y", but I can say that it is *my* first choice -- and by
a wide margin.

So, David, please don't take any of my comments personally, they are
really directed more at TI than anyone else.

-p.

On 05/06/2011 02:09, Peter Johansson wrote:
> On Sat, Jun 4, 2011 at 6:43 PM, David Brown <david@...
> <mailto:david%40westcontrol.com>> wrote:
>
>  > The Mac is fine if you can fit entirely within the Mac way of thinking -
>  > everything should come from Apple, it should look stylish, but it should
>  > not do anything that is not blessed by Apple.
>
> I'm not quite sure where you get that from. Apple has provided a
> first-class development environment and complete documentation for
> quite some time now. And until recently, it was completely free.
>
> It also ships with gcc, libraries, and headers, allowing you to
> compile, install, and use just about anything available under Linux.
>

You've convinced me that I'll need to read/think/learn a bit more here.

Perhaps I just don't know enough Mac users, or that those that I do know
are very specific and limited in their use.


>  >  It is rare that something
>  > goes wrong - the closed hardware and software (and reseller and service)
>  > environment has a big advantage there.  But if something /does/ go
>  > wrong, you have a much bigger problem and self-help is far harder.
>
> Apple is remarkably good about simply replacing things when they fail,
> even if due to the user's fault, and even when out of warranty.
>

I've no doubt of that - I'm sure Apple's service is good.  My point was
just that it is only Apple, or their very close partners, that can do
any service - you don't have the flexibility that you find in the PC
world.  And I also appreciate that this has its benefits too - you are
guaranteed a certain level of service.

> Granted, a large portion of Apple's sales are unified devices
> (laptops, iPads, and iMacs) using almost entirely proprietary
> components, obviously making end-user repair difficult if not
> impossible. But this should be obvious to anyone with even minimal
> technical knowledge.
>
> Yes, I have built plenty of desktop PCs over the years. I have run
> into a ridiculous number of problems just getting things to work from
> the start.
>
> Case in point: My most recent build, about three years ago, was an
> OpenSolaris/ZFS file server. I chose a mobo with 6 SATA ports as well
> as IDE. Because OpenSolaris cannot boot from ZFS, the plan was to
> boot from a single IDE drive and use the six SATA ports in RAIDZ2
> configuration. Turns out that when you have drives on all six SATA
> ports, the BIOS won't let you boot from IDE. So I have to boot from a
> USB attached drive. I spent an entire day tracking that down before
> just giving up and waiting for a BIOS update -- that never came. I
> have run into similar problems with just about every single PC I've
> ever built.
>

That's both the advantage and the disadvantage of an "open" platform -
you have the choices, you can mix and match and do advanced stuff if you
want.  But you also take chances there, and have to be more
self-reliant.  You have to know more about what you are doing yourself,
or be willing to find out, or to know where to find people that can help
you.  That's the price you pay.  But I'm sure that when you built this
server (perhaps booting from a USB stick), it's cost price would be a
fraction of a Mac server with equivalent space and speed.  Whether the
whole thing is cost-effective or not depends on the results, and the
time spent on it.

I've seen the opposite problems - I recently bought an IBM server at
work, and wanted to use the drives individually instead of using the
built-in raid.  It was basically impossible - it could not be done with
out all sorts of re-building that would invalidate all the guarantees.
So I am stuck with the limited IBM-official raid setup, instead of using
a more flexible, faster and safer Linux mdraid setup.


> I am one of the people who does appreciate the styling of Apple
> products, and when I purchased my first MB Pro, I was actually
> planning to wipe OS/X and dual-boot Linux and Windows. Turns out that
> not only did I not do that, but I never felt the need to re-install
> the factory O/S myself -- something I always did to new machines
> previously.
>

With Windows, you always have to re-install the "pre-installed" OS if
you want to get the best out of the machine.  The "pre-installed" images
usually take many hours to install themselves, and include endless
realms of junk, trial-ware, ad-ware, etc.  It's normally much faster to
install from scratch than to use the "pre-installed" system.


>  > For some uses, the Mac is a great solution.  My mother has one, and I
>  > think it was the right choice for her.  She can do her email, use the
>  > web, and do the word-processing and spreadsheet work that she needs.  It
>  > looks good, and apart from a couple of occasions, it has had no hardware
>  > or software problems.  But it definitely has its limitations - she has
>  > the Apple word processor that came with it, and can't swap documents
>  > with anyone else.  When something did go wrong with the hardware, there
>  > was no choice but to take it to the Apple shop and send it off for
>  > service.  And she is paranoid about avoiding anything non-Apple blessed
>  > on the machine - that may be just my mother, but I understand it is not
>  > uncommon for Mac users.
>  >
>  > Basically, Macs are a poor choice if you need to have something
>  > (hardware or software) a little unusual, or not commonly used by Mac
>  > users.
>
> Yes, a lot more people know Windows. It should be statistically
> obvious that since the ratio of PC to Mac users is 10:1, you will know
> to PC fix-it people for every one Mac person. The reality is that
> Macs "break" far less often, so far fewer Mac people ever need to
> develop the skills required to "fix" them.
>

These too facts conspire to give a very small number of people who are
able to help you out with Mac problems (if they do occur).

There is a similar situation with Linux.

> For a long time I was the go-to guy for PC repair (which is to say,
> mostly Windows repair) among my friends. While I wanted them to
> switch to Linux, I knew that was not a viable option. However,
> shortly after "going mac" I developed a new policy. "If you buy a
> Mac, I'll help you with you Mac problems. If you want PC help, I'll
> do that at market rates." Those who listened and bought Macs have
> required almost zero assistance.
>

I can appreciate that!  I have converted my nearby family to Linux to
avoid problems, after battling to remove viruses from my mother-in-law's
windows machine.  I suppose a Mac would have been an alternative for
her, but I'm not familiar enough with it to give her much support, and
the nearest Apple shop is a long way from here.

All I need to do now is convert people at the office to Linux...

>  > It is certainly /possible/ to dig into a Mac system and install
>  > anything you like - it's FreeBSD at heart, and you can install an X
>  > server and pretty much everything else you want.  But then you have gone
>  > very far from the Mac philosophy, and you lose most of the benefits of
>  > having a Mac - you can no longer be confident that everything will work
>  > together and keep working together.  In that case, why bother with the
>  > Mac?  Because you like paying extra for having a mouse that looks like
>  > it has only one button, and a cool fruity logo on the screen?
>
> That's not how I look at it at all. I look at the Mac as a very
> capable Unix machine with a *commercial* operating system and
> *professional* support that just happens to have a very nice GUI and
> just happens to be supplied on aesthetically pleasing hardware.
>

You are showing me a viewpoint that I thought was very unusual in the
Mac world.  I understand what you mean here - it's just that you are far
from what I understood to be the "average" Mac user.  (I know that the
"average" Windows user hasn't a clue how their system works, but there
is, I think, a higher proportion of "power users" amongst Windows users.)

> One of my biggest complaints about Linux is lack of standardization on
> the desktop, standardization among distributions, and standardization
> between GUI and command-line configuration. It just drives me
> friggin' nuts.
>

Yes, sometimes there is such a thing as too much choice, and sometimes
it is best with a single system that you are familiar with and does the
job (that's partly why I prefer XP to Win7).

> And don't even get me started with the people who like to "theme"
> their systems. I'm glad there is an option for people who like that
> stuff, but I don't want any part of it!
>

I agree here.  I can't understand why people would want to change their
"wallpaper" - you can't see the wallpaper if you are actually /using/
the machine!

>  > If you like it, and it works for you, then that's great - I'm all for
>  > freedom of choice.  I think software should often be written
>  > cross-platform, because it should be the user's choice which OS they
>  > want (and also because it's a natural consequence of good software
>  > development, and good forward planning on behalf of the developer).
>
> Me too. My initial response was merely to point out the factual
> inaccuracies in your initial statements.
>

And I thank you for correcting and educating me - I've learned a few
things here in this thread from you.

>  > But different OS's have different strengths, and there can be little
>  > doubt that the Mac is /very/ weak for technical and engineering work.
>
> It is only weak because corporations do not support it, and not for
> any technical reason. It is the classic chicken-and-egg problem.

I never doubted the technical capabilities of the Mac (though I might
have been unclear there) - just the technical abilities and interests of
most Mac users (and I don't mean that as a slight against Mac users - if
you are not a "computer person" then Macs make a good choice).

It is, as you say, a chicken-and-egg problem - there is the same issue
in the Linux world.  There is a noticeable difference, however.  Some
commercial software developers /do/ provide software for non-Windows
platforms, but there is a clear distinction on their choice of a second
platform.  If they are making graphics or multimedia programs, or games,
they will pick the Mac as number two (or as platform number one).  If
they are making technical software (development tools, engineering or
scientific software, server or network software, etc.), then Linux is
the number two (or number one) platform.

Again, there are no /technical/ reasons for that distinction.  The days
when Macs were naturally better at graphics and multimedia than Windows
or Linux are at least ten years behind us, and Macs are as capable for
development software as Linux machines.

> Fortunately, that is beginning to change. MicroChip, for example, has
> made the commitment to go fully cross-platform with MPLAB-X.
>
>  > You might prefer the Mac's desktop to other systems (I don't, but I
>  > haven't used it much, and it's a matter of taste anyway) - but that
>  > won't change reality.
>
> Um, are you even listening to yourself here? How can you possibly
> state that you do not like something that you have not even taken the
> time learn?
>

It doesn't take long to find that you don't much like a system.  I'll
accept that the Mac desktop might grow on me, and that perhaps I'd
prefer it if I worked with it for long enough.  And there is no doubt
that my preferences are heavily based on what I'm used to, rather than
some intrinsic or objective measure.  That's really what "a matter of
taste" means.

If you want specifics, I don't like the unified menu on the top of the
screen, I don't like the huge icons on bottom of the screen, I don't
like the mixup between application-start icons and running application
icons, and I don't like the window decorations.  I like a desktop
interface that keeps out of the way and makes it as easy as possible for
me to work with the applications I am using.

Maybe one day I'll get a bit more familiar with Macs.  There are a few
programs and pieces of hardware made by my company that would be of
interest to Mac users, and I'm sure at some point we'll get a Mac to
test with.  And I have no doubts as to who will get the job of learning
the system and getting everything to work with it.

But I can't see the point of spending time learning about a new desktop
and a new OS just on the off-chance that I will prefer it after using it
for a while - especially when I know it will make it much harder for me
to use the programs I need.  If I want to try different desktop
arrangements, I can install a dozen different ones on my Linux machines
and still use the same application software.

> Sure, there is a bit of a learning curve with OS/X as there is with
> any product. The big difference for *me* is that the more time I
> spend working with OS/X, the more I like it. This is in stark
> contrast to many other systems, where the more I use them the more I
> hate them.
>

Again, this is personal preference.  The more I use Linux (with a Gnome
2 desktop, though I spend a lot of time at the command line), the more I
hate Windows...

>  >  Personally, I prefer to use Linux - but I know I
>  > can't avoid Windows for a number of programs.  Using a Mac would make
>  > life even harder - and without the advantages of Linux for technical
>  > software.
>
> I agree, it really sucks when vendors do not provide OS/X support, or
> provide only half-hearted or token support. But that's not a fault of
> OS/X. The reality is that I probably know *more* about the internals
> of Linux than I do about OS/X, but only because there are so many
> things I have had to go in and fix over the years. The reality is
> that I have learned so little about OS/X because I have not needed to.
>
> Personally, I am a believer in choosing the best tool for the job.
> Unfortunately, the development options for the MSP430 *do* rather suck
> under OS/X at the moment. But they don't suck appreciably more than
> they do under Linux.
>

I must admit I haven't tried msp430 development with Linux - in fact, I
haven't done that much msp430 work in the past few years.

>  > And if it sounded earlier like I have a bee in my bonnet about Apple,
>  > then it's because I do - but that's getting /really/ off-topic for this
>  > group!
>
> The problem is that as long as people continue to propagate complete
> un-truths, companies like TI are going to shy away from supporting
> OS/X. This is what I have a problem with and this is why the
> discussion *is* appropriate for the list.
>

Fair enough - and I hope TI (and others) listen more to you than me!

> As a professional software developer of some 20+ years now, and as
> someone who has both used and developed for most of the major
> platforms, I can state unequivocally that OS/X is a first-class
> development environment. I am not going to state that it is better
> than "x" or "y", but I can say that it is *my* first choice -- and by
> a wide margin.
>
> So, David, please don't take any of my comments personally, they are
> really directed more at TI than anyone else.
>

Some of your comments were definitely directed at /me/ - but they were
definitely appropriate.  I don't know if anyone at TI has been
listening, but /I/ have certainly learned more about Mac usage from your
posts.

You haven't convinced me to buy a Mac - but if it is possible to get a
trial version of MacOS for a Virtual Box machine, then I would certainly
stretch that far.

mvh.,

David