8051 C Compiler

<cbarn24050@aol.com> wrote in message 
news:1156714970.374382.41100@i3g2000cwc.googlegroups.com...
>
> Stephen Pelc wrote:
>> On Fri, 25 Aug 2006 11:15:46 -0400, Walter Banks
>> <walter@bytecraft.com> wrote:
>>
>> >Compiler development is a labour intensive industry by very skilled 
>> >people providing
>> >products that service a small market. We have looked very closely at the 
>> >issues
>> >around piracy and it is rarely an issue of price. We have not found 
>> >piracy to have
>> >a price threshold.
>>
>> Having provided Forth compilers for 25 years, I make two observations
>> 1) You'll always get ripped off,
>> 2) According to some people, the "best" price is half that of the
>>    lowest cost product on the market.

I would argue, actually, that the best software is usually either free or 
relatively high-priced.  Bargains are what you want to beware of, unless 
they are cut-down versions of high-end products.

>> Now that gcc and friends have reduced the acquisition cost of C
>> compilers to zero, you will always bleed.
>
> It's always been a mystery to me why C compilers aren't free anyway. If
> youv'e spent millions producing a chip why make it hard for people to
> use it?

The PIC people are doing this with a compiler for their lowest-end chips --  
they've bought a commercial one and are distributing it with MPLAB.

> Compiler writers can easily protect their product with
> a dongle, no excuse now we have usb sticks. The real downside with that
> is when the customer recieves crappy software he can send it back and
> get his money back, which I guess is why they don't use them.

That is a factor! :)

cbarn24050@aol.com wrote:
> Stephen Pelc wrote:
> 
>>On Fri, 25 Aug 2006 11:15:46 -0400, Walter Banks
>><walter@bytecraft.com> wrote:
>>
>>
>>>Compiler development is a labour intensive industry by very skilled people providing
>>>products that service a small market. We have looked very closely at the issues
>>>around piracy and it is rarely an issue of price. We have not found piracy to have
>>>a price threshold.
>>
>>Having provided Forth compilers for 25 years, I make two observations
>>1) You'll always get ripped off,
>>2) According to some people, the "best" price is half that of the
>>   lowest cost product on the market.
>>
>>Now that gcc and friends have reduced the acquisition cost of C
>>compilers to zero, you will always bleed.
> 
> 
> It's always been a mystery to me why C compilers aren't free anyway. If
> youv'e spent millions producing a chip why make it hard for people to
> use it? Same arguement applies to the FPGA market. Some companies do
> give them away free of nearly free, a trend I expect will become the
> norm eventually. Compiler writers can easily protect their product with
> a dongle, no excuse now we have usb sticks. The real downside with that
> is when the customer recieves crappy software he can send it back and
> get his money back, which I guess is why they don't use them.

Quite a lot of compilers DO have free versions, with some ceiling.

  If you look at the newest systems from Freescale and Zilog, they do
have free compilers, (as in bundled with the low cost eval/development 
boards) and not just compilers, but free OnChipDebug as well.

  On chip debug support already an important selection criteria.

-jg

Jim Granville wrote:
> Quite a lot of compilers DO have free versions, with some ceiling.

Yes I know but why do they have a ceiling? To be fair not all have one.

>
>   If you look at the newest systems from Freescale and Zilog, they do
> have free compilers, (as in bundled with the low cost eval/development
> boards) and not just compilers, but free OnChipDebug as well.
>
>   On chip debug support already an important selection criteria.

Iv'e never seen the need for on chip debug but then everything I do is
real time. Most of the really stupid stuff shows up on the simulator.

cbarn24050@aol.com wrote:
> Jim Granville wrote:
> 
>>Quite a lot of compilers DO have free versions, with some ceiling.
> 
> 
> Yes I know but why do they have a ceiling? To be fair not all have one.
> 
> 
>>  If you look at the newest systems from Freescale and Zilog, they do
>>have free compilers, (as in bundled with the low cost eval/development
>>boards) and not just compilers, but free OnChipDebug as well.
>>
>>  On chip debug support already an important selection criteria.
> 
> 
> Iv'e never seen the need for on chip debug but then everything I do is
> real time. Most of the really stupid stuff shows up on the simulator.

  Yes, the silly logical-error stuff can show up in the simulator, but 
SFR (mis)handling does not, and with the newest USB Debug pathways, 
speed is much better : you now effectively get a REAL_CHIP in your 
simulator.

  The biggest problem with simulators, is you are never sure when they
start giving you false information. Very hard to beat doing debug using 
the actual chip you are going to ship!

  It also means suppliers can focus on the Debug support, and the
resource that was put into the simulator, can go elsewhere.

-jg

mc wrote:
> I would argue, actually, that the best software is usually either free or
> relatively high-priced.

Dont EVER assume that a high price is any indication of quality,
fitness for purpose or any other usefulness. It just isn't.

Jim Granville wrote:
>   It also means suppliers can focus on the Debug support, and the
> resource that was put into the simulator, can go elsewhere.

Yes another revenue stream. I'm sure they have their uses I just
haven't needed one just yet.

Op Fri, 25 Aug 2006 10:51:06 +0200 schreef Chris Hills  
<chris@phaedsys.org>:
> In article <op.tetyxsu3y6p7a2@ragnarok.lan>, Boudewijn Dijkstra
> <boudewijn@indes.com> writes
>> Op Wed, 23 Aug 2006 17:01:57 +0200 schreef seba <seba@nospam>:
>>> <cbarn24050@aol.com> wrote:
>>>>
>>>> C and 8051 are a very poor marriage especially it terms of speed.
>>>> Unless you have some overwheming reason to use both together find
>>>> another option.
>>>
>>> I'm just looking for a list of C compilers compatible with Aduc8xx
>>> microcontrollers; I think I will use Keil compiler but, before starting
>>> development, I liked to know if there are good substitutes, other than
>>> Keil...
>>
>> Keil is about to die.

After some more thought, I take the above back.  Please read below.

> Highly unlikely. Why would ARM kill it?  The Keil 8051 and C166 are
> stable and mature products. I can see they may not do much more
> development but I can't see them disappearing.... They are a good
> revenue stream.

I agree.

> What makes you think Keil is about to disappear?

Well it is a possibility.  The reasoning was as follows: {past:} ARM  
licenses their cores to take over more of the 8/16/32 bit market.   
{recently:} ARM takes over Keil to gain compiler knowledge for ARM and to  
control part of the compiler market for '51.  {future:} An evil ARM might  
want to stop the '51 compilers and so persuade '51 users towards ARM.

But that makes no sense.  There are much better ways of persuading '51  
users to move to ARM.

> Even if Keil/ARM push the Cortex in favour of the 166, the 51 market
> will still continue.



-- 
Gemaakt met Opera's revolutionaire e-mailprogramma:  
http://www.opera.com/mail/

Boudewijn Dijkstra wrote:

> Op Fri, 25 Aug 2006 10:51:06 +0200 schreef Chris Hills

> 
>> What makes you think Keil is about to disappear?
> 
> Well it is a possibility.  The reasoning was as follows: {past:} ARM
> licenses their cores to take over more of the 8/16/32 bit market.
> {recently:} ARM takes over Keil to gain compiler knowledge for ARM and to
> control part of the compiler market for '51.  {future:} An evil ARM might
> want to stop the '51 compilers and so persuade '51 users towards ARM.
> 
> But that makes no sense.  There are much better ways of persuading '51
> users to move to ARM.
> 

I think there are very few reasons to persuade 8051 users to us ARM. There
will always be applications for which the 8051 is ideally suited,
especially with its highly integrated peripherals. The 8051 is an order of
magnitude simpler than the ARM such that getting your head around the
entire device and its instruction set is still possible and beneficial.

If ARM kills Keil none of this will change. People will just use a different
compiler.

Ian

Op Mon, 28 Aug 2006 12:19:35 +0200 schreef Ian Bell  
<ruffrecords@yahoo.co.uk>:

> Boudewijn Dijkstra wrote:
>
>> Op Fri, 25 Aug 2006 10:51:06 +0200 schreef Chris Hills
>
>>
>>> What makes you think Keil is about to disappear?
>>
>> Well it is a possibility.  The reasoning was as follows: {past:} ARM
>> licenses their cores to take over more of the 8/16/32 bit market.
>> {recently:} ARM takes over Keil to gain compiler knowledge for ARM and  
>> to
>> control part of the compiler market for '51.  {future:} An evil ARM  
>> might
>> want to stop the '51 compilers and so persuade '51 users towards ARM.
>>
>> But that makes no sense.  There are much better ways of persuading '51
>> users to move to ARM.
>>
>
> I think there are very few reasons to persuade 8051 users to us ARM.  
> There will always be applications for which the 8051 is ideally suited,
> especially with its highly integrated peripherals. The 8051 is an order  
> of magnitude simpler than the ARM such that getting your head around the
> entire device and its instruction set is still possible and beneficial.

For many applications, complexity tends to increase over time, creating a  
need for faster processors.  The step from Keil for 8051 to RealView for  
ARM is not so big, because it's all ARM.  And with a good IDE and  
C-compiler you don't need to get your head around the entire device and  
it's instruction set; at least not all of the developers in a group.

> If ARM kills Keil none of this will change. People will just use a  
> different compiler.

Exactly.


-- 
Gemaakt met Opera's revolutionaire e-mailprogramma:  
http://www.opera.com/mail/

In article <44f1b89c@news.wineasy.se>, David Brown <david@westcontrol.re
movethisbit.com> writes
>The number of people who illegally copy a software tool is irrelevant,
>at least in the short term (it is important if your aim is to damage
>other manufacturers as much as to increase your own income - I don't
>think that applies in the embedded tools market).  What is relevant is
>the number of people who use an illegal copy who would otherwise have
>paid for it.

Or bought someone else's. I.e. I didn't buy Keil because I got a hacked
IAR.

>  That's a very different number - and I believe in many
>cases, the number is so low as to be hardly worth considering.

No so. In some markets piracy is 90% of the market. There are some areas
where if there are several SW engineers on a team you only get to sell
one compiler for any given target....

I have seen companies in Europe who try to get out of buying compilers.
USB dongles and parallel port dongles on hubs/switches are not uncommon.


>  But
>then, my opinions are based on personal use, people I've talked to, and
>sources like this newsgroup - people who sell these tools presumably
>investigate the issues more scientifically.

A bit difficult... will everyone who uses pirated SW please register so
we can count you :-)  the other problem is that compiler companies are
not going to shout about it either.  Though there are ways of
identifying illegal compilers is many cases.

Due to the institutionalised piracy in some places it does not occur to
some that they should not call the compiler company for support even
with a hacked copy!



-- 
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
/\/\/ chris@phaedsys.org      www.phaedsys.org \/\/\
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/