Developing/compiling software

In message <1fvcb51mrf4o6o1ai4dcrf09aalgjl6kd8@4ax.com>, Jon Kirwan
<jonk@infinitefactors.org> writes
>On Sun, 20 Sep 2009 16:18:42 +0100, Chris H <chris@phaedsys.org>
>wrote:
>
>>>Someday, we should meet and have lunch.
>>
>>I know who to have a s a referee.
>><snip>
>>...and mix with people a bit more
>
>Ironic, coming after suggesting you'd need a referree to meet.

It was a recommendation from people who know you.

-- 
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

In message <n4tcb5lbk7m7stpqrfkmvhfje8ut4gebth@4ax.com>, Jon Kirwan
<jonk@infinitefactors.org> writes
>On Sun, 20 Sep 2009 20:19:13 +0200, "Lodewicus Maas"
><wicus.maas@gmail.com> wrote:
>
>>In the country I'm living - PIC's are really hard-to-get ,and at a price. I
>>started looking at Ebay and found the cheapest option available which can
>>give me 32 I/O ports, and the best offers I could get was on the AT89S52,
>>and this is how I ended up with the ATMEL product
>>
>>I definately agree with a few posts that PIC might be easier and maybe
>>cheaper, but like I said, I had to look at availability/price first, and now
>>I must move on to the next step, which is compiling the code I already
>>written over the past 2 months - without having any compiler or hardware. My
>>programmer arrived  on Friday and as soon as I made up my mind on a
>>compiler, then I can test(compile) the code which is currently only in a
>>.txt file, and hope there is no compilation errors.
>>
>>My AT89S52's should arrive within the next 2 weeks, and only then will I see
>>if the past 3 months was a total waste of time.
>>
>>Thank you for all the input
>>Much Appreciated
>>Lodewicus Maas
>
>Best of luck.  The AT89 is a fine chip for some uses.  I'm not sure
>why you haven't tried to compile the code, though.  As Chris has
>mentioned, there are demo versions of commercial c compilers that are
>available.

There are demo version of the Keil and IAR.  Though I would suggest Keil
in this case as a first stop. It is the one area where they really
specialise.

>And besides that, there is SDCC which you could also do
>some trial compilations with.

Not worth the effort. Start with Keil

>  I'm not sure if the Keil IDE can do
>this (it may work, just fine) but Silicon Labs has an IDE as well for
>their 8051 core cpus and their IDE (and SiLab's web site discusses
>this in an appnote) can integrate SDCC into it, so you should be able
>to run some tests that way.  T

Again not worth the effort. This is a lot of work for a system that will
not even come close to Keil

>here are some slight differences in
>syntax for ports, if I recall, but that's also documented.  At least
>you could have tested for compilation before receiving parts.

The Atmel stuff will be set up to compile with the Keil extensions.  You
will have to port for the SDCC.

BTW there are some very cheap dev kits with the 4K version of the Keil.
However....

The code size limit is not the killer here. It is the DATA space. As
Keil does aggressive data overlaying in the DATA space you can get it to
compiler programs that the SDCC can't get to fit in the space.

I have seen cases where 119 bytes of DATA  was compiled into 23 bytes of
DATA space. So in that case the SDCC would run out of space and the 2K
eval Keil will compile and run the app.

That is without the optimisation of the code the Keil can do compared to
the SDCC.

Then there is the fact the Keil is VERY heavily tested compared to the
SDCC and WILL do what it says.

The you have the fact that most companies, like Atmel, write their code
to be Keil compliant (not SDCC compliant)

As Jon pointed out changing development teams can have an effect.
Whereas the Keil has had a single team all the way though SDCC has AFAIK
changed complete development teams several times.


-- 
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

On Mon, 21 Sep 2009 10:34:49 +0100, Chris H <chris@phaedsys.org>
wrote:

>In message <n4tcb5lbk7m7stpqrfkmvhfje8ut4gebth@4ax.com>, Jon Kirwan
><jonk@infinitefactors.org> writes
>>On Sun, 20 Sep 2009 20:19:13 +0200, "Lodewicus Maas"
>><wicus.maas@gmail.com> wrote:
>>
>>>In the country I'm living - PIC's are really hard-to-get ,and at a price. I
>>>started looking at Ebay and found the cheapest option available which can
>>>give me 32 I/O ports, and the best offers I could get was on the AT89S52,
>>>and this is how I ended up with the ATMEL product
>>>
>>>I definately agree with a few posts that PIC might be easier and maybe
>>>cheaper, but like I said, I had to look at availability/price first, and now
>>>I must move on to the next step, which is compiling the code I already
>>>written over the past 2 months - without having any compiler or hardware. My
>>>programmer arrived  on Friday and as soon as I made up my mind on a
>>>compiler, then I can test(compile) the code which is currently only in a
>>>.txt file, and hope there is no compilation errors.
>>>
>>>My AT89S52's should arrive within the next 2 weeks, and only then will I see
>>>if the past 3 months was a total waste of time.
>>>
>>>Thank you for all the input
>>>Much Appreciated
>>>Lodewicus Maas
>>
>>Best of luck.  The AT89 is a fine chip for some uses.  I'm not sure
>>why you haven't tried to compile the code, though.  As Chris has
>>mentioned, there are demo versions of commercial c compilers that are
>>available.
>
>There are demo version of the Keil and IAR.  Though I would suggest Keil
>in this case as a first stop. It is the one area where they really
>specialise.
>
>>And besides that, there is SDCC which you could also do
>>some trial compilations with.
>
>Not worth the effort. Start with Keil
><snip>
>Again not worth the effort. This is a lot of work for a system that will
>not even come close to Keil

I think it took me a half hour to follow the instructions.

>>here are some slight differences in
>>syntax for ports, if I recall, but that's also documented.  At least
>>you could have tested for compilation before receiving parts.
>
>The Atmel stuff will be set up to compile with the Keil extensions.  You
>will have to port for the SDCC.

True.  But it's been minor issues so far, in comparing them.  Some of
the differences are talked about in the SiLabs appnote.

>BTW there are some very cheap dev kits with the 4K version of the Keil.
>However....
>
>The code size limit is not the killer here. It is the DATA space. As
>Keil does aggressive data overlaying in the DATA space you can get it to
>compiler programs that the SDCC can't get to fit in the space.
>
>I have seen cases where 119 bytes of DATA  was compiled into 23 bytes of
>DATA space. So in that case the SDCC would run out of space and the 2K
>eval Keil will compile and run the app.
>
>That is without the optimisation of the code the Keil can do compared to
>the SDCC.
>
>Then there is the fact the Keil is VERY heavily tested compared to the
>SDCC and WILL do what it says.
>
>The you have the fact that most companies, like Atmel, write their code
>to be Keil compliant (not SDCC compliant)
>
>As Jon pointed out changing development teams can have an effect.
>Whereas the Keil has had a single team all the way though SDCC has AFAIK
>changed complete development teams several times.

The OP has yet to say what this is all about and money may yet be an
issue.  Hard to say.

SDCC is probably not the direction to head if you are expecting the
kind of sole-source support phone numbers and email you get when you
lay down some cash.  No doubt.

Jon

On Mon, 21 Sep 2009 10:24:59 +0100, Chris H <chris@phaedsys.org>
wrote:

>On Sun, 20 Sep 2009 19:24:32 GMT, Jon Kirwan <jonk@infinitefactors.org> wrote:
>
>>On Sun, 20 Sep 2009 16:18:42 +0100, Chris H <chris@phaedsys.org>
>>wrote:
>>
>>>>Someday, we should meet and have lunch.
>>>
>>>I know who to have a s a referee.
>>><snip>
>>>...and mix with people a bit more
>>
>>Ironic, coming after suggesting you'd need a referree to meet.  Still,
>>I'm sure you meant this last part in a positive way so I'll just say
>>you shouldn't worry.  Just met with my Representative over lunch, a
>>few days ago, for example.  Plenty on my plate in that regard.  No pun
>>intended.
>>
>>Jon
>>
>>P.S.  I'd still think it would be helpful to others if you'd expand
>>even slightly on your earlier comment about SDCC.  I'm curious and I
>>did check to see what I could find about what few words you offered,
>>found something that seemed close, and if so it seems to be a bit out
>>of date.
>
>It was a recommendation from people who know you.

Obviously, not as well as you imagine.  In any case, this whole thrust
of yours is little other than an attempt at being personal.  Which is
exactly how you started with your responses here, again.  Rather than
dealing with a direct question, the answer to which may help others if
you'd cared to.

Oh, well.

Jon

OK. So Keil is NOT an option anymore
.
The Demo/Eval version can only compile up to a max of 2K - which I reached 
already. I then requested a quote from the local suppliers of Keil software, 
and the quote is ...
.
.
I hope you're sitting ...
.
.
.
R 39,335.81 ( this is equal to 5,326.93 USD) - for a single user license
.
.
My whole outlook on life is "value for money", and I don't invest in 
anything if this requirement is not met, but unfortunately there is no way 
in which I can justify this as a hobbiest
.
I'm now busy looking at ImageCraft
.
The "doors" keep on closing, but eventually I'll get there. 

Lodewicus Maas wrote:

> .
> R 39,335.81 ( this is equal to 5,326.93 USD) - for a single user license
> .
> .

around 4000 ukp, per seat ?...

That's outrageously more expensive than I remember it last time I 
looked. I would have a hard time keeping a straight face recommending it 
to some clients, as the first response would be "go and find a different 
processor" :-(. There's plenty of choice out there.

I have some sympathy for the vendors as i'm sure they are dedicated 
professionals in every sense, but the availability of good quality open 
source tools for most of the cpu landscape means a diminishing client 
base and less money for development. Unless they can come up with a 
revised business model, they seem condemned to niche markets and long 
term decline. If you want evidence, compare the diversity of tool 
vendors a decade ago with the same now.

The inescapable logic is that proprietary tool vendors can't really 
compete with the vast number of dedicated open source developers who 
continuously review each others work and improve the product with each 
iteration. While some may claim that proprietary tools produce better 
code than gnu, on balance, i'm not convinced. If open source produces 80 
or 90% of the quality of the proprietary tools, based on any metric, 
does it really matter, assuming modern high performance micros with 
large code spaces ?...

Regards,

Chris

Lodewicus Maas wrote:
> OK. So Keil is NOT an option anymore
> .
> The Demo/Eval version can only compile up to a max of 2K - which I 
> reached already. I then requested a quote from the local suppliers of 
> Keil software, and the quote is ...
> .
> .
> I hope you're sitting ...
> .
> .
> .
> R 39,335.81 ( this is equal to 5,326.93 USD) - for a single user license
> .
> .
> My whole outlook on life is "value for money", and I don't invest in 
> anything if this requirement is not met, but unfortunately there is no 
> way in which I can justify this as a hobbiest
> .
> I'm now busy looking at ImageCraft
> .
> The "doors" keep on closing, but eventually I'll get there.

ImageCraft does not make an 8051 compiler - the make (amongst others) an 
AVR compiler.

For hobby use, you have two options.  You can use SDCC for free, or you 
can drop all thoughts about using the 8051.

The Atmel AVR devices are a much better choice for most uses - 8051 is 
good when you have a history of using them, or when you are using 
specialised chips with an 8051 core.

If you go for the AVR, you have several choices.  ImageCraft is a good, 
cheap, user-friendly compiler.  WinAVR is a project based on avrgcc and 
thus is entirely free.  There are a fair number of other compilers 
available for the AVR, some low price and some high price - far more 
/realistic/ options than for the 8051, especially for hobby or 
low-budget use.

In message <3ZidnbQvZPgxSCTXnZ2dnUVZ8g-dnZ2d@saix.net>, Lodewicus Maas
<wicus.maas@gmail.com> writes
>OK. So Keil is NOT an option anymore
>.
>The Demo/Eval version can only compile up to a max of 2K - which I
>reached already. I then requested a quote from the local suppliers of
>Keil software, and the quote is ...
>R 39,335.81 ( this is equal to 5,326.93 USD) - for a single user license

It should be half that.....


-- 
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

In message <jHlum.172734$z%1.21899@newsfe12.ams2>, ChrisQ
<meru@devnull.com> writes
>Lodewicus Maas wrote:
>
>> .
>> R 39,335.81 ( this is equal to 5,326.93 USD) - for a single user license
>> .
>> .
>
>around 4000 ukp, per seat ?...

No. It if half that price in the UK

>I have some sympathy for the vendors as i'm sure they are dedicated
>professionals in every sense, but the availability of good quality open
>source tools for most of the cpu landscape

If that were the case there would be a problem but so far most of the
open source compilers are no where near as good as the top commercial
ones.

>means a diminishing client base and less money for development. Unless
>they can come up with a revised business model, they seem condemned to
>niche markets and long term decline. If you want evidence, compare the
>diversity of tool vendors a decade ago with the same now.

True... So eventually programmers will be expected to produce their own
tools and RTOS on their own time at their own expense.

>The inescapable logic is that proprietary tool vendors can't really
>compete with the vast number of dedicated open source developers who
>continuously review each others work and improve the product with each
>iteration.

Who funds these open source programmers? They have to eat.

The problem is the main users of the open source tools are not the ones
who are producing them...

-- 
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/

In message <4ab9e8bc$0$26317$8404b019@news.wineasy.se>, David Brown
<david@westcontrol.removethisbit.com> writes
>Lodewicus Maas wrote:
>> OK. So Keil is NOT an option anymore
>> .
>> The Demo/Eval version can only compile up to a max of 2K - which I
>>reached already. I then requested a quote from the local suppliers of
>>Keil software, and the quote is ...
>> .
>> .
>> I hope you're sitting ...
>> .
>> .
>> .
>> R 39,335.81 ( this is equal to 5,326.93 USD) - for a single user license
>> .
>> .
>> My whole outlook on life is "value for money", and I don't invest in
>>anything if this requirement is not met, but unfortunately there is no
>>way in which I can justify this as a hobbiest
>> .
>> I'm now busy looking at ImageCraft
>> .
>> The "doors" keep on closing, but eventually I'll get there.
>
>ImageCraft does not make an 8051 compiler - the make (amongst others)
>an AVR compiler.
>
>For hobby use, you have two options.  You can use SDCC for free, or you
>can drop all thoughts about using the 8051.

Or use the 2K Keil or the 4K Keil if you can find one.  These can
compile programs the SDCC can't. However for hobby use the Keil is
expensive but then it is not aimed at hobby users.

for commercial use one of my customers worked out the SDCC cost them
about 5K in time and resources.  So free to buy does not always mean
completely free. Though for hobby use time is free.



-- 
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\
\/\/\/\/\ Chris Hills  Staffs  England     /\/\/\/\/
\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/