[anyone]

Switching from 5V designs to 3.3V designs is already a burden. Is more expensive and designs require more space or higher density. Having to deal with lower currents just increase the complexity. But this mean only two things:

1. Many are not ready to make the switch so, the 3.3V boards are not for them (not that they are not "intellectual" capable of making the switch);

2. For those beginners, brave enough to dive into, there must be a solid documentation and I mean not only formulas. There must be detailed examples of interfacing with various peripherals to avoid damages to the board as much as possible (is not an easy job to "move" folks to a "new system"), as you find on PICAXE documentation.

The "pros" who waited for Due are not considered in my comment - they know what to do with an ARM and a 3.3V board. And also, they must understand that not everybody is a pro and there can be differences from person to person (considering that anyone will take by storm the Arduino store and the local distributors to get a Due).

A good hardware reference about how it should be a noise immune design is here.

Thank you DSilva, you're fantastic! The 2) and 3)are accomplished. Now it should be easy for Arduino team to move the core to it. The Atmel will follow us soon after that

Vasi

P.S. Maybe they (Atmel) will reconsider and switch to Eclipse IDE to cover all the operating systems (and that will be a smart move). Microchip is using NetBeans IDE for their MPLAB X and they are doing a great job (in both the IDE and C compilers which are now cross platform). It would be pitty for Atmel to consider that engineers are tided only on Windows world. Anyway, this is another discussion - let's do things step by step.

Don't know how aggressive... maybe not at all. But as long as I am aware of a better avr-gcc compiler than the current one, then I'm starting to complain, making requests, and in the end, making the switch. It doesn't need to be exactly 4.7.2, but right now, 4.8 produce bigger code and incorrect in some cases.

Regarding to frequency of changes, if 4.7.2 proves to be better and stable (until now, it does), then I want the switch and, until 4.8 proves to be better, it can stay there for years.

Thank you for the info.

1) So, the first move is expected from Atmel... I think that will take a little longer than expected...
2) This has been done. MHTools, but it needs to be updated with Georg-Johann Lay's package.
3) The 4.7.2 version is fresh, not yet an OS X version... I think...
4) I think I saw somewhere on Internet a tutorial about migrating to gcc 4.7 (including core changes) ... I have to search for the link again (I started to use Arduino 1.0.1 with gcc 4.7.0)... found it

If Atmel is excluded, then maybe the migration to 4.7 will happen sooner? Well, I forgot that also Linux must be able to "deliver" 4.7.2 in the most important distros (but I've seen Arduino for Linux packaged with his own version)... Fedora will have it at version 18.

This is a post I'm watching on avrfreaks.net

Didn't, thanks for the link. Actually, I'm using Arduino IDE only occasionally. I use Eclipse with avr-gcc 4.7.0 on Linux Debian Testing (not yet 4.7.2 in a linux distribution to my knowledge) and with avr-gcc 4.7.2 (thanks to Georg-Johann Lay ) under Vista. 

Smaller code. Also it comes with a lot of bug-fixes. Maybe is not so obvious for the few avr's used by Arduino language but, considering that the trend is to extend to other microcontrollers, then 4.7.2 will be better. And for many, Arduino is supposed to be a launch pad to AVR's world and "raw" C/C++ programing. Is good to have a correlation.

I'm always for the better version of gcc compiler (and for "shrinking the projects"  ).

It seems that avr-gcc 4.7.2 version is the greatest for avr's. When will Arduino IDE migrate to it? The sooner, the better!

Vasi

http://www.arduino.cc/playground/Learning/CommandLine

However I suggest to not go for make but for scons.

https://github.com/suapapa/arscons

scons is much easier to pickup and maintain.

Hi Klein,

Consider also this scenario, from the MPIDE forums:

I would like to send a hex file to a customer and give them the ability to upload without sending the whole project.

. I've posted the link on the previous post. Make is (a solution but) definitely not for them (clients), and is scaring for beginners.
Anyway, is not my problem (even if it is a real one considering all the cases) as I'm successfully using avrdude-GUI with the last version of avrdude. The OP from the MPIDE forums said also:

I wish MPIDE could be used to upload hex files without compiling, that would be easy!

Best regards,
Vasi

In other words, why Arduino IDE does not have an interface similar to avrdude GUI? Considering that one of the goals is to support other AVRs as well, I think is time to have the required tool to upload the .hex files.

Why we need it... Imagine that maybe I'm making boards for others, boards which use different ATmegas and I need to upload hex files to test the board functionality, frequently switching from one ATmega to another (and obviously, from a hex file to another). For those hex files, I need to compile only once. But there can be various scenarios...

For the start, it will be good to have a GUI interface to Optiboot bootloader.
BTW, Optiboot bootloader support ATmega168P which is not yet supported by Arduino IDE and such interface will help if you decided to go beyond Arduino language.

_____________
Edit:
A funny thing: the users of MPIDE wants the same thing!

1. I'm relating to realities in my zone (is all what it matters).
2. For RepRap is there a project started for Chipkit boards. And expect more where speed is needed.

My post suggest progress to Arduino platforms (and that you can't ignore "the signs"). Google chose Arduino platform for their ADK assuring Arduino existance for a long time from now on but also here are more powerful (and cheaper?) platforms for that.

But I'm sure Arduino developers will take into account all of these.

Vasi

[prices are already high]

I see Mircochip has come out with the chipKIT32....its compatiable with arduino shields and has more I/O pins then the ONO,,,should I be worried that this could be the end of the Arduino???????

There is a lot more to the Arduino then just the hardware. I'm certainly not concerned, are you?

Lefty

Indeed, Arduino can be in trouble if:

• - there are supply problems and prices are increasing (or prices can't be lowered); - prices are already high
• - all the projects which made Arduino famous will be fully adapted (or their authors will migrate) to Chipkit and other compatible boards - is happening already

Frustrations in making transition (software and hardware problems) to Arduino clones can strengthen the Arduino position. But, something must be done for sure. You can't just sit on your leader chair without progress. Still, 5Vcc is a barrier to many hobbyists and Chipkit example in adapting a 3.3V chip to be a 5V board is to be followed.

Also, "the model" Microchip is making in quietly supporting  the Chipkit boards with specialized engineers is definitely to be followed. I hope Atmel is reading (or someone pointing them to) this.

Vasi

JAL language started as a subset of VHDL language and now it is close to Pascal. No chance to compile C language . But is the best free compiler around here which can fight with commercial ones from equal positions. It is the best for PIC12/16 families and good enough for PIC18 family. Going down to assembler? With a good high level compiler as JAL you don't need to. But if you want, JAL include also assembler. JAL is easy to learn.
The free C variants (we also have C++ for PICs) here are in no way better than JAL in the resulted HEX file.
Unfortunately, and I don't know why, Microchip don't want to release for free his PIC C compilers and I mean the ones with the full optimization. We always wanted to have a GNU C compiler as AVR programmers have.

V-USB is great indeed, it would be great to be ported also for PICs. It can't compare with a HW USB (you obtain with V-USB only 1Mb/s transfer and the HW can do 12Mb/s) but is still great.

It is a little unfair to compare the prices for ATmega328 when it can't be supplied anymore in DIP package - is missing also form two online shops from my country, where I have access. We have plenty of AVRs here - Romania is in Europe .
 
But is not about me. Is about South America and other countries with difficulties in obtaining ATMEL microcontrollers. And when this is happening, an ATmega will be always expensive. If you read the thread from the beginning, you can have a right image about. For them, Pinguino is a nice solution. For others, is fun to play with it and to contribute to a better language compatibility.

Yes, is Google's fault - I'm from Romania.
Working with peripherals and registers can be easy when you understand the mechanism. It was easy for me to (after I saw an example) set a digital pin as output and input (beyond pinMode() and digitalWrite() ) when I tried for the first time an ATmega32/644P after months of PIC programming. No big deal. Only at configuration bits I got toasted   . But only once.

Thanks to mikroElektronika, I can program in C, Pascal, Basic, both types of microcontrollers, and using even the same board layout. So, no problems at all. See my site:
 - ATMEL online only Arduino language for now.
 - Microchip various languages. BTW, bank problem is automatically hidden by many compilers/languages so programmer is not aware of.

Yes, here are PICs with hardware UART and others without (in this case you can use only software serial).
Now Microchip produce a PIC at 16MIPS, with ATmega328 performance but double flash and RAM memory. You can build with it a compatible Arduino/Sanguino board, at hardware level, using the same FTDI chip or using a variant with USB native. But I think you saw already many boards Arduino like which use different micrcontrollers - also here not big deal.

What I'm trying to say is, you must be able to deal with any microcontroller available on your zone, in a similar way or better. Arduino language was chosen because of the huge existing database. And it proved to be a bridge between the worlds.  

Só para adicionar que existem ainda chips que comunicam através de UART... coisa que um chip com USB provavelmente não terá.

Nope, UART is also there, ready to be used. Yes, you can have two variants of ported libraries which use UART - with Serial USB, or with UART if you want to adapt the code for yourself.

Mas cada caso é um caso e deve ser ponderado no inicio dum projecto. No entanto, só de pensar no pesadelo que é configurar um PIC com as mudancas de bancos, prefiro ter mais um chip na placa e usar AVR.
Existem já AVRs com USB, mas são de 32 bits e mais virados para o mundo dos DSPs.

I don't argue on this. For a poor student, the availability and price of AVRs are the main problems where he lives. So, an alternative had to be created. This is what is Pinguino today. An alternative with a cheap and common microcontroller. If ATMEL problems (the ones above) will ever change, be sure that the right micrcontroller will be chose.
But we can't wait so we must continue to prepare students, learning them Arduino language, being able one day to address any of microcontrollers, better prepared than others... I hope you now understand the big picture.

Vasi