embeddedgurus.com: "An open letter to the developers of the MPLAB IDE"

[Rufus]

atmel understands something mcp missed: planting seeds.

And maybe Microchip understands something Atmel missed: stoney ground.

[retrolefty]

 I also think that the 'arduino' platform has not had nor ever will have an impact or caused any major changes to the micro controller industry.

 Rather the point is that it has had a very big impact on many thousands of new users that would have otherwise never had any exposure or use for a micro-controller. It simply created a new very successful niche market.

 I'm sure the original Arduino developers were as surprised as anyone else with the results of their initial project. Some of those new users will go on from there and enter the more traditional path to become professional EEs where they might not have ever been interested in the field.

 But for most arduino users it's just a simple and inexpensive way to have lots of fun building stuff. The Arduino project should be applauded for what it has accomplished, not scorned or berated for not being professional enough way of getting started. At least they get to start by learning C/C++ rather then Basic. 

[dannyf]

its about the learner and home diy'er.

Then maybe you should have said something like:

I see no new *****learner and home diy***** projects that are pic based compared to the numerous ones that are arduino based.

this is a very smart business approach.

Yeah. Except that that approach, if 100% successful, wouldn't amount to a rounding error for Atmel.

How many such "smart" approaches Atmel has to make to overcome its competition?

[dannyf]

atmel is totally winning over the new generation of kids

My initial comment wasn't really about Microchip vs. Atmel - both, in my view, have a bleak future.

This year is the 3rd year over the last 5 years that I have visited China for mcu-related seminars. PICs are practically not followed there; AVRs had some followings and many people there are switching to ST (STM8 or STM32). You would say the biggest and most enthusiastic participation at those 32-bit ARM seminars (Freescale and Renesas have their own following, also very strong). And the enthusiasm got bigger each time as time went by.

If I were Atmel, particularly Microchip, I would be very scared.

I think there will continue to be a market for 8-bit mcus, just as there is still a market for 1/4-bit mcus here now. I would be surprised if it is substantial enough to support a handful of shops: how many of you know of anyone working on MC14500?

The glory days of 8-bit mcus are probably behind us.

[linux-works]

it seems that every controller based board I get from ebay/china uses the VERY ancient 8051 series chips!  big ugly wide-DIP chips, not even small nice ones like atmel and pic

their toolchain is at least 20 yrs behind the west.  I can't understand the 8051 love over there, other than maybe its sourced by lots of makers and its cheap as, well, chips.

the western DIY market, though, shows no signs of slowing on 8bit cpus.  especially the thruhole ones.  I avoided the mega-arduino chips because they are not diy solder friendly and I want my users to be able to solder and repair their own boards.

as long as there is a DIP version of a controller that has full open source support, I'm happy.  it does not have to be an arduino, but anyone who plays games with me by charging me to REMOVE no-ops - I will refuse to do business with such companies if I can at all help it.   each time I think of them inserting slow-me-down instructions as punishment, it fills me with anger!!  let them sell hardware.  the HELL with trying to sell software; its not their market and they are stupid to think this has to be a profit-center.

[dannyf]

it seems that every controller based board I get from ebay/china uses the VERY ancient 8051 series chips! 

Does that include the "fake" Arduino boards you got from China?

I can't understand the 8051 love over there,

Maybe it has something to do with their functionality? Value? Availability? ...? Try to find a PIC or AVR with 24-bit onboard ADC; or onboard CPLD? ...

the western DIY market, though, shows no signs of slowing on 8bit cpus. 

Nor did those MC14500 guys in 1982. Where are they now?

I avoided the mega-arduino chips because they are not diy solder friendly and I want my users to be able to solder and repair their own boards.

That's a valid concern if you don't have access to affordable PCB services.

the HELL with trying to sell software; its not their market and they are stupid to think this has to be a profit-center.

The beauty of a free market is that you can do whatever you feel like, including being irrational.

Me? I look at what I get vs. what I pay and make a decision on that. The vendors can make whatever they make on me, however they make that on me, as long as I get good value for my money.

[janoc]

atmel is totally winning over the new generation of kids

My initial comment wasn't really about Microchip vs. Atmel - both, in my view, have a bleak future.

This year is the 3rd year over the last 5 years that I have visited China for mcu-related seminars. PICs are practically not followed there; AVRs had some followings and many people there are switching to ST (STM8 or STM32). You would say the biggest and most enthusiastic participation at those 32-bit ARM seminars (Freescale and Renesas have their own following, also very strong). And the enthusiasm got bigger each time as time went by.

If I were Atmel, particularly Microchip, I would be very scared.

Atmel has their own ARM line too. However, I do agree that most new stuff is easier and cheaper (!) to design with something like STM32F1xx or F3/4 series than an 8bit PIC or an Atmel micro. Where the 8bit chips have advantage is the simplicity. Even STM32F103 series can be totally overwhelming for someone upgrading from a smaller micro.

The glory days of 8-bit mcus are probably behind us.

Probably not new fancy projects, but cheap, tiny and simple to program 8bit micros will have their place for a long time to come for simple things - that's why the ancient 8051 architecture (and the lack of royalties) still hangs on. This is where Microchip has a certain advantage - their chips are relatively cheap. On the other hand, ATMegas and XMegas are not at all priced competitively, IMO - I can get many times more powerful STM32 for half of the price of an ATMega328 ...

[linux-works]

it seems that every controller based board I get from ebay/china uses the VERY ancient 8051 series chips! 

Does that include the "fake" Arduino boards you got from China?

I can't understand the 8051 love over there,

Maybe it has something to do with their functionality? Value? Availability? ...? Try to find a PIC or AVR with 24-bit onboard ADC; or onboard CPLD? ...

I avoided the mega-arduino chips because they are not diy solder friendly and I want my users to be able to solder and repair their own boards.

That's a valid concern if you don't have access to affordable PCB services.

I don't order arduino boards from china; I have my own arduino board and I buy real chips from US suppliers.  maybe you were thinking of someone else?

I have not found where the audio boards I sometimes buy from china need ANY of the things you mentioned.  the answer is simple: they are cheap, the tools have long been pirated (ie 'free' now) and the chips are probably very cheap.  otoh, they are huge chips and waste a lot of board space.

the boards I have bought were DACs or preamps and all they needed to do was to receive some IR, decode it and set some registers in the main chips.  NOTHING that could not be done with ANY other controller.  8051 gave no magic here, sorry to disappoint you.  they were chosen, like I said, because they were cheap, multiply sourced and the software is most likely pirated and they have enough people who know the 8051 well enough.  I have not heard of anyone seriously using 8051's on any new projects in well over 15 yrs.

as for soldering, if you are making a DIY kit, you should assume thruhole ability only.  to assume more than that is to limit your market and make things harder than they have to be on your users.

[dannyf]

I have not heard of anyone seriously using 8051's on any new projects in well over 15 yrs.

Time to broaden your horizon again,

[janoc]

go to hack-a-day

I said this earlier so let me be blunt: if you want to know how a product does in its market, go look at how it does in its market, not how it does in *****some segments***** of its market.

Hope it helps.

I think it is massively unfair to compare a general market share versus "projects by kids". The latter is overwhelmingly dominated by Atmel at the moment, thanks to Arduino and its clones, both ATMega and ARM based.

I don't think anyone was trying to generalize this to a total market share - Atmel is a flea when compared to Microchip's dinosaur there.

However, it is not really by anything of Atmel's doing, really - they simply had a suitable product on the market for which open/free 3rdparty tools were available. If Microchip had reasonable tooling, they would have had that market segment too.

I don't see why so many people feel compelled to express their disdain and scorn for Arduino users -  sure, these are rarely well engineered projects and many of these people are naive/ignorant but it is a tool that allows many more "mere mortals" to start building stuff. Give them a break ...

[gregariz]

The glory days of 8-bit mcus are probably behind us.

I heard that in 1990.

Microchip responded with the PIC10 series. As of yet I haven't seen any serious ARM contenders that compete with the PIC10 price point for low end appliances/gadgets.

[violet]

It always seemed to me that atmel are good for bringing people into the digital electronics world, and some of the things being produced are great.

 Personally I do prefer microchip, they feel much more solid and yeah probably a bit more hard work but so much more satisfaction at the end of a project (that's the important thing to me regardless of what I use). The products also seem more robust, and time tested but that just my personal opinion.

 I've never had to pay for using microchip products in my personal projects, their samples system allows me to get the physical devices for free and if i'm not doing anything too difficult, then stick to the PIC18 series and there's always been free software. If you want to take something to production then yeah of course youve gotta pay, your'e going to make some money so don't you think they deserve some for giving you that opportunity? It's like someone giving you a room to design a shop layout in for free but if you want to start selling things from it then you've got to pay..

 Microchip also have a very large product portfolio which gives them experience...

 What Atmel have is an easy to use product, and it's easy to become comfortable and famliar with that. I'm used to being challenged with a lot of different controllers so don't feel any real commitment to a single solution, plus prototyping on an atmel because it's easy for presenting a proof of concept but do they really stand up to production costs? I don't think so... (not to mention custom branding).

 I don't want to be positive or negative to the whole microchip/atmel argument but it really does depend which way you look at it.

[linux-works]

I have not heard of anyone seriously using 8051's on any new projects in well over 15 yrs.

Time to broaden your horizon again,

care to point to any current (in the last 2 yrs, say) 8051 diy projects?

I can provide literally hundreds of arduino project that are prominent in diy over the last 2 yrs.  I could find a percentage of that for pic.  and maybe a percent of that percent for 8051.

can you give examples as to this so-called existence of 8051 projects that I somehow have not seen or heard of?  I'm pretty active in the diy world and I have not heard of a single 8051 project, but please, do let me know if I've missed any good ones.

[dannyf]

I'm pretty active in the diy world...

If you still cannot get it, there is no point in getting it for you.

[gocemk]

This letter was written back in 2011!!! And it is about the old (outdated) MPLAB IDE. The main IDE now is MPLAB X and because it's based on Netbeans it is supported on Linux too. And it is MUCH better than the old MPLAB IMHO. Although there are still some folks out there who claim the IDE is awful, i have to say that my experience with it is VERY positive. Lately i'm doing some work with Keil, and it crashed on me twice, without any warning! One of the crashes corrupted my project file, so i had to start over by creating a new project! Never had that experience with the MPLAB X.

Microchip offers FREE editions of their compilers for all their families, and XC8 has become better since version 1.21. And if you want you can still use the legacy HI-TECH C compiler for the 8 bits in MPLAB X in lite mode. And PIC18 + XC8 free or HI-TECH C lite will OUTPERFORM the Arduino anyday, so it is your loss for not using the PIC.

As for the demise of the 8-bitters, it will happen, but only for more complicated projects. Currently i'm working on 2 projects that are so trivial, (simple user interfaces) even the PIC16 series is overkill. I think that the 8-bitters will continue to hold this (low end) market for the years to come.

[linux-works]

is it true that they do insert no-ops to cripple the free versions?  is that only on some architectures?

it would be great if that was a false rumor.  I just can't fathom using a tool that knowingly slows down your generated code.

[dannyf]

they simply had a suitable product on the market for which open/free 3rdparty tools were available.

I would argue that in the 8-bit world, Atmel's products are light year ahead of Microchip's: speed, hardware multiplier, vectored interrupts, adc, etc. Basically the latecomer's advantage.

If Microchip had reasonable tooling, they would have had that market segment too.

Microchip benefited from being the first mover - specializing in the mcu market: 8051 has a considerably better architecture but was slow in developing a full set of (onboard) peripherals. This allowed Microchip to build up a huge user base with its broad product offering.

It is that very advantage that in the end will drag Microchip down. I think their strategy over the last 10 - 15 years has focused too much on milking that cow (playing defensive), rather than developing a way to allow their customers to migrate up market (playing offensive).

[gocemk]

is it true that they do insert no-ops to cripple the free versions?  is that only on some architectures?

it would be great if that was a false rumor.  I just can't fathom using a tool that knowingly slows down your generated code.

This is/was true for the XC8 (which is actually rebranded HI-TECH C). I am using the old HI-TECH C 9.80 lite, because i'm used to it and it supports the PIC's i use more frequently.

[dannyf]

This is/was true for the XC8 (which is actually rebranded HI-TECH C).

Not intentionally inserting NOP. But simply generating overly lazy code.

I am using the old HI-TECH C 9.80 lite

You may want to try pre-OCG compilers, both standard or pro versions.

The newer XC8 compilers are actually very respectable. I tried it on a PIC18 and the free XC8 is just shy of the pro PICC18 and the pro XC8 is quite a bit faster than the pro PICC18.

[mikeselectricstuff]

Do you really think Microchip gives a toss about OSS projects using their chips or not?

Microchip however have always been DIY friendly, way before Open Source was trendy - parts are always readily available, and who else makes a wide range of 16 and 32 bit parts in DIP?

They definitely screwed up on the free 8 bit compiler, but the free versions of their 16 and 32 bit compilers are just fine.
 

[mikeselectricstuff]

I would argue that in the 8-bit world, Atmel's products are light year ahead of Microchip's: speed, hardware multiplier, vectored interrupts, adc, etc. Basically the latecomer's advantage.

But when 16 & 32 bit parts are cheaper, who cares?

[mikeselectricstuff]

Microchip clearly think it's more important to cater to the needs of commercial users than some random project seen on Hackaday still on a breadboard.

Despite a few minor detours into AVR and ARM in t hepast, nowadays my first choice for any product, or 1-off design is Microchip, and many of the reasons are more applicable to commercial products than DIY, and almost none related to the architectural details so often argued over :

I can buy preprogrammed chips from MicrochipDirect for a few pennies more. AFAIK no other MCU manufacturer does this, at least for small (<1000) qtys, or as easily. Even if all that gets preprogrammed is a bootloader, this can save a lot of time and hassle in production. 

The peripherals and devtools for the whole range form PIC10 to PIC32 are sufficiently similar that there is minimal effort to switch between them. I still use MPLAB 8.something, and I can use it for the whole range. 
In general the compilers are set up by default so you don't need to do any dicking around with header files or linker configs unless you're doing oddball stuff like bootloaders. Just #include <pic.h> or <xc.h> and off you go.

I can use the same programmer across the whole range, with an identical hardware interface.
For production test jigs to give to an assembly house , I can just throw in a PicKit 3 in standalone mode for single button idiot-proof programming.

Most parts available in multiple packages, often DIP,SO,SSO and QFN are all available. (and I mean actually available, like I can get a few hundred tomorrow, or preprogrammed in less than a week)   

[gregariz]

its not always the best chip for the task, but the fact that the devel tools work on all 3 platforms and are fully free (and pretty mature, too) seals to deal for me.

I have on occasion used arduino code from the web to try out a new display or chip. For that I have an Uno lying about somewhere's. When I have confirmed the hardware works and am ready to start developing with that new hardware I plug it (usually a shield) into a Microchip based 'Arduino-ish' board and port the code into my preferred compiler. It's usually fairly easy to bring the code across and get it running. Once I have a basic framework running I will then lay out my custom cut-down hardware.

I tend to stay away from chip-vendor supplied tool-chains. I find third party compiler vendors do a better job of documenting the compiler. I also find they do a better job of consistently producing device drivers and libraries. So compilers I will happily spend some money on if they save me time. But now I have a couple I can't see myself going to MPLAB or any other chip vendor. Last time I checked the arduino environment didn't even have a debugger.

At the end of the day you are wanting to release a product into the wild at as low a cost as possible and with the ability to be upgraded by the user. So with a huge selection of chips (some rediculously cheap and many pin compatible upgrade paths) and a good supplier of bootloaders Microchip have the commercial market.

Any hobbyist may prefer to use the arduino to cook up something and that will probably continue, but for the reason's above that's probably the ceiling for the penetration of the platform.

[miguelvp]

Renesas (Hitachi, Mitsubishi & later NEC) has nothing to worry about at least not for a while.

Atmel has been growing strong and surpassed Microchip in 2010 or so.

http://www.dataweek.co.za/article.aspx?pklarticleid=7412

But not because of the maker community. Microchip went from way down the list to the top on the previous decade leaving Atmel, Zilog and others in the dust.

The funny thing about that article is that you gotta see Cypress (minus smartcards and automotive) growth of %139. Love their PSoC chips. And that is for MCUs it doesn't count USB chips. Then again Automotive is huge for sales, but it shows that Cypress is climbing up big time.

Edit: forgot to attach the picture, doesn't include smartcards and automotive mcus.



8bit mcus are pretty much at no growth but not loosing shares, 16 bits are declining and 32 bits are climbing (mobile stuff of course)

But 8 bit is going to always be king in price and battery life. All those cheap mp3/mp4 players have an 80C51 embedded in their multimedia chip, some PSoC chips have it too.

TI has been stale for a while but not loosing ground. Motorola?? well maybe they get royalties or something, Zilog, Toshiba, etc, might be getting royalties as well.

But even the lonely Z80 probably has shipped more chips than Arduinos and derivatives.

In any event, this was about an open letter, not about who can pee further away.

Myself I don't like the PIC branch instruction set (lack off) for anything lower than the PIC24 and they are very spartan as for the instruction set, but that implies less power consumption and longer battery life at a lower manufacturing cost.

Different strokes for different folks.

[scientist]

I can buy preprogrammed chips from MicrochipDirect for a few pennies more. AFAIK no other MCU manufacturer does this, at least for small (<1000) qtys, or as easily. Even if all that gets preprogrammed is a bootloader, this can save a lot of time and hassle in production. 

IIRC Digikey can do this for most of the atmels they have in stock, at no expense in basically any quantity (but I haven't tried it yet).