Hi All

(Free and/or Open Source Software)

As the proprietary vs FOSS battle heats up and the US legal system is challanged with history and technology details they are ill equipped to deal with, I have been trying very hard to forecast what contortions FOSS introducers will be exposed to before the inevitable adoption. How, in this, one of the most fiercely proprietary arenas, where extraordinary means of "protection" are the norm, will the transition take place? Will we see the sue-crazy suicidal tactics of SCO? Will we
be ignored and slowly replace the holdouts? Will automation customer demand "pull" the market before pressure from corporate IS and IT "pushes" change? Or will we simply trail 10 years behind the times? Put on your futurist hats and weigh in. The dialog has changed already
to put Open as opposed to "open" as a very desirable thing. Will it happen or not and how? I'm sure our vendors will be interested as well :^)

Regards

cww

 

I'm not good in weather forecasting more than a week. And I feel nobody is.

We have to do it.
My contribution is:
http://pvbrowser.org

regards,
Rainer Lehrig

 

Well, Curt, as we don't really have any viable OSS product offerings yet (other than at the margins), perhaps its a little early to tell? I've been hearing on this list for more years than I care to remember how soon MatPLC (or was it PuffinPLC?) is going to be great, but I must confess I've never seen one anywhere, and mentioning it to customers tends to provoke a bit of a blank stare. And nobody I have ever met who is in a position to actually sign cheques in payment for kit is remotely bothered that configuration tools run on Windows rather than Linux.

I'd respectfully point out that you have asked a very leading question, anyway, which presents a classic false dichotomy (i.e. Open is always good, everything else is part of an industry/microsoft conspiracy to repress the customers and that evil tactics will be required to put the valiant 'FOSS' down). The truth is much more shaded.

Automation technology does trail consumer electronics, probably not by 10 years, but it does trail. There are good reasons for this, firstly the teams working on the products are smaller and skilled in areas other than the use of leading edge technology, i.e getting real time systems working robustly and reliably. All the same good ideas do get pulled into the mix. But much more importantly, and something you seem consistently to have failed to understand in all the years you have been posting on this list is that the investment cycle is much longer, and the users don't see reasons to change. They are in the business of making stuff more cheaply than they can sell it, not supporting technological ideologies; the safest bet is always the best bet, and is usually to do what worked last time. There are some technologies - fieldbus for example, probably radio networking next - that offer benefits that will shift them up a gear, but the norm is that they don't, and this is how it should be.

Notwithstanding this, the industry is not dominated by a single company like Microsoft, but is viciously competitive. There is constant pressure for box price reduction and product enhancements, and as we've seen in other posts today, companies like Wago (or PLCDirect) are snapping at the big guys' heels perpetually. And here's the rub: ultimately this price competition will move the focus of product development to companies where labour rates are cheaper, and all of these programmers in the West with the luxury of some free time to work on writing code for nothing will find they have a lot more free time to do it in, but considerably less luxury.

Anyway, closed has benefits, as I've argued before. The existence of a standard industry wide platform from Microsoft has meant that smaller vendors don't have to spend time chasing a myriad of offshoot targets for their toolset - ultimately this means better tools. And most actual users (the poor deluded fools) seem to think that the child-stealing conspiracy that is OPC is actually pretty neat.

Best, Tim Linnell

 

I see industrial automation products lagging behind IT for years to come. However, I envision the current problems with free & open software will dissipate.

The Legal systems of North American and yes, the European countries too, are all struggling to catch up to the full implications and impact of Open Source Software. However, they're all learning. Eventually, I expect that the two models will remain, each as a viable option.

I expect that vertical market (meaning specialized application) software will remain mostly closed source. Commodity software will find plenty of open source advocates to support it. Clearly one can see where each has it's place.

The problem within automation is the lack of open standards. This is something which many are dealing with right now. For example, it wasn't that long ago that the coordination for the '1131 standards would have been almost unthinkable.

Fieldbus is yet another standard which has taken forever to become a reality. In fact, it took so long that some vendors such as Siemens decided to fork the standard toward something they thought they could get a smaller subset to agree on. That's what begat Profibus.

Without these standards, nobody is going to write open source code. Without open source code, nobody is going to keep these vendors from inserting "customized tweaks" unique to their gear.

Some day in the distant future, I see standards bodies actually writing open source code to illustrate the behavior they expect their standards to adhere to. An example of this is the OPC specification which includes some VB code snippits on exactly what they expect to call their routines with and how it could be used.

Having example code is just the beginning, however. I envision a day when new protocols come about, based upon an ANSI reference language and with a complete sample implementation in that reference language. At that point, the only question is what hardware some manufacturer would use and what features they would enable.

We are finally graduating from the proprietary everything systems and moving toward systems with standardized open source parts. In the future, the Integrator will be the market maker in this industry, not the device manufacturers...

There. Those are my ideas. They're just mine; they're full of lots of delusional nonsense; and most likely, they're probably all wrong. But it's free and my source is my very own limited experience. YMMV, IANAL, IDSFME (I don't speak for my employer).

 

Hi Tim

You bring out one of the things that will drive adoption if experience in the rest of the embedded industry is considered. It would be difficult to compete with a competitor with drastically lower IP costs. Although I don't know if PLC volumes are high enough where this is
significant. There is a substantial probability that one of the chasers will see advantage in a me-too world by being innovative. Biting sarcasm aside, and the point about leading taken, I'll classify this as a "we'll ignore it and see if it goes away" response.

Regards

cww

 

Hello,

On May 19, 2004, Tim Linnell wrote:
> Well, Curt, as we don't really have any viable OSS product offerings yet
> (other than at the margins), perhaps its a little early to tell? <

Well, Curt did ask for your "futurist hat"...

> I've been
> hearing on this list for more years than I care to remember how soon MatPLC
> (or was it PuffinPLC?) is going to be great, but I must confess I've never
> seen one anywhere, and mentioning it to customers tends to provoke a bit of
> a blank stare. <

Your customers are hiring you for your expertise in the field; you are the one who should keep up with emerging trends and add them to your toolbox, or not, as seems appropriate.

As for the MatPLC, it'll be ready when it's ready. For some purposes, it's probably ready now, for others it may yet take a few years. If you agree with the concept, we invite you to join us and help us forward. If you find what we have already useful, we invite you to take advantage of it - you know our terms (basically, give source to the customer).

> I'd respectfully point out that you have asked a very leading question,
> anyway, which presents a classic false dichotomy (i.e. Open is always good,
> everything else is part of an industry/microsoft conspiracy to repress the
> customers and that evil tactics will be required to put the valiant 'FOSS'
> down). The truth is much more shaded. <

This is probably the wrong forum to discuss morality; still, it tends to be true that Free software is good and proprietary software is evil. If nothing else, proprietary software represents a delegation of power which few would
accept if it were spelled out explicitly.

Jiri
--
Jiri Baum <[email protected]> http://www.csse.monash.edu.au/~jirib
MAT LinuxPLC project --- http://mat.sf.net --- Machine Automation Tools

 

Jiri,

If you read my post, you'll see that I don't discuss morality. Curt does (perpetually). And I'm glad that you agree that this is not the right forum for it.

He does, as usual, introduce a false dichotomy, which means that the discussion he is trying to start is fatally skewed from the outset. He does present 'FOSS' as a 'good thing' that the 'bad industry' has to conspire against to repress. False dichotomies are a classic technique used to
mislead - they are widely used, for example, in introducing pyramid scams (though I don't suggest OSS is a scam), or in politics to persuade people that there is 'no alternative'. So if we're going to have this debate, then
let's lay out honestly the two opposing positions rather than indulging in rhetorical tricks.

Which positions are (to my mind), the OSS community has so far not provided any viable alternative to commercial software in the automation field so the question of a 'shut out' really doesn't arise. Whereas on the other side the intense competition for customers with at least 4 major players (AB, Siemens, Mitsi, PLCDirect) is working for the benefit of customers to drive prices down. There is a myriad of smaller suppliers producing specialist kit
which use the more or less open interfaces of interfaces available from the big players (whether this be Profibus, DeviceNet, Modbus, 4-20mA, OPC or whatever), and no real clamour from people actually paying for this for anything different. In fact many big accounts ask for proprietary 'open' interfaces, such as Ethernet/IP or Profinet. And as I explained, there are good reasons why customers are broadly satisfied with kit which uses relatively old technology.

Best, Tim Linnell

 

On May 20, 2004, Curt Wuollet wrote:
> Biting sarcasm
> aside, and the point about leading taken, I'll classify this as a "we'll
> ignore it and see if it goes away" response <

There is no sarcasm in my post - read it again. Sometimes the truth hurts, and perhaps this is the source of the 'bite' you are feeling? Since 'it' hasn't shown any signs of actually arriving for the very many years you have been announcing its coming, it might be better to classify my post as a 'don't worry about imaginary creatures hiding in the shadows' response. ;-)

The key point here though is that you are setting out to mislead by presenting a false dichotomy in your question - this may be deliberate or just a function of idealistic overenthusiasm, I don't know, but misleading it most certainly is. Neither you nor Jiri has addressed the substance of the (detailed) alternative analysis I gave, but has rather indulged in ad hominem attacks around the margins.

I am a professional and highly skilled software engineer who expects to make a living from producing complex embedded products. I do the best to provide good products for my customers at a price that meets both of our objectives (making a living), and take some pride in visiting them, understanding what they do and what their motivations are, and *listening* to them. The competitors I really fear (and respect) do precisely the same. It really does irritate me then to hear that I'm part of some industry wide conspiracy to shut out 'open' software to do the customer down. It is simply not part of my working day to indulge in dirty tricks to destroy FOSS, nor of the
working day of my bosses who frankly are much more concerned about stealing a march on their traditional competition than worrying about something that doesn't actually exist yet.

Without rancour, you would do a lot better to turn yourself from an output device to an I/O device, i.e. to listen sometimes rather than just broadcast.

Best, Tim

 

Hi Tim


> On May 20, 2004, Curt Wuollet wrote:
>
>>Biting sarcasm
>>aside, and the point about leading taken, I'll classify this as a "we'll
>>ignore it and see if it goes away" response <
>
>
> There is no sarcasm in my post - read it again. Sometimes the truth hurts,
> and perhaps this is the source of the 'bite' you are feeling? Since 'it'
> hasn't shown any signs of actually arriving for the very many years you have
> been announcing its coming, it might be better to classify my post as a
> 'don't worry about imaginary creatures hiding in the shadows' response. ;-)

I'll let the reader be the judge on the sarcasm point. But if by "it"
you mean FOSS, it has arrived, and shown obvious advantage in several
roles. In fact Apache is pretty much the standard for HTTP servers.
And Linux itself has become important to some extremely large and well
thought of pillars of the computer industry. I think it is quite
reasonable to argue that penetration is directly related to the
"openness" of the market segment. With the web being, so far, beyond the
control of any cartel and the general computing market seeing a glimmer
of hope after a close call with monopolization. By the same argument,
one could expect penetration to be inversely proportional to the degree
of control exerted over a segment by a company or companies. Automation,
by it's nature and current practice is very tightly controlled and
vertical. That is, one company sells the hardware, the software, the
tools, the peripherals, the cables, the documentation, the support,
the right to connect and interoperate, etc., etc. The fact that there
are several such companies doesn't really change things much as they
all operate in pretty much that same way. Between them perhaps there
is competition, but from the outside it isn't any easier crossing the
mountain range than climbing an individual mountain. For our purposes
here, ridiculing our effort because we haven't been able to leap over
in a single bound, or fly to the summit is to make light of the task
as well as asking the reader to presume that it's possible but we
simply aren't making the effort. After all, a great deal of effort has
been expended in greasing the slopes and removing any footholds. But
we are working on inventing a helicopter, so anyone can get to the
top. Hard work yes, but far more practical and innovative than building
another mountain to get there. And it holds the potential to get you on
top of any mountain rather than settling for one.

>
> The key point here though is that you are setting out to mislead by
> presenting a false dichotomy in your question - this may be deliberate or
> just a function of idealistic overenthusiasm, I don't know, but misleading
> it most certainly is. Neither you nor Jiri has addressed the substance of
> the (detailed) alternative analysis I gave, but has rather indulged in ad
> hominem attacks around the margins.

Idealistic enthusiasm is a necessity I'm afraid, without it no one would
challange the status quo. And there is a touch of morality. But no
intent to mislead. There are some exceptionally sound and practical
reasons why I believe we will eventually get there, But doing so is
certainly more problematic than conquering web service. I liken it to
the bad old days of general computing. We need a universal platform
to make indenture to a particular company unnecessary. This is already
happening with only a few facets remaining to be cracked.

>
> I am a professional and highly skilled software engineer who expects to make
> a living from producing complex embedded products. I do the best to provide
> good products for my customers at a price that meets both of our objectives
> (making a living), and take some pride in visiting them, understanding what
> they do and what their motivations are, and *listening* to them. The
> competitors I really fear (and respect) do precisely the same. It really
> does irritate me then to hear that I'm part of some industry wide conspiracy
> to shut out 'open' software to do the customer down. It is simply not part
> of my working day to indulge in dirty tricks to destroy FOSS, nor of the
> working day of my bosses who frankly are much more concerned about stealing
> a march on their traditional competition than worrying about something that
> doesn't actually exist yet.
>
Well I get a bit irritated that trying to solve some of the problems
with that model in favor of the users and to make things easier for
integrators and generally much less of a PITA is percieved as something
other than that. And to the extent that that is in conflict with making
the maximum profit possible at all levels from the user, I am
unapologetic. My "business model" simply eliminates the higher levels.
Which means the lower levels, the provider and the customer should do
better with a much lower overall cost.

> Without rancour, you would do a lot better to turn yourself from an output
> device to an I/O device, i.e. to listen sometimes rather than just
> broadcast.

I am listening Tim, you want to make money the way things are. The
strange part is that I don't see at all how that would change if we
and FOSS were completely sucessful. I doubt that there would be a
free version of what you are doing. And you might even save a buck
or two a unit with an open platform and a FOSS toolchain. You are
a mountain climber, not a mountain.

Regards

cww

>
> Best, Tim

 

Hello,

I'm combining responses to about three of Tim Linnell's posts here.

I also get a bit philosophical in places, but it's all tied to the A-list and automation. I'm wearing my futurist hat, to use Curt's phrase.

Tim Linnell:
> Neither you nor Jiri has addressed the substance of
> the (detailed) alternative analysis I gave, but has rather indulged in ad
> hominem attacks around the margins. <

I've re-read your post, and it still seems to answer the question 'what will the future be like?' with 'it's not here yet'. It hardly seems to do justice to the question.

Curt does indeed ask a leading question in that he assumes that FOSS is the eventual destiny of the Automation world. Within that framework, though, it is still possible to ask important questions: Is the assumption likely? How quickly will it come? Will it be good or bad for us? If good, how can we help it? If bad, how can we hinder, or at least soften the impact?

My personal answers are: FOSS is coming; it will be a few years yet; it will be good overall; and I'm helping by participating in MAT.

> Automation technology does trail consumer electronics, probably not by 10
> years, but it does trail. There are good reasons for this, <

Agreed with this paragraph. The question is whether FOSS will be one of those technologies that will "shift them up a gear" or not. (I predict: yes.)

> Notwithstanding this, the industry is not dominated by a single company like
> Microsoft, but is viciously competitive. <

That's certainly better than a monopoly, but FOSS was never about defeating a monopoly; it's about supplanting proprietary software in general.

> If you read my post, you'll see that I don't discuss morality. <

That was more a note on myself than on you. Still, in dismissing terms like "good", "evil", "repress" and "conspiracy", you at least touch on the topic. More importantly, though, it's difficult to discuss FOSS without mentioning morality, because morality forms one of its foundations.

On the A-list, people help each other with no expectation of being paid for it, no restrictions placed on the use to which they put what they learn, and very few on what they download from the PLC Archive on control.com.

Like the Automation list, Free Software is "a true community, a place where peer discussions revolve around common (and uncommon) problems, where people help each other, where the important issues facing [the members] can be discussed and debated" (to quote The Control.com Manifesto).

The GPL is specifically and consciously desinged as a means to foster such community, discussion, and helping each other. (On the A-list, this is done by the moderators.)

The practical side, Open Source, is more commonly mentioned on this list: higher quality and lower cost, compared with proprietary software. This is as it should be; this list is for practical people with practical problems and solutions, so that quality and cost have a much more immediate impact than idle musings on the future of humanity. In posts explicitly tagged as futurist hat, though, it seems appropriate to shift the emphasis somewhat, and point out the original intent of the GPL.

> Which positions are (to my mind), the OSS community has so far not provided
> any viable alternative to commercial software in the automation field so
> the question of a 'shut out' really doesn't arise. <

The question has arisen already. A viable product must interoperate with at least a few major protocols and industrial busses. If the vendors make it difficult for a FOSS project to obtain permission to do so, FOSS is shut out.

Even modbus.org is apparently incapable of granting this permission, and the others are generally worse. This is also our complaint against OPC: it's impossible to obtain permission to implement it in FOSS - which is a pity, because it's otherwise "pretty neat" (to quote your words).

Jiri
--
Jiri Baum <[email protected]> http://www.csse.monash.edu.au/~jirib
MAT LinuxPLC project --- http://mat.sf.net --- Machine Automation Tools

 

Tim,

your response is quite interesting, but I think that it is based on a few serious misconceptions. Let me try to explain...

On May 21, 2004, Linnell, Tim wrote:
> If you read my post, you'll see that I don't discuss morality. Curt does
> (perpetually). And I'm glad that you agree that this is not the right forum
> for it. <

Morality should have IMHO role in any forum. Even if it is not discussed openly, morality - or lack of it - or (possibly seriously distorted) morality as a basis of law afftect all human activity, including automation. It will be sad day indeed when people will be ashamed to acknowledge that morality plays some role in their decision-making process.

> He does, as usual, introduce a false dichotomy... <

What is your involvement with FOSS, Tim? In my experience there is marked difference in perception of FOSS between people who use it and people who don't. People who have no or limited experience with FOSS software tend to treat it as just another brand of software - probably a bit amateurish - which differs from "mainstream" mainly by the strange fact, that owner of copyright forgot to send invoice. People who use FOSS OTOH understand that absence on invoice (or presence of invoice) is secondary to the main issue: differece in basic design & development philosophy of FOSS software. This difference is difficult to describe but it is real - it allowed international ad hoc organised motley crew of Linux programmers to challenge best and most complex proprietary software system known to humankind. I am not claiming that one or the other is momentarily better - I am just saying that there is a dichotomy real enough between both systems.

> Which positions are (to my mind), the OSS community has so far not provided
> any viable alternative to commercial software in the automation field so
> the question of a 'shut out' really doesn't arise... <

AFAIK it is almost vice versa: absence of "viable alternative to commercial software" can be partly or fully atributed to 'shut out' you are attempting to deny. If there is absence at all - how much time did you spent
familiarising yourself with later versions of MATplc, BTW?

> ... and no real clamour from people actually
> paying for this for anything different.
...
> explained, there are good reasons why customers are broadly satisfied with
> kit which uses relatively old technology. <

Yes, of course. I am sure that the same statement would be applicable to almost any obsolete - but working - technology. No points for this one...

Cheers

Petr
--
<[email protected]>

Petr Baum, P.O.Box 2364, Rowville 3178
fax +61-3-97643342

 

I will leave the philosophical discussions to others, and just discuss a few
practical points with Mr. Baum.

A question I've asked before about your MAT project is "who is your target
market?". Just because you aren't planning on charging any money for it
doesn't mean you don't need a market. What problem is it you are solving for
people? If so far your plan is to target "anyone who needs a PLC", then I
don't think you'll generate a lot of interest. Soft logic systems (which MAT
is) in general have not had a lot of success in the market so far. I believe
there are a number of reasons for this, of which cost is only one.


I in turn predict that I/O will have to be more of a commodity item before
something like MAT will have an opportunity to be a threat to the
conventional PLC in the general market. I think I/O will become a commodity,
but not for a few years yet.


My own opinion (for what that may be worth), is that if at this time you are
going after the general PLC market, then you are targetting the wrong market.
Most people buy conventional PLCs from the market leaders because they are
easy, simple, reliable, and maintainable. There are good distribution
networks for spare parts, lots of training courses, and plenty of nice thick
manuals. Maintenance departments can stock a particular part number and know
that it will fit into a machine they have. Regardless of what you may think
of prices and vendor domination in the automation market, it is not even
close to the problems in the computer software market which are driving
customers there to seek alternatives.


The sort of market that I would forsee at this time as being interested in a
MAT-like system would have the following characteristics.

A) It would be part of a larger system, rather than just a PLC, and it would
be a system where the type of PLC is NOT customer specified. This criteria
avoids current customer specs for particular brands.

B) It would be integrated by a company with more than average engineering
ability. This avoids current limitations on documentation and lack of
training courses. Essentially, the people in a company like this would be
able to figure things out for themselves without a lot of hand-holding, and
could also offer their own contributions for improvement.

C) The system would already include a PC, which is closely integrated into the
system. Since a PC is already present anyway, the designers would be avoiding
adding a PLC and interfacing it to the PC. This is an easier decision to make
than choosing a PC over a PLC.

D) They would be selling dozens of broadly similar (but not necessarily
identical) systems. They would be selling enough systems that spending the
time to learn something new would have a return in terms of cost and in
making a better product. It would also be worth while for them to add some
missing features that they want.


The markets that I see meeting the above are computerised test equipment, and
certain types of standard OEM machines. A number of companies in these
markets have already written their own proprietary soft logic systems, so
they do have a need for them.

The computerised test equipment builders probably have the simpler needs, and
so would be a better initial target. I don't know of any third party
proprietary soft logic systems which address this market very well, so you
wouldn't be trying to displace any well established alternatives. Starting in
a niche market doesn't mean being limited to a niche forever. Once you have a
base system which is in use, you can add features to it to address particular
needs which are not currently being met.


P.S. If a genuinely "open" soft logic system does enter into the general PLC
market, I think it is quite likely to be as part of a proprietary product
line. You would buy it as part of a product the same way you would a
conventional PLC. The company selling it could charge money for it because
they would be selling it as a single complete package already loaded onto the
CPU hardware, together with support contracts, manuals, training courses, and
distribution.

The advantage to the customer would be the ability to change suppliers easily
if they got fed up with their current one. We wouldn't have the IEC 61131-3
fiasco because it wouldn't be a "compatible" logic solver, it would be the
*same* logic solver. Your PLC programs would be completely portable because
the same run-time software would be used.

Given the above, it is not entirely impossible that an "open" soft logic
system could originate from one that is currently proprietary. A company
which makes its money selling I/O (or a second or third tier PLC vendor)
could "open" an existing product (either one they already have, or one they
buy for the purpose) and use it as a tool to undermine the market dominance
of the first tier vendors. If they are initiating the project, they can make
sure that it has full support for their I/O (and other devices) product line.
If they can change the market to separate the I/O from the CPU sales, they
could have a lot more to gain from I/O sales than to lose from CPU sales by
this strategy.



--

************************
Michael Griffin
London, Ont. Canada
************************