# Software Quality

C

#### Curt Wuollet

Hi All

Living with bugs, missing features, unsupported versions? Have you been coerced to upgrade before you even got that last version working? I saw this article and it sounded so familiar I thought I would share. What's coming up and some remedies are discussed. Relevent since almost everyone in this market uses proprietary software.

http://www.cio.com/archive/101501/wasting.html

Regards

cww

--
Free Tools!
Machine Automation Tools (LinuxPLC) Free, Truly Open & Publicly Owned
Industrial Automation Software For Linux. mat.sourceforge.net.
Day Job: Heartland Engineering, Automation & ATE for Automotive
Rebuilders.
Consultancy: Wide Open Technologies: Moving Business & Automation to
Linux.

R

#### Rob Hulsebos Philips CFT

Yeah, and I wonder if it is getting worse with the new policy of Microsoft - a new version or update of Windows, will it also require updates of the other software installed? Which in turn may force other software to be updated? Does a whole avalanche of updates have to be installed
at once? Questions, questions...

Rob Hulsebos

C

It'll probably depend on a quota. Once you hit $500.00 per license per year for your share of bugs, they call off the enforcers. After all, a protection racket doesn't work if it's more painful to pay than not. I'm sure thay don't care if you install them or not as long as you pay for them. Regards cww A #### Anthony Kerstens Sounds just like Rockwell, making you pay thousands of dollars for each of their software packages with little hope of inter-compatibility. It's alway peeved me to have to pay 4k for RSLogix5 and then pay another 4k for RSLogix500. It peeved me even more to pay for RSLogix5000, and then see all the features of the other two not present in 5000. You'd think since hundreds of thousands of dollars is being sunk into hardware that at least they'd offer the software for free, or at least <1k. And make one software package that would do it all. Is anybody at Rockwell listening???? Hello!!!! Anthony Kerstens P.Eng. (Oh, how I wish my customers would smarten-up and start using Modicon. J #### Jake Brodsky So you must have seen this blurb on Slashdot too... As I see it, the real problem here is social. Software is not a physical thing. Selling, licensing, copyrighting, and/or patenting are all ideas which don't quite fit the situation. The major advantage of the GPL, as I see it, is that it gets around all these things by turning the situation on its head. It opens up the code and the development process to the whole world instead of restricting it to a small cabal of monkish programmers who have no time to experience the worlds where the software may propagate. However the GPL is still a rough and raw concept. It has yet to gain much acceptance among business folk who don't understand what they're really buying when they choose a software platform. Ergo, the article you cited. Until the world agrees that we have a new branch of intellectual property that simply doesn't fit any of the old models, and until we all agree what this model should be, we are going to have some very poor software because our motives for writing good software simply don't push us in a productive direction. A #### Alex Pavloff Interesting article -- it really does point out how the big IT vendors make money and the frustration that it engenders in their customers. Some points: In the case of IT companies mentioned in the article, the "subscription" model does appear to make sense. However, I don't think that this model applies that well for those of us in this industry. With a few exception (Intellution et al), nearly everything in the automation industry is a piece of physical something that you can hold in your hand (and throw back at the vendor if it becomes too annoying). It's hard to charge a subscription for something that I've already got. Now, as another poster here complained, Rockwell is charging yearly for their RS* software, in addition to charging a premium for their hardware. Why can they get away with it? Easy -- they're a big company selling to big companies. Eason Technology (my company), is small (not completely by choice, of course <g>). As such, we're selling to small-medium companies. Our software is now selling for$99 -- a tiny fraction of the cost that we've invested in it, and we have no yearly fees and have never charged any money for an upgrade. Our software, is, of course, completely useless without the hardware we sell. Going to a subscription model makes absolutely no sense -- and the customers at our end of the market wouldn't go for it.

Now, in regards to open source. To make open source work, you need to have smart people, and you need them on staff. Its great if you have the money for that. The reason that open source hasn't taken off in the automation market is because, well, software is one of many elements to a project. Generating enough work for someone good enough to pay for themselves would be difficult. Curt, you're one of the main people behind the LinuxPLC project, but you still do "other stuff" to pay the bills, right? Large
companies are those that generally "sponsor" open source projects (Hi, IBM!). As commented in the "Integrator" thread, large manufacturer (Allen
Bradley et al) are trying to get into the integration game in order to make some money. What incentive do they have to open source any of their stuff (or sponsor, say, the LinuxPLC) -- they just want to sell their own stuff. IBM, in contrast, is sponsoring many open source projects because then they can turn around and sell their services using this software.

Until something changes, open source development in this industry will be mainly an after-work project. My company will be using Linux & open source software on our next generation of HMI panels, but our product is a lot closer to a computer, than say, some sort of motor. With a few exceptions, hardware doesn't lend itself to the open source paradigm.

In general, I don't think that we can take arbitrary lessons from the IT field and apply them to our field.

Alex Pavloff
Software Engineer
Eason Technology

C

#### Curt Wuollet

Hi Alex

List Manager wrote:
>
> ---------- Forwarded message ----------
> From: Alex Pavloff <[email protected]>
>
> Interesting article -- it really does point out how the big IT vendors make
> money and the frustration that it engenders in their customers.
>
> Some points:
> In the case of IT companies mentioned in the article, the "subscription"
> model does appear to make sense.
>
> However, I don't think that this model applies that well for those of us in
> this industry. With a few exception (Intellution et al), nearly everything
> in the automation industry is a piece of physical something that you can
> hold in your hand (and throw back at the vendor if it becomes too annoying).
> It's hard to charge a subscription for something that I've already got.
>
> Now, as another poster here complained, Rockwell is charging yearly for
> their RS* software, in addition to charging a premium for their hardware.
> Why can they get away with it? Easy -- they're a big company selling to big
> companies.
>
> Eason Technology (my company), is small (not completely by choice, of course
> <g>). As such, we're selling to small-medium companies. Our software is
> now selling for $99 -- a tiny fraction of the cost that we've invested in > it, and we have no yearly fees and have never charged any money for an > upgrade. Our software, is, of course, completely useless without the > hardware we sell. Going to a subscription model makes absolutely no sense > -- and the customers at our end of the market wouldn't go for it. > > Now, in regards to open source. To make open source work, you need to have > smart people, and you need them on staff. Yes and no. Yes, Open Source would be invaluable to integrators with the expertise to add or modify a product to get those functions that you need, but no one has thought of yet. And contributing such work back to the public pool would build very rich functionality and broad application quickly. But no, simply being Open Source doesn't imply any extra skills or talent. If the products you use today suddenly released their source, would that make them any harder to use? I don't think so. Implicit in the authors argument is that the Open Source be self supported. Community support would make up the difference. This list is an example of how well that can work. Especially since many people on the list get the impression that their current tools are more or less self supported ;^). But a lot of Open Source is used exactly the same as closed source, by installing the binaries and leaving the source on the CD as insurance. I use many OSS products every day that I have never read the source for. It's nice that I could fix them if I want to and I can never be left hanging or forced to upgrade and can maintain them indefinately, but I am not interested in improving those outside my interests. Its great if you have the money > for that. The reason that open source hasn't taken off in the automation > market is because, well, software is one of many elements to a project. > Generating enough work for someone good enough to pay for themselves would > be difficult. Curt, you're one of the main people behind the LinuxPLC > project, but you still do "other stuff" to pay the bills, right? Yes, I do captive automation with OSS for my employer. For the type of work we do, integrating existing machines and equipment, proprietary off the shelf technology is not cost competitive because by design, no two machines or closed systems can interoperate. It is possible to do it OTS in some cases but you end up buying many adapters, bridges and general kludges which eat up much more time and money than custom programming with generic hardware. Some new work is done with PLC's etc. some is done with Linux. PLC's are very good for a few things. For most other things Linux requires a lot less hardware and time. This is a tremendous competitive advantage that we have almost exclusively to ourselves because most folks are still trying to make PLC's do everything. Large > companies are those that generally "sponsor" open source projects (Hi, > IBM!). As commented in the "Integrator" thread, large manufacturer (Allen > Bradley et al) are trying to get into the integration game in order to make > some money. What incentive do they have to open source any of their stuff > (or sponsor, say, the LinuxPLC) -- they just want to sell their own stuff. > IBM, in contrast, is sponsoring many open source projects because then they > can turn around and sell their services using this software. For the great majority of the readers of this list, the money _is_ in services. You typically don't get rich selling AB or GEF hardware or software. What you get paid for is expertise and solutions. That's why we aren't looking for AB or GE or even IBM as sponsers. We want you to join us in building stuff that you can use to build better solutions for higher profitability. The big guys can take care of themselves. > Until something changes, open source development in this industry will be > mainly an after-work project. My company will be using Linux & open source > software on our next generation of HMI panels, but our product is a lot > closer to a computer, than say, some sort of motor. With a few exceptions, > hardware doesn't lend itself to the open source paradigm. > > In general, I don't think that we can take arbitrary lessons from the IT > field and apply them to our field. No, not generally but OSS and generic hardware fit perfectly in the problem areas that the Big guys not only don't address but have in fact created. Like communications, integration, networking, computation and any area that requires flexibility and interoperation. That's plenty of work to keep me busy and we're a small shop. It's the most useful tool in my toolbox. Doing what a PLC does isn't that much to add. Regards cww -- Free Tools! Machine Automation Tools (LinuxPLC) Free, Truly Open & Publicly Owned Industrial Automation Software For Linux. mat.sourceforge.net. Day Job: Heartland Engineering, Automation & ATE for Automotive Rebuilders. Consultancy: Wide Open Technologies: Moving Business & Automation to Linux. M #### Michael Griffin At 16:35 18/10/01 -0400, Alex Pavloff wrote: <clip> >Now, as another poster here complained, Rockwell is charging yearly for >their RS* software, in addition to charging a premium for their hardware. >Why can they get away with it? Easy -- they're a big company selling to big >companies. They can charge a high price because they are perceived as having dominance in a particular sector of the market. People are willing to pay extra for AB hardware partly because they believe it is easy to find lots of companies (integrators, etc.) who will have the software and so can support them. Integrators are willing to pay high prices for the software because they believe there are lots of AB PLCs being installed in machines, and so they can pick up lots of work programming them. However, this whole positive cycle can go disastrously (for AB) into reverse if people start to perceive any weakness in the company's market position. Integrators will become reluctant to continue investing in the software if they think there will be less AB hardware around to make money off of (and so pay for the software) in future. Customers will become reluctant to pay a premium for AB hardware if they think it is going to become less easy to find integrators who have the software to work with it. The two trends then feed off each other to continue the decline. Once this sort of downward spiral begins, the only thing Rockwell can do is cut prices on everything, and I don't think they have the cost structure to let them do that. >Eason Technology (my company), is small (not completely by choice, of course ><g>). As such, we're selling to small-medium companies. Our software is >now selling for$99 -- a tiny fraction of the cost that we've invested in
>it, and we have no yearly fees and have never charged any money for an
>upgrade. Our software, is, of course, completely useless without the
>hardware we sell. Going to a subscription model makes absolutely no sense
>-- and the customers at our end of the market wouldn't go for it.

For a small company, inexpensive software makes a lot of business sense because by lowering the software investment cost, it lowers the risk

M

#### Michael Griffin

The article was interesting, but I think their idea that the subscription model presented there will put more control into the hands of
customers is just wishful thinking. A single customer really only has any leverage if they represent a significant piece of revenue for a software company and they have the ability to hold back payment until they are satisfied.

I don't think that the large software companies would be interested in the subscription business model unless they felt it would bring in more revenue at a lower cost to themselves. I could see customers (especially the smaller ones) being coerced into paying ever higher monthly fees for an irregular stream of updates of questionable value or utility.
The mainframe computer era operated almost entirely on the "subscription" model being proposed (it used to be called leasing), and I
don't recall hearing that very many software customers were satisfied with the results of that.

**********************
Michael Griffin
**********************

D

#### David McGilvray

Thursday, October 18, 2001 10:21 PM Michael Griffin wrote:
<big clip>

I don't see that open source necessarily means "free". You might
still have to pay someone (i.e. a consultant) some money to do something for
you. I could theoretically build my own "open controller" using bits and
pieces of hardware that I bought from various places, and software that I
LinuxPLC). However, why would I do this, other than for the educational
value? Whatever it is I may be good at, it isn't piecing together these
kinds of systems and then testing and documenting them.
It would make much more sense to buy a standard tested system
(hardware plus software) from someone who is spreading their engineering
cost over a much larger number of systems than I would ever use in a
lifetime. I would rather simply pull the complete controller out of the box
and install it and load in my PLC program.
Since I could buy equivalent systems from several different
companies I would still have the advantage of more choice as to whom I deal
with. I would also enjoy a lower risk since different suppliers would be
offering essentially similar and compatable systems. I could switch between
them if I grew dissatisfied with one, or if one of them dropped out of the
A supplier would in turn have the advantage that this lower risk
would make customers more willing to do business with them to begin with.
This would be a quicker and easier way for a new entrant to get into the PLC
business than creating their own system entirely from scratch.
<clip>

Michael,
Although I can't speak to the LinuxPLC product specifically, an open source system need not necessarily be different from a proprietary alternative. I expect the LinuxPLC, when finished will be an open source option similar to
the proprietary automationX. The scenario you describe of "building" your own controller with "bits and pieces of hardware" and "down loaded software" grossly misrepresents the situation for server based systems. First, the
fact that you have choice is generally considered a good thing. Second, there is no more "building" involved and conceivably substantially less "building" (certainly the case with automationX - try up to 80%) compared
to conventional PLC's - it's generally called configuration. Third, server based systems typically have far less hardware - resulting in much less costly installation & maintenance.

Look to the IT industry for inspiration. Once upon a time, IBM dominated the industry selling complete system, including hardware, software, etc all from big blue. Today, the landscape is substantially different with a huge amount of choice covering virtually every conceivable features and function set. And best of all, the final solutions are much better and much less
costly.

David McGilvray
M&R Automation

M

#### Mark Wells

Michael,

Interestingly, for OEM's like Eason, the licensing costs of including some open source products may actually be more expensive than old-fashioned proprietary products. I was going to test MySQL (open source database product) to replace a proprietary SQL database product that is used in one of my company's products. I just needed a single workstation version that supported SQL. The licensing terms for NON-END-USERS make licensing MySQL more expensive than my original proprietary database product, especially for low product volumes.

In the case of MySQL, if the end-user configures and installs the database (or hires a consultant to do it), then there is no licensing fee for my
company's product if I want to include MySQL connectivity. Unfortunately, the industry that the product is targeted at typically has factories with fewer than 50 people, which means that they usually have very little IT support, and very little money in the consulting budget.

Sincerely,

Mark Wells
President
Runfactory Systems Inc.
http://www.runfactory.com1235 Bay Street, Suite 400
Ph. 416-934-5038
Fax 416-352-5206

M

#### Michael Batchelor

This licensing cost is only one of several problems I've had with MySQL, so I've started doing some work with Postgres. The open source license is really open source. There is a Windows build, but I haven't started experimenting with it yet. And there is an ODBC driver freely available, too. My guess would be that it's not going to be easy to set up the way Wonderware just installs MS SQL Server, but I think it will be workable.

E

#### Ed Mulligan

In this case, aren't you serving as the consultant to install the software for the end user, so you are covered? Have them write their P.O. so that they are buying the system _and_ the consulting service. For your development box, aren't you the end user and also covered?

I won't let never having read MySQL's agreement stop me from answering. 8^)

Ed

Speaking for me, not for Starbucks. . .

M

#### Michael Griffin

We have a piece of custom test equipment we wanted duplicated (with minor changes) in a hurry by the original integrator. It is a custom Visual Basic application running on a PC with a Windows NT operating system.

To make a long story short, the integrator spent his entire on-site time allocation (3 days) trying to solve Windows NT computer hardware dependencies (problems with talking to various boards), and had to switch to
Windows 98 just to get things working. The actual integration work (actually solving the problem he was getting paid for) ended up being done on his own time (with the customer breathing down his neck).
He of course still has to spend additional time figuring out the Windows NT problem to deliver on the terms of the contract. He will likely solve this by replacing the computer. All of this extra work (and hardware)
is of course at his own expense. This isn't unusual when dealing with PC systems though.

This example brings up an interesting question. It has been asked in the course of this discussion as to where someone would turn to for support with open source software.
The fellow mentioned above was using Microsoft Visual Basic with Microsoft Windows NT, which is of course "closed" and proprietary. Where
could he turn to for support with his software problems? No where as far as I could see. He had to solve everything himself. Who *would* he have talked to? The computer worked. The boards worked. Windows NT worked. Visual Basic worked. They just didn't work together.

I guess the only solution to this sort of problem is "to have smart people on staff". Have you got any better ideas? Should we have called
Microsoft to complain about their software? Do you really think that would have helped?

I don't know if we would have been any better off using open source software in this sort of situation. I find it hard to imagine though how things could have been any worse.

**********************
Michael Griffin
**********************

M

#### Michael Griffin

David McGilvray wrote:
<clip>
>The scenario you describe of "building" your
>own controller with "bits and pieces of hardware" and "down loaded software"
>grossly misrepresents the situation for server based systems.
<clip>

I don't think I made any misrepresentations about server based systems in general or your product in particular, since I didn't mention them. I was referring to a normal PLC (especially small ones) used in typical small machine applications.

>Look to the IT industry for inspiration. Once upon a time, IBM dominated
>the industry selling complete system, including hardware, software, etc all
>from big blue. Today, the landscape is substantially different with a huge
>amount of choice covering virtually every conceivable features and function
>set. And best of all, the final solutions are much better and much less
>costly.
<clip>

Alright - I suggest we do look at the IT industry. How many people (non-hobbyists) build their own computers out of parts they found in surplus stores and software they scrounged up on the internet? I'm sure you could build a perfectly good system that way (if you knew what you were doing), but is it economical? Most people would rather just buy a computer.
If I am going to buy an industrial controller, I don't want to have to figure out which board works with what driver. I am willing (because of cost) to pay someone else to do that.
This is particularly important with low cost systems where there is very little margin to spend on unexpected problems. This may be less
significant on larger systems.

**********************
Michael Griffin
**********************

M

#### Mark Blunier

We have to pay for the software over and over again through support contracts, because we expect to make changes to the system, and know that if we do, we are likely to run into problems, and our only hope to get help from Rockwell is to have a support contract. We also figure that they will gradually eliminate a few of the bugs in the newer versions of the program. This is also some of the reasons some of us prefer Modicon. They don't charge us for support beyond the cost of the equipment, and the
software (at least the old stuff that I'm using) doesn't seem to be full of bugs.

Mark Blunier
Any opinions expressed in this message are not necessarily those of the
company.

> >Now, as another poster here complained, Rockwell is charging yearly for
> >their RS* software, in addition to charging a premium for their hardware.
> >Why can they get away with it? Easy -- they're a big company selling to
big
> >companies.
>
> They can charge a high price because they are perceived as having
> dominance in a particular sector of the market. People are willing to pay
> extra for AB hardware partly because they believe it is easy to find lots
of
> companies (integrators, etc.) who will have the software and so can
support
> them. Integrators are willing to pay high prices for the software because
> they believe there are lots of AB PLCs being installed in machines, and so
> they can pick up lots of work programming them.

J

#### Jeffrey D. Brandt

Well, I always said that public domain software is worth what you pay for it. (nothing, get it?). By that, the AB/R$should be worth its weigh in gold. And, would'nt you expect at least something that works with ALL AB/R$'s stuff??
The previous comment about "so they can pick up lots of work programming " is also out the window these days, because integrators are competing with
Allen Bradley for work. Check out AB's web site. GTS used to be a VERY thinly veiled integration arm of AB.....NOW they (AB) tout their
'application programming assistance' as a chargable service. Geez, as if business wasn't hard enough.

I'm with Mark. Software isn't the great dividing point that it once was. 'It ain't no ICOM' (for you kids: 'ICOM' was the name of the very good
company that originally produced the 'best ladder programming and documenting software ever known to man', later to be gobbled up by AB and called 'A.I.', and later to be canned in favor of R$Logix) used to be my mantra for pronouncing relative quality of programming packages like Omron's LSS and almost anything windows based. These days, there are some very good competitors out there, S7 (hated it, loved it), VersaMax (Starting to not hate it so much), and the even better 'FREE' tools. I remain as mystified as the rest on this list ) except for the obvious AB cheerleaders). We all know that we already pay more for AB because of their 'great support' (another topic), so....WHY do we have to pay a premium for almost everything else from this outfit? I guess, we do be cause we will, and they do, because they can. Remember, you can always buy better, but you'll never pay more. And, BTW, I've already shared this with AB's management........... Jeff R #### Robert McDonald I can only agree with the lock-in R$ seems to have over its software.

Compounded by the fact that the newer hardware can only be supported by the newer software, so you MUST upgrade. Take a micrologix 1200 for example, we had a dozen of so used in a neat OEM project. Tried to build another recently, now its series C, sorry, you need to upgrade R\$500 to program it, please deposit money here.

Robert McDonald

J

#### Jiri Baum

David McGilvray:
> >The scenario you describe of "building" your own controller with "bits
> >and pieces of hardware" and "down loaded software" grossly misrepresents
> >the situation for server based systems.
...
> >Look to the IT industry for inspiration.
...

Michael Griffin:
> Alright - I suggest we do look at the IT industry. How many people
> (non-hobbyists) build their own computers out of parts they found in
> surplus stores and software they scrounged up on the internet?

Exactly. The same thing would apply to building your own open-source controller - few would do it, and then only if they have special needs.

> If I am going to buy an industrial controller, I don't want to have to
> figure out which board works with what driver. I am willing (because of
> cost) to pay someone else to do that.

Yup. So you pay your local VAR, or a national chain if you prefer, to put an industrial controller together for you from standard parts. Most likely you'll get one of the standard configurations off the shelf, maybe with minor variations.

Because anyone has access to the parts and the software, though, you'll have several sources[1] - which is good for price, but more importantly
good for risk management. If your VAR goes out of business, you can get substantially the same product from next door.

Proprietary software, by its nature, is single-sourced. You should always think twice before using a single-sourced component.

Jiri

[1] in the worst possible scenario, the alternative source will be doing it yourself, which would be expensive, but should be rare and in any case will be better than nothing.
--
Jiri Baum <[email protected]> http://www.csse.monash.edu.au/~jirib
MAT LinuxPLC project --- http://mat.sf.net --- Machine Automation Tools