Microsoft .Net's impact to Automation Industry

M

Thread Starter

Mark Meng

Dear list members,

As Microsoft's new software development tool .net will be launched soon, has anybody in the industry evaluated the impact it could put on
automation, especially the area of HMI:

1) The entire .net technology is claimed to be web-based, how will it help to develop or improve web-based HMI? Is there any HMI package supplier preparing to take the advantage of it?

2) The new language C# is obviously aimed at Java, and it has high chance to win. I remember the Java topic has been hot in this forum for these two years, but now is everybody preparing to shift to Microsoft or stay with Sun?

Best Regards,
Mark Meng
 
R

Ranjan Acharya

I suppose we are about to have lots of hot air from the automation solution vendors. In the mean time we still have to deal with whether or not existing tools will work with Win 2000, Win 2000 SP1, Win 2000 SP2, Win XP, Win NT SP5, Win NT SP6a, DOS boxes under any flavour of Windows and GNU/Linux with its various X-window interfaces

Perhaps the IEC might even sit down and write a standard to cover .NET

For the most part, I am still reading about the PC-based automation controller revolution, the XML revolution, the unified single field bus
revolution, and the IEC 61131.3 revolution. In the mean time I keep working with incompatible non-interoperable tools.
 
W
Here's my analysis, Mark.

First, the HMI technologies are not as important as what applications they are used on. This is emphatically true when you consider all types of control systems. We often get interested in the "howness of what" instead of the bottom line. You can see this in the fact that all the automation companies are moving to being integrators, and all the good integrators are moving up and out from their bases on the factory floor. HMIs and digital control systems have been available now for a decade...a _very long time_ in the business world of today. Managers and C-level company leaders take the capabilities of these systems for granted, often because they have never directly worked with them, and they assume that they work fine on the factory floor. They are more interested in how well they can integrate to the rest of the enterprise...and what the _benefits_ are in doing that.

Enterprise integration is _not_ cheap. By courtesy of Interwave Technology, Shari Worthington and I adapted a paper Charlie Gifford wrote as a sidebar in our new book _eBusiness in Manufacturing_. This sidebar details all of the planning and strategic work that has to be done, before the HMI is even bought, let alone installed, field tested and blest. When you read that, you get a clear understanding of just how costly even supply chain integration is.

Faced with those costs, and with hugely unsatisfactory records in most integration projects from CRM to ERP (CRM is worst, with an average failure rate of 65% of all projects) managers are reluctant to buy _any_ technology,
regardless of who developed it, unless the payback is clear and immediate.

Second, there is Microsoft itself. As a stockholder, I personally am a foaming madman, because I see them actually _forcing_ the government to break them up, and I don't think most parts of a broken up Microsoft will stand. I expect to lose a significant chunk of my retirement income, unless I sell soon. I can only hope that Gates and Ballmer have figured out that
we'll all come out ahead if the company is broken up...

Microsoft has simply terrified most IT professionals with the new "reporting" and "authorizing" features of Windows XP and .Net. There is no way on earth I will install an operating system that gives the vendor access to individual documents on my hard drives _on purpose_. Suppose, as was the case a few years ago, I was bidding a project to Microsoft, itself? It would be very easy for them to just lift my calcs, read them and know exactly what I was bidding and what margin I'd be making.

No corporate executive could possibly allow him or herself to be in that position.

So, it doesn't matter how cool .NET is. For the immediate future, Microstupid has shot themselves in the foot.

All that will happen in the near term is that people will continue to use the tools they've been using, because THEY WORK WELL ENOUGH to do factory floor projects, and people will continue to write ODBC and driver library hooks for the integration to SCM and ERP systems. Lots of people are using Microsoft's most stable platform, Windows NT4/SP6a, instead of upgrading.
Why should they?

Part of the reason for this, too, is that development costs money. In a time when all of the major automation companies are reporting reduced revenues, and all but two of the majors are in deep financial trouble, _and_ at least three of them are in danger of nonsurvival, there is nobody who is spending any significant amounts on R&D.

Add to this the intense development work going on in the open source community, and you can only begin to get a picture of a seriously fragmented situation becoming even more fluid.

We are likely to be back in the early '70s soon: incompatible OSs, fragmentary and fragmented versions with limited to poor version control, difficulty porting programs, difficulty interfacing programs. Part of the issue is that while Microsoft has committed suicide, most of their competitors are too stupid to do anything about it. And while the open source movement is making great strides, Linux is simply not ready for prime time, _yet_. It will be, but it simply is three or four years away.


Walt Boyes
co-author of "e-Business in Manufacturing: Putting the Internet
to Work in the Industrial Enterprise" ISA Press--September 2001 ISBN:
1-55617-758-5
____________________________________________
---------------------------------------------
Walt Boyes -- MarketingPractice Consultants
[email protected]
21118 SE 278th Place - Maple Valley, WA 98038
253-709-5046 cell 425-432-8262 home office
fax:801-749-7142 ICQ: 59435534

"Strategic marketing, sales and electronic
business consulting for the small and medium-sized
enterprise: http://www.waltboyes.com"
---------------------------------------------
 
A

Alex Pavloff

> > As Microsoft's new software development tool .net will be launched soon,
> > has anybody in the industry evaluated the impact it could put on
> > automation, especially the area of HMI:

.NET is a marketing buzzword. Keep that in mind when you see anything from Microsoft. They'll brand everything as .NET if they think it'll help. When I'm talking about .NET, I'm talking about their technological doodads they've created. Their technical plans are, quite frankly, pretty cool. Their plans to sell services and become the Info Portal Of The World is doomed to failure, because I don't think that people trust Microsoft with their personal info after all the Hotmail security holes and IIS security holes.

They've got a single runtime (the CLR or Common Language Runtime), which has the capability (well, what Microsoft says it has) to allow for easy communication between C++/Fortran/Cobol, etc etc. That's pretty neat actually.

C# the language is aimed right at Java. It's a COM based (or whatever they call COM nowadays) language with that looks like C++, but more closely tied to a Microsoft platform. It's easier to make Windows applications with it, but don't count on it being useful for anything else (which I why it won't replace C++).

They've also created a whole pile of library functions that you can call from any language. This is all well and good. What it really means though is that you're hitching your cart to the Microsoft cart, and good luck trying to get off. For some people, this is acceptable. For others (I'm just waiting for Curt to jump in here <g>), it won't be.

> > 1) The entire .net technology is claimed to be web-based, how will it help
> > to develop or improve web-based HMI? Is there any HMI package supplier
> > preparing to take the advantage of it?

It's web-based, but tied to Microsoft. This is great if all you want is PCs running Windows. I don't like PCs running Windows on the factory floor. I like thin (comparatively) clients running Linux or some other OS with a standard web browser. I suppose Windows CE could go there, but all indications and rumors that I've been hearing paint all the existing Windows CE devices as pig-slow.

> > 2) The new language C# is obviously aimed at Java, and it has high chance
> > to win. I remember the Java topic has been hot in this forum for these two
> > years, but now is everybody preparing to shift to Microsoft or stay with
> > Sun?

How many people are using Java in Automation? I dunno. What I see happening is Intellution and all the other Big Software That Runs On PCs using C# as a step up from VB.

But that's my view. I am a C++ programmer. I am not planning to use C# or .NET in any way, shape or form.

Alex Pavloff
Software Engineer
Eason Technology
 
C

Curt Wuollet

I can't imagine a more idiotic scenario than to keep your process control and automation data on someone else's server. Especially a morally
bankrupt partner like Microsoft. Why not simply publish it to your competitors?
That aside, it will be ballyhooed and eagerly purchased simply because MS says so. It's where you want to go today.

Regards

cww
 
T

Tom Tuddenham

I don't think .NET is a particularly scary ogre. There are some great ideas coming out of .NET that J2EE (or other solutions) has to catch up on. Even if much of .NET is vapourware, is still sets a benchmark we can aspire to. <br><br>For my part, I'm looking forward to the XML-OPC spec when it's finally delivered to the general public. In my job I have to think about how to integrate plantfloor and business systems and anything that promotes the XML "glue" technology is a good thing from where I stand.

Cheers
Tom
 
V
There are those who are technically locked-in to Microsoft and there those who innovate.
Vic Ellescas
Sverdrup Technology
Windows are closed...Linux opens the future
 
W
No, Curt, it will be bought because none of Microsoft's competitors, including the open source movement can market their way out of a torn wet paper sack.

There is a huge market for turn-key, plug-and-play non-Microsoft computer operating systems. There is a huge market for an office suite that doesn't report to Mother Redmond every ten minutes. Watching Windows and Office try to get through my firewall is hysterical.

But most people aren't programmers. Even many systems integrators don't like doing custom programming. It raises the bar to attract new customers. The more "function blocks" that are already written, the cheaper each job becomes.

There is an analogy in the back office there too.

What is needed is somebody to package an opensource OS/office suite and sell it. And all the Linux purists need to get on board with it instead of sitting back and taking shots at Microstupid and everybody else who has a clue.

Walt Boyes
co-author of "e-Business in Manufacturing: Putting the Internet
to Work in the Industrial Enterprise" ISA Press--September 2001 ISBN:
1-55617-758-5
____________________________________________
---------------------------------------------
Walt Boyes -- MarketingPractice Consultants
[email protected]
21118 SE 278th Place - Maple Valley, WA 98038
253-709-5046 cell 425-432-8262 home office
fax:801-749-7142 ICQ: 59435534

"Strategic marketing, sales and electronic
business consulting for the small and medium-sized
enterprise: http://www.waltboyes.com"
---------------------------------------------
 
It depends on what you're after.

If you're looking for equivalents of Word or Excel, no, they're not there yet. You have a choice between feature-poor wysiwyg editors (gnome, kde) and the feature-rich but obscure TeX system (or LaTeX, if you prefer).

For a lot of other things, Linux is indeed ready. This is most obvious in the server area, especially with Apache; but in many ways automation is more similar to servers than to desktops anyway.


Jiri
--
Jiri Baum <[email protected]>
http://www.csse.monash.edu.au/~jiribvisit the MAT LinuxPLC project at http://mat.sf.net
 
No! No, it isn't.

Automation _used_ to be about hardware and process control.

Now, those things are tools, subsets of where the real action is for the future.

The real action is in providing the integration capability to as inexpensively as possible, integrate the entire enterprise, cross-plant and
vertical.

Microsoft has made a huge effort in this area, that is just starting to pay off for them. UDDI, for example.

The open source movement is still at the tweaking hardware stage. It _can_ catch up, it _may_ catch up, but not until somebody with a clue says,
"Here's $100 million. Make an OS that is packaged and clean, and better than Windows for the 80% application."

Give me a guaranteed user base of 200 million, and a $25 royalty on each OS sold, and I'll do it myself.

Get a clue, friends. It isn't about automation, or process control. It is about generating economies in manufacturing and distribution. If Linux or Snoopux or Charliebrownux can do that, easily and cleanly, it will win. And if there are no paybacks, it is doomed. It is about return on investment, not how cool the OS is.

Walt Boyes
co-author of "e-Business in Manufacturing: Putting the Internet
to Work in the Industrial Enterprise" ISA Press--September 2001 ISBN:
1-55617-758-5
____________________________________________

---------------------------------------------
Walt Boyes -- MarketingPractice Consultants
[email protected]
21118 SE 278th Place - Maple Valley, WA 98038
253-709-5046 cell 425-432-8262 home office
fax:801-749-7142 ICQ: 59435534

"Strategic marketing, sales and electronic
business consulting for the small and medium-sized
enterprise: http://www.waltboyes.com"
---------------------------------------------
 
Walt wrote:
> No! No, it isn't.

Not sure what you're disagreeing with here...

That linux is ready or that automation resembles servers more than desktops?

If the latter: Like servers, and unlike desktops, automation is characterized by requirements for unattended, reliable operation, relatively qualified administration, lock-down requirements.

To be sure, it has some specifics all its own.

> The real action is in providing the integration capability to as
> inexpensively as possible, integrate the entire enterprise, cross-plant
> and vertical.

Linux supports many open protocols (and a few closed ones). In contrast, Microsoft provides a single-vendor solution, perhaps more comprehensive, but not greatly so.

> It _can_ catch up, it _may_ catch up, but not until somebody with a clue
> says, "Here's $100 million. Make an OS that is packaged and clean, and
> better than Windows for the 80% application."

You mean like IBM?

> It isn't about automation, or process control. It is about generating
> economies in manufacturing and distribution. If Linux or Snoopux or
> Charliebrownux can do that, easily and cleanly, it will win.

(FWIW, Linux is not named for the Charlie Brown character, but for its original author, Linus Torvalds.)

> And if there are no paybacks, it is doomed. It is about return on
> investment, not how cool the OS is.

I'm not sure what paybacks you refer to here - from manufacturing or from the OS itself?

For the OS itself, the argument is moot: the fact of the matter is that Linux+Apache is already one of the best combinations for Web serving. The
economic explanation may be interesting, but only to economists. As far as we are concerned, Linux simply is.

For manufacturing, the OS used for control doesn't matter, as long as it provides the services, reliability, flexibility etc required. So - what do you look for in an OS you're going to use for control?


(This last is a serious question, btw - if there's something specific lacking from Linux, we want to know about it, so that we can fix it.)


Jiri
--
Jiri Baum <[email protected]>
http://www.csse.monash.edu.au/~jiribvisit the MAT LinuxPLC project at http://mat.sf.net

 
E
Jiri wrote:

> If you're looking for equivalents of Word or Excel, no, they're not there
> yet. You have a choice between feature-poor wysiwyg editors (gnome, kde)
> and the feature-rich but obscure TeX system (or LaTeX, if you prefer).


Not even StarOffice?

http://www.sun.com/staroffice/
Ed
Speaking for me, not for Starbucks. . .
 
Curt Wuollet wrote:
> I can't imagine a more idiotic scenario than to keep your
> process control and automation data on someone else's server.

Stop the FUD, Curt. There is absolutely nothing about .NET or the .NET vision that includes a requirement of storing your own corporate data on
another companies server.

Just because .NET happens, does that mean that suddenly Wonderware's historical logger is going to start streaming process data to Microsoft?
Does that mean that RSSQL will only connect to Microsoft Corporate SQL Servers and that you won't be able to install a SQL server in your own
facility? Does that mean that we'll never need our own hard drives again?

No, that's not the Microsoft vision. Sun's vision, maybe - but not Microsoft.

Jeff Dean
[email protected]
 
Walt wrote:
> No, Curt, it will be bought because none of Microsoft's competitors,
> including the open source movement can market their way out of a torn wet
> paper sack.

Perhaps... That there monopolist that'll stomp over anyone preinstalling it on any sizeable scale doesn't help, either.

> There is a huge market for an office suite that doesn't report to Mother
> Redmond every ten minutes.

Yes, like I wrote, equivalents of Word and Excel just aren't there yet. Neither Gnome nor KDE is up to it yet, and TeX is obscure. For the time
being, the best we can offer is StarOffice (not open source).

> But most people aren't programmers. Even many systems integrators don't
> like doing custom programming. It raises the bar to attract new
> customers. The more "function blocks" that are already written, the
> cheaper each job becomes.

Obviously. That's why a standard linux install includes between one and two thousand utilities and commands, for dealing with text files, images, etc.

It's also the secret behind the Apache webserver - a bunch of webmasters got together and decided that instead of each of them doing custom programming, they'll all do it together. Spread the work and the benefits.

> There is an analogy in the back office there too.

Yup, Apache.

> What is needed is somebody to package an opensource OS/office suite and
> sell it.

Red Hat ? SuSE ? Debian ?


BTW, I note you don't address Curt's security concerns; perhaps he phrased them rather irreverently, but they are reasonable.

Jiri
--
Jiri Baum <[email protected]>
http://www.csse.monash.edu.au/~jiribvisit the MAT LinuxPLC project at http://mat.sf.net
 
J

Joe Jansen/ENGR/HQ/KEMET/US

Jiri,

I'll bite on that one. I am all over the open source idea. My problem is that I am not a C programmer. If I want to write something that will gather data from a PLC, pack it up, send it off through the WAN to a also-freshly-written client for display on my desktop, I whip out Visual Basic 6.0 Enterprise. I have not seen anything in Linux that can let me do it as easily as VB. That is what *I* am looking for.

None of my controls are PC based. Any critical data is stored in the PLC because I know that Wonder-Where? may crap out at any second. That is how I handle both reliability and security. Don't trust the PC.

Ease of programming, tho, is king for simple data collections that are temp jobs to catch a fault and send me what I need to know. I realize that DDE to RSLinx sucks, but it is pretty simple to set up. Winsock may not be the greatest implementation in the world, but drag, drop, punch in some port and IP info, and it is sending my data off into the great beyond. At my client end, set my port and IP, connect, and viola! Here comes my data. Simple and straight forward to a non-C programmer.

I am willing to sweat out coding elegance and good design work for a machine control. But for a utility that I am going to use for a week or
two and toss, what can I use in Linux that offers that ease of setup? (That, too, is a serious question. If something exists, I'd love to hear).

Let's face it: Not everyone knows C. And that seems to be a requirement for working seriously with Linux. I have bought RH 6.2, downloaded and
installed RH 7 and mandrake 7. My computer at home is a dual boot. Yet even though I don't *think* I'm an idiot, none of the downloaded packages seem to ever work for me. NONE. I tried getting a driver for my digital camera. Found it on the web. It was an exact match to my camera's model number. Didn't work. I tried getting StarOffice to install. Complains about libraries. Get all the library rpm's installed, and staroffice simply won't launch. I try it under KDE, under Gnome, from command line. nothing. Linux is not ready for non-geek home use. Curt, I realize that you have deployed it across your entire workforce. You are also there to make it work. I gave up. I haven't booted to Linux for months now. Mainly because I cannot get anything to run that doesn't install with the
distro. Bottom line is Linux is not a system for someone who doesn't want to spend their life doing it.

</Tirade>
</RANT>

--Joe Jansen
 
A
Curt, read my earlier post, and do a little reading on .NET. Only one facet of .NET is the "things run on our server" part a la Microsoft's Passport. I would like to point out that Ximian, led by Miguel De Icaza, of the GNOME
project, is doing an open-source initiative to make their own version Microsoft's Common Runtime Language, a compiler for C#, and a set of class
libraries. Check it out at http://www.go-mono.com/index.html. There's another one, but I forget the name.

From the Mono FAQ:
> Question 8: You guys should innovate instead of copying.
>
> In this particular case, we see a clear advantage in the platform
> and we are interested in using the features of the CLI on open source systems.

Your hatred for everything Microsoft blinds you to the fact that not everything they do(technically) sucks. It's like the converse of the Not Invented Here syndrome... the "It Better Not Be Invented There" syndrome.

Alex Pavloff
Software Engineer
Eason Technology
 
My point exactly, Joe. Thanks for making it so well.

Walt Boyes
co-author of "e-Business in Manufacturing: Putting the Internet
to Work in the Industrial Enterprise" ISA Press--September 2001 ISBN:
1-55617-758-5
____________________________________________

---------------------------------------------
Walt Boyes -- MarketingPractice Consultants
[email protected]
21118 SE 278th Place - Maple Valley, WA 98038
253-709-5046 cell 425-432-8262 home office
fax:801-749-7142 ICQ: 59435534

"Strategic marketing, sales and electronic
business consulting for the small and medium-sized
enterprise: http://www.waltboyes.com"
---------------------------------------------
 
interspersed at @@@@

Jiri Baum said:

Walt:
> No! No, it isn't.

Not sure what you're disagreeing with here...

That linux is ready or that automation resembles servers more than desktops?

@@@@Linux is not ready, and automation is only a tiny fraction of what is going on here.

<lots of argument snipped>
Linux supports many open protocols (and a few closed ones). In contrast, Microsoft provides a single-vendor solution, perhaps more comprehensive, but not greatly so.

@@@@It isn't about Linux vs Microsoft! It is about the Linux contingent getting a clue about how the world works. If you want to beat MS, you have to do what they do well: provide simple solutions. Remember the joke about Linux Airways, where you can get anywhere you need to go, but they only hand you a chair and the manual? If open source wants to really overcome the evil empire, it is going to have to do that by marketing products that people want and need. You can't just hand them a toolbox and a manual and tell them how much better it would be if only they took the time to learn how. Why do you think AOL has so many users? If your answer is, "who cares? They are all stupid," you have missed the thrust of what has quietly made Steve Case wealthy, and much less hated than Gates.

> It _can_ catch up, it _may_ catch up, but not until somebody with a clue
> says, "Here's $100 million. Make an OS that is packaged and clean, and
> better than Windows for the 80% application."

You mean like IBM?

@@@@Sure, look what they did to a perfectly good OS, twice.

> It isn't about automation, or process control. It is about generating
> economies in manufacturing and distribution. If Linux or Snoopux or
> Charliebrownux can do that, easily and cleanly, it will win.

(FWIW, Linux is not named for the Charlie Brown character, but for its original author, Linus Torvalds.)

@@@@God, don't any of you Linux people have a sense of humor? Boyscouts, all of you! Of course I know that. But from the way Linus and the rest of you are acting, you might as well be, in comparison to the way Microsoft runs its business.

> And if there are no paybacks, it is doomed. It is about return on
> investment, not how cool the OS is.

I'm not sure what paybacks you refer to here - from manufacturing or from the OS itself?

For the OS itself, the argument is moot: the fact of the matter is that Linux+Apache is already one of the best combinations for Web serving. The
economic explanation may be interesting, but only to economists. As far as we are concerned, Linux simply is.

@@@@I understand that you all believe that Linux simply is. It is a matter of faith. But the fact remains that there is more to automation, and more to the issues of manufacturing than what OS you use, and which OS has the best support. Yes, Linux+Apache is the best I've found for Web serving. In a high maintenance environment, surrounded by techies and programmers, it
works exceedingly well. It sucks bigtime for small web businesses that are started by non-programmers, though. That's why MS is selling lots of FrontPage, and why MS is selling lots of server software. People buy what they trust, and they know and trust MS...granted, that is becoming rather frayed, but it is still there.

For manufacturing, the OS used for control doesn't matter, as long as it provides the services, reliability, flexibility etc required. So - what do you look for in an OS you're going to use for control?

@@@@First of all, I don't look for an OS that is only good for control. I look for an OS that I can use to completely integrate my enterprise. I want to look for an OS that leading HMI manufacturers and leading SCM, ERP, and factory floor integration software makers write for. I want to look for an OS that is all those things, and reliable, flexible, etc. When you can show
me a Linux with a complete suite of software and applications that will allow it to be used to integrate an enterprise as easily as I can do it in Windows, I'll say that Linux is _here_.

Understand me, I am not shooting at Linux, and I am NOT supporting MS. I am commenting as an analyst on the situation as I see it. Open Source is not addressing the issues that won Microsoft the war. And I don't mean the monopoly, either. They won the war before they lost the peace.

Walt Boyes
co-author of "e-Business in Manufacturing: Putting the Internet
to Work in the Industrial Enterprise" ISA Press--September 2001 ISBN:
1-55617-758-5
____________________________________________
---------------------------------------------
Walt Boyes -- MarketingPractice Consultants
[email protected]
21118 SE 278th Place - Maple Valley, WA 98038
253-709-5046 cell 425-432-8262 home office
fax:801-749-7142 ICQ: 59435534

"Strategic marketing, sales and electronic
business consulting for the small and medium-sized
enterprise: http://www.waltboyes.com"
---------------------------------------------
 
Top