automationtechies.com column for December is up

A
Once more into the breach!

Curt, I'm not doing this because I think Linux is a bad platform by any means -- I just think that some of your statements are wrong, and others
don't really convey the whole story.

> From: Curt Wuollet [mailto:[email protected]]
> To: [email protected]
> Subject: Re: INFO: automationtechies.com column for December is up
>
> That, Bob, is exactly what we are attempting to do. Exactly! But because
> of lack of familiarity, misperceptions and a good deal of FUD,
> the feedback I get is that something hosted on Linux can't possibly be
> easy to use.

Linux can be easy to use. That's a matter of software. At this point, Linux is only easy to use for techies like you and I. To make something easy to use, you need to have a GUI interface. The quality of the GUI apps available out there for various tasks is varied. Some apps (like The Gimp, an excellent and free alternative to Photoshop) are pretty good. But they're rare, and overwhelmed by the applications designed by programmers, for programmers. There aren't a lot of programmers in the world -- and most of them aren't in the automation industry.

> And any software that someone is willing to give to you
> can't have any value. And that Windows is somehow important
> to, and good for, automation and control when it's the least reliable,
> most problematic OS ever offered and has been for it's history.

Curt, I've been using Windows since 3.1. It's gotten a hell of a lot better. Windows 2000 and XP are, in my experience as both a developer and network administrator for my small company, the most solid desktop operating systems I've ever seen. Heck, I compare my Windows NT workstation to the Mandrake 8.1 workstation sitting right next to it, and as a desktop OS, I'd say they're running neck and neck in reliability, with Windows getting the nod for usability.

> And lately they are bent on exploiting their monopoly to greatly increase
cost.

In 1995, a Windows 95 upgrade cost $95 and I believe a "full version" cost $195. Today, a Windows XP Home upgrade costs $99 and a full version of XP Home costs $199.99.

Where's the great increase in cost?

> Fortunately, some companies most notably IBM, see this and are busy
> porting their crown jewels to Linux to offer least cost, Openness
> and reliability. Their reasoning is that, if given a choice between a
> Free, Open and reliable OS or an expensive,kinda good enough,
> but very popular OS, people will do the sensible thing.

The analogy doesn't work. IBM is in the solutions & support industry. IBM jumped on the Linux bandwagon because it means they don't have to spend as much development time/money on AIX (their version of Unix). They go to Linux and can then offer their customers lower cost packages running on x86 hardware, while still signing them up for a nice service contract. The money comes on the service contract.

Besides, when you get to large installations, whether they be in the automation or IT industry, the cost of the software is but a mere fraction of the entire machine.

> Bearing the fresh scars of that battle, I think they overestimate people
> in general. I have been hoping that engineers and technical people are
> more likely candidates. It's an indirect sell because most folks here
> use Windows for editing and peripheral functions but use PLCs
> for control.

That is correct. Now, I don't want to be mean here, but if you've got a SoftPLC that's open source and feature laden, then by all means, give it to us. The LinuxPLC project has been pre-alpha for a long time now, and you've been telling us how great it will be to have an open source control package. I know that you & Jiri & the other contributors don't have a lot of time to work on it, but if you think there's a market for open source Linux automation, then go after it, make a pile of money, and then climb to the top of this pile of money and taunt us Microsoft-using peasants before driving off in your brand new Ferrari.

Don't tell us -- show us.

> The combination is "good enough".

Yup. Good enough -- job gets done. Why should anyone switch?

> And they address the Openness issue by simply folding and having all
<vendor> shops
> and simply not doing what <vendor> doesn't want them to do. That leaves
cost. Cost
> is what is responsible for the slump. Cost is responsible for putting out
quotes
> and losing the jobs. And cost isn't very important until no one is doing
any capital
> investment or improvements because it costs too much. Now it's important.
How much
> can you charge people before they say no? How much off the bottom line
before you
> get the job? Would taking all the licenses and software costs off do it?
Would
> commodity networks do it? Would generic hardware do it? Wouldn't charging
what
> was left plus your billable hours do it? And wouldn't _your_ profit be the
same?
> Seems like an idea worth supporting to me.

Curt, arguing that the current economic slump is due to the increased cost of software is ludicrous.

Alex Pavloff
Software Engineer
Eason Technology
 
C
My point here Alex, is that easy to use automation software can be written on top of Linux. Yes, Linux has a lot of stuff that appeals to techies. We do seem to have a fairly large compliment of techies here. And it's not like the quality of the existing software most have to use is universally excellent because it's on Windows, or even just universally excellent for that matter. Since we would be talking about new applications, I don't see what the intent or quality of the existing non-automation apps has to do with it. The commercial stuff I have seen on Linux has been pretty good. If we could harness the creativity and talent reading this, there is simply no reason why our community owned apps couldn't be outstanding. A community like this would have the knowledge and experience to do an excellent job and the resources would be far greater than any one company could ever afford to devote to GUI design. And the low level "program that serial port" and "talk to that machine tool" stuff would just be the icing on the cake.

> > And any software that someone is willing to give to you
> > can't have any value. And that Windows is somehow important
> > to, and good for, automation and control when it's the least reliable,
> > most problematic OS ever offered and has been for it's history.
>
> Curt, I've been using Windows since 3.1. It's gotten a hell of a lot
> better. Windows 2000 and XP are, in my experience as both a developer and
> network administrator for my small company, the most solid desktop operating
> systems I've ever seen. Heck, I compare my Windows NT workstation to the
> Mandrake 8.1 workstation sitting right next to it, and as a desktop OS, I'd
> say they're running neck and neck in reliability, with Windows getting the
> nod for usability.

Well I'll agree it has got a lot better. But for what I do, Linux functionality and flexibility far outweighs any usability difference. They don't make shrinkwrap to do a lot of things that integrators and people with diverse hardware need to do. An awful lot of the problems I see come up on the list are trivial to solve with the tools and access that come with Linux. Look how often people have to go outside the shrinkwrap anyway. All the little glue bits and tweaks will never dissappear. We might as well make them easy. Linux has it's strength in areas that are important to automation. The other guys are kinda going the other way, away from hardwareaccess, serial ports, headless boxes, bit twiddling, simple filters data conversion and the like. My last project would have been really ugly with a maze of usb-serial converters, media converters, protocol bridges and the like instead of one Linux box and a couple of pages of glue. And a _lot_ more expensive.

> > And lately they are bent on exploiting their monopoly to greatly increase
> cost.
>
> In 1995, a Windows 95 upgrade cost $95 and I believe a "full version" cost
> $195. Today, a Windows XP Home upgrade costs $99 and a full version of XP
> Home costs $199.99.
>
> Where's the great increase in cost?

Yes but that's not a realistic typical installation. It's getting hard to do anything with Windows with just one box. And then there are the applications and seat licenses and it always seems to grow. For tools you would have 1 notebook, 1 OS, and an application of some sort. For some reason, to go beyond that always seems to require an extra box or two in the mix. I have a $600.00 notebook with no dongles and all the tools I could possibly want for aproximately $603.00. I doubt that any current Windows version would even boot on it. It all adds up.

> > Fortunately, some companies most notably IBM, see this and are busy
> > porting their crown jewels to Linux to offer least cost, Openness
> > and reliability. Their reasoning is that, if given a choice between a
> > Free, Open and reliable OS or an expensive,kinda good enough,
> > but very popular OS, people will do the sensible thing.
>
> The analogy doesn't work. IBM is in the solutions & support industry. IBM
> jumped on the Linux bandwagon because it means they don't have to spend as
> much development time/money on AIX (their version of Unix). They go to
> Linux and can then offer their customers lower cost packages running on x86
> hardware, while still signing them up for a nice service contract. The
> money comes on the service contract.
>
> Besides, when you get to large installations, whether they be in the
> automation or IT industry, the cost of the software is but a mere fraction
> of the entire machine.
>
> > Bearing the fresh scars of that battle, I think they overestimate people
> > in general. I have been hoping that engineers and technical people are
> > more likely candidates. It's an indirect sell because most folks here
> > use Windows for editing and peripheral functions but use PLCs
> > for control.
>
> That is correct. Now, I don't want to be mean here, but if you've got a
> SoftPLC that's open source and feature laden, then by all means, give it to
> us. The LinuxPLC project has been pre-alpha for a long time now, and you've
> been telling us how great it will be to have an open source control package.
> I know that you & Jiri & the other contributors don't have a lot of time to
> work on it, but if you think there's a market for open source Linux
> automation, then go after it, make a pile of money, and then climb to the
> top of this pile of money and taunt us Microsoft-using peasants before
> driving off in your brand new Ferrari.

First of all, it's not about money. I'll drive my 91 Dodge and try to do something good for the automation community, thank you. My choice. We have been quite clear that we are providing the stone for the stone soup. If folks will throw in a little of what they've got, we'll have a fine stew fairly quickly. If they don't, you're all still welcome to share, but it's gonna be kinda thin for a while. If we work together a little bit, we can change the industry and drastically lower costs, etc. Or we can be marketed to and herded in the direction deemed most profitable for some very large corporations. I'm fairly confident we can do things better for us.

> Don't tell us -- show us.

Grab the cvs snapshot.

> > The combination is "good enough".
>
> Yup. Good enough -- job gets done. Why should anyone switch?

Perhaps to make the combination really good and much handier.

> > And they address the Openness issue by simply folding and having all
> <vendor> shops
> > and simply not doing what <vendor> doesn't want them to do. That leaves
> cost. Cost
> > is what is responsible for the slump. Cost is responsible for putting out
> quotes
> > and losing the jobs. And cost isn't very important until no one is doing
> any capital
> > investment or improvements because it costs too much. Now it's important.
> How much
> > can you charge people before they say no? How much off the bottom line
> before you
> > get the job? Would taking all the licenses and software costs off do it?
> Would
> > commodity networks do it? Would generic hardware do it? Wouldn't charging
> what
> > was left plus your billable hours do it? And wouldn't _your_ profit be the
> same?
> > Seems like an idea worth supporting to me.
>
> Curt, arguing that the current economic slump is due to the increased cost
> of software is ludicrous.

I'm not arguing that. I'm arguing that the automation slump would benefit greatly from even 20% lower total solution costs. Or that even 20% off your bottom line would get you more work. Or when everybody has all the work they can handle, 20% better margins. And 20% isn't a very ambitious target with what a shared community platform can save over the status quo. I can't do much about the rest of the economy. I suppose I could buy a little newer Dodge. Well, maybe next year.

Regards

cww
 
M
> Curt, I'm not doing this because I think Linux is a bad platform by any
> means -- I just think that some of your statements are wrong, and others
> don't really convey the whole story.

I'm not saying Linux is perfect, but I think that some of your statements are wrong.

> Linux can be easy to use. That's a matter of software. At this point,
> Linux is only easy to use for techies like you and I.

My parents have been using Linux for the last 3 years, and they certainly aren't techies. It came pre-installed on their system. Mom needs W95 for some software she used with her sewing machine, but doesn't want to switch to W95 for email and internet stuff, as she doesn't want to have to install all this complicated software for viruses and security patches.

> To make something easy to use, you need to have a GUI interface.

Linux has several choices in GUI interfaces.

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