Communications - A funny thing happened....


Thread Starter

Roger Irwin

.....on the way to the Forum.......

Usually when I question the wisdom of using Microsoft for everything I receive quite a few off-list replies.

The recent discussions concerning communications protocols were no exception, what was different was that instead of being off-list because the content would not get through moderation, this time the response was of genuine enquiry as to "so how do you do your communications". Presumably people are embarrassed to ask such
questions in public;-)

However, one respondent really threw down the gauntlett, summarised, he stated:

"If you realy believe the only rational solution is to do your own protocolls at the machine end, then there must be a lot of other people in the same situation. Therefore if you believe your protocols are any good, you may as well publish
them. Even if only one other person in the world adopts them, you may gain some synergy and feedback.
......I am suprised that somebody that advocates open source keeps something potentially useful to himself".

Well he does have a point there, so publish I will, and be damned!

You will find a description of my TCP/IP based protocols, along with an introduction, at:

COM is not going away, Roger. I guess everyone should follow you and your white paper rather than all of the companies that are supporting OPC?
No they should not! OPC does not connect to the field device, it connects the PC that is connected to the field device to the desktop.
The PC still communicates with the field device using PROFIBUS/MODBUS.....or in my case, my protocoll. The reason I did it like this
instead of say MODBUS, is that on a device with a simple TCP/IP stack I can connect directly to desktop applications for simple requirements OR I can wrao it in a COM object, even an OPC one, then my 'whitepaper' would become an OPC solution,
IF required.

You seem to be a bit confused between 'outward appearance' and manifestation.

Also, you are confused between COM and DCOM. I have never said COM is going away, I have said DCOM (as a COM transport) is going away. My protocol could be viewed as one of many alternatives for transporting COM objects. OPC themselves are proposing alternative means of transporting COM objects, but they are not ready yet and when they are I am afraid they will not necessarily be implementable on devices with
minimal TCP/IP stacks, which is what is actually emerging onto the market and what I am looking to use.

As I said in 'Why do it', I did it because I could find no equivalent, but I have done it in a way I can make it coexist with other standards. I made it available because asked. It does not compete with OPC.

BTW, I think COM and DCOM will hang around in the IA industry for some time, but that belief is primarily based on the principle that people in this industry have little knowledge of what else is out there, not the fact that their are 250+ corporate members of OPC. OPC charges relatively little corporate membership, many of these members sign up because they have some OPC support, that does not mean they are univocally committed to it.

Let's face it, Microsoft are corporate members of Xopen, but don't let us expect them to ship X servers anytime soon;-)

The biggest threat to DCOM, and probably the reason Microsoft want to upgrade it to SOAP and XML, is the fact that it has only had limited take-up. You see 250+ corporate members is actually nothing in the ocean of corporate
computing, and the fact of the matter is that IA is one of only a few areas were DCOM has been universally accepted for anything and everything, in other sectors DCOM's 'competitor' CORBA is very widely accepted, look at

There is propably the whole fortune 500 in there. Don't get me wrong. I am not proposing we use CORBA in IA, but it's acceptance in the general
market place means that OS and Development tools developers cannot ignore it. This currently manifests itself with 'DCOM' being offered in
Professinal level tools and applications and CORBA being added at the enterise level.

Now Microsoft want to introduce SOAP technology. How long may we expect serious support for DCOM? If we do not use DCOM, what do we use? I am
keeping my options open, and I am certianaly not going to use DCOM in sitations were it is not adapt just so I can lock my products into a technology which does not necessarily have such a solid future. 250 IA corporations who have some
interest in DCOM are by no means enough to assure it's future support, not by a long chalk.