Anyone can take information from this weblog as long as (s)he refers to it origin properly.
April 14th, 2006: Independent is not the same as Neutral
Comments received from:
Rob Weemhoff, Senior Consulting IT-Architect at IBM
Sent: vrijdag 25 augustus 2006 14:23
Subject: Weblog April 14th, 2006 independent vs neutralSteven,
I represent IBM in NEN committees, and I assure you the ISO standards process is by design based on interest, and not independent in your definition of the term.
---
Regards,
Ing. Rob G. WEEMHOFFSent: zaterdag 26 augustus 2006 14:45
Subject: RE: Weblog April 14th, 2006 independent vs neutralHi Rob,
Well, I’ve worked, on national and international level, in a number of ISO working groups in the period 1985 up to 2000 too. You’re right, in that timeframe there was a lot of influence from commercial organisations. Because ISO, at least in those years, worked with country delegations this influence was neutralised to a certain degree. I don’t know about the last years, but until the year 2000 ISO was a pretty open (be it unknown) community where everyone could contribute through the proper country delegation. This may have changed when NEN started to ask money from contributors, your words suggest it has.
In my opinion ISO comes/came closest to an organisation able to write independent standards and norms. Including the fact the way of doing means it takes a long time before results are available. Standards introduced by organisations like OMG and The Open Group, in fact combinations of commercial organisations for commercial purposes, are at most neutral (as The Open Group correctly states itself). But some activities claim independence, and that is, in my opinion, misleading.
In my opinion there is a lot of good work to be done by consortia like OMG and The Open Group in the field of agreeing on points of view by IT-vendors. Although, there are a number of basic issues that cannot be standardised at that level because they are not up for neutral compromises through different commercial interpretations. A method like TOGAF, for instance, may be a standard from the organisations that form The Open Group. The fact a conceptual schema contains 9 concepts, as defined and proved in ISO, is above that kind of standard because there may be many ways to work with those 9 concepts, TOGAF is just one of them.
At the end of the day all of this comes down to what can be done for organisations that need information (IT-vendors call them customers). They buy and maintain a lot of IT to get this information. Standardising real professional issues, like the 9 concepts I mentioned as an example, should be above all discussion. I don’t see another organisation then ISO to do this. Application of these kinds of concepts, like in TOGAF in the example, can be done in the other organisations. In other words: basic professional issues/concepts ought to be standardised by another organisation then application of these issues/concepts.
There is another, more practical issue with neutral/independent. Currently IT-vendors, supply-side, also offer services at demand side. It happens quite often that people from IT-vendors (or people in alliance with these organisations) work on both demand and supply side. This leads to, at least, ethical and business problems. Like in the world outside the Information and IT-sector we should introduce independence between demand- and at supply-side. Do you agree with this? It would mean IT-vendors would have to become real contractors.
Best regards,
Steven
Sent: maandag 28 augustus 2006 10:47
Subject: RE: Weblog April 14th, 2006 independent vs neutral[...]
> This may have changed when NEN started to ask money from contributors, your words suggest it
> has.Yes, it has as long as I remember. When I replaced Rob van der Stap (CMG) in NEN SC27 I insisted to pay what they asked while he always refused.
> In my opinion ISO comes/came closest to an organisation able to write independent standards and norms.Agree, because all different interest can be represented. (If they are able to pay)
[...]
> In other words: basic professional issues/concepts ought to be standardised by another organisation then
> application of these issues/concepts.Agree. ISO gives a standard, another organization, preferably demand side, writes a profile. Supply side uses the profile. For example Common Criteria Protection Profiles.
> There is another, more practical issue with neutral/independent. Currently IT-vendors, supply-side, also offer services at demand side. It happens quite
> often that people from IT-vendors (or people in alliance with these organisations) work on both demand and supply side. This leads to, at least,
> ethical and business problems. Like in the world outside the Information and IT-sector we should introduce independence between demand-
> and at supply-side. Do you agree with this? It would mean IT-vendors would have to become real contractors.You mean in the standardization process?
Or IT-vendors should not offer consultancy services?Sent: dinsdag 29 augustus 2006 14:43
Subject: RE: Weblog April 14th, 2006 independent vs neutralHi Rob,
We agree on most issues. I have also refused to pay NEN because I have always paid for my journeys and hotels myself. Travelling 2 or 3 times a year all over the world has cost me about HFL.100.000,- per year without any help form anyone, and that was quite enough for me. What NEN is doing now can only be done by large companies, and that is a great pity. In my experience most knowledge and innovation is in the people and the smaller companies.
> You mean in the standardization process?
> Or IT-vendors should not offer consultancy services?
I mean in practice. I do not propose IT-vendors should not offer consultancy services in general, I mean they should not offer services specifically targeted at demand side issues.
Even stronger: there is a real large area for consultancy on IT, and I do think IT-vendors may be the best organisations to do this kind of work in a large number of cases (I can also imagine some of these services are offered by independent parties). Call it IT-solution- or IT-supply-oriented services.
In practical terms I mean services around requirements and the knowledge an organisation needs to have and manage about their information. These subjects are the basis, for instance, for procurement activities, because this knowledge contains what needs to be in a specification.
And I do mean there should be a discussion on the translation between demand and supply issues. Not a vendor-neutral discussion, but an essential, professional discussion where demand and supply will meet. We tried to have this discussion in ISO ODP and other working groups, but in that timeframe it was not possible to have this kind of discussion. Today it may be.
From: Hans Bool
Sent: maandag 4 september 2006 1:23Hello steven,
Interesting topic, this standardization dilemma. I Have a question about the passage:
"usually contain trade-offs regarding professional rules and principles."
What are these trade-offs exactly, could you give an example?
Do you also mean that the standard only suits / serves the development parties (developing the product and services) and not the maintenance parties? Otherwise I do not see a threat when all stakeholders are equally involved in the development of the standard...
The case against the independent development is that this independent organ is also less committed to the standard. I do think that standardization is infrastructural and therefore serves a central-kind of development-implementation approach, but all the parties should be somehow involved in this process.
Besides, is there a way that we can measure the quality of a standard?
From: Steven
Sent: dinsdag 5 september 1:46The problem is that usually only a small group of people takes part in the development of a standard. These people work for organizations and are in the standard development out of interest and out of commercial issues.
Let me give you an example. Suppose we are talking about the SQL standard. The work on this standard is based on the people that take part. People from IBM input their knowledge together with their experience with the SQL products they deliver. So does Oracle, Sybase etc.
Now suppose a change to the existing standard is proposed. For instance another way a specific statement works. It may be, that IBM and Oracle have a different implementation, and when the standard is changed they will need to change their products in the marketplace to conform to the new standard. This will usually also affect their installed base, and it is not only the changing of the SQL-product that will cost money, but also the change of the installed base, the applications that are built by the current solution. There are examples where a change can cost a 500 million dollar for an IT-vendor.
Now, if you are going to discuss the new standard, their may be a number of participants who have these kind of problem. I say kind of problem, because they will usually have a different implementation of the standard and therefore have different problems. The usual way to solve is to try and compromise. So, a kind of solution that will have less cost for everyone, at least the strong vendors. If you are working with country delegations, and vendors have people in several of these delegations an extensive negotiation process frequently starts. And in the case of SQL not one but tens or hundreds of these situations may occur. You can imagine the tradeoffs that are discussed, and the consequences such tradeoffs may have for the vendors.In this process we have not yet talked about the way SQL should work. As you know all is based in mathematics and the real working of things is very much fixed. But in the tradeoff process, especially when we are talking about relatively new standards, the commercial issues are usually very strong, and the tradeoff may be in the working of the standard itself. To be solved a next time.
Back to your questions:
… There are very many standards. Some are directed towards development, others may be bound by operation. This may even depend on the crowd that is developing the standard, because these crowds usually do not have all groups available. Just look at the cost, no vendor can participate in all of the 100's/1000's of standard activities at the same time. I, for instance, have worked about 10 years on Open Distributed Processing (ODP). At a certain point in time we had to write what we called the information and the enterprise viewpoint. But the crowd consisted of IT-engineers whose education and experience was in system software. The only thing they could say about enterprises was that these enterprises have to formulate policies and levels of quality of service. This is why these two viewpoints are so weak in this standard. And this was a standard where the system software products and their development were prominent. It was the wrong crowd for that part.… For independent standard development: should you accept "standards" from the Open Group as such? They are a combination of 6 IT-vendors and a number of involved organizations who look at subjects from their viewpoint. Sure, there will some fundamental discussions, but in the end they will each have to earn money with the result. This is what is called neutral, based on agreement and negotiation. And 6 organizations is just a small part of all.
Well, you say that standardization is infrastructural and therefore serves a central-kind of development-implementation approach, but all the parties should be somehow involved in this process. This is why I made the difference between neutral and independent is important. Standards like SQL, the real length of 1 metre, the definition of application etc. should be independent: "everybody" involved and agreeing (impossible, of course, but one should try). Standardizing a method like TOGAF, The Open Group Architecture Framework, can be neutral. It is the use of basic principles in a method that is to be used in the market. Conformance of TOGAF to the basic standards should be guaranteed, but usually is not.
Measure the quality of a standard? Good question, but a very difficult answer. Can you determine the quality of the definition of the term application? I have worked on Conceptual Schema Modeling Facilities. At a certain point in time prof. John Sowa was able to give mathematical proof based on object orientation and set theory that a conceptual specification is only based on 9 concepts. Is this a high quality standard? The fact was to hard for world, because in the end they abandoned the standard. Too many changes to be made in practice, although the mathematical proof was there. Another example: ISO 9000. Is this a high quality standard? It has to be translated to practice before it can be used, but it is widely used? You tell me.
Another issue is the conformance issue. Every standard has a paragraph on conformance: what to do if you have created something to be able to say it conforms to a standard? And who checks this? I know, a few years ago, of 26 organizations who check ISO 9000 in the Netherlands. And they must be accredited to do this work.
It would be great to be able to measure the quality of a standard. But, again, also this is up for discussion because of the cost of it all.