Issue 17474: Allow multiple operations for a direct measure (multiplicity 0..*) (smm-rtf) Source: Cordys (Mr. Henk de Man, hdman(at)cordys.com) Nature: Revision Severity: Summary: Allow multiple operations for a direct measure (multiplicity 0..*). In addition add attribute “type” (String) to DirectMeasure. When “type” is filled, it corresponds to observation type Resolution: Revised Text: Actions taken: July 13, 2012: received issue Discussion: End of Annotations:===== s is issue # 17474 From: Henk de Man To: Juergen Boldt , "issues@omg.org" , "smm-rtf@omg.org" Subject: RE: issues 17472 - 17475 -- SMM RTF issues Thread-Topic: issues 17472 - 17475 -- SMM RTF issues Thread-Index: AQHNYRB5ZHhwo3QPgUiSoRbKmbLbHZcpGa+g Date: Sat, 14 Jul 2012 18:40:01 +0000 Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.24.11.7] X-OriginalArrivalTime: 14 Jul 2012 18:40:02.0870 (UTC) FILETIME=[1490ED60:01CD61F0] With respect to #17472: These attributes seem better fit for Measure, the method of measurement. A measurement is a single data point obtained by applying a measure. The measure may have precision, confidence bounds and a distribution. With respect to #17473: Each of these types seem to be categories of measures. Perhaps SMM should include some built-in measure categories such as Estimators, Simulators and Benchmarks. With respect to #17474: Measures can be equated. If there are multiple, equally good direct measures for a given characteristic then they are be directly associated by the measure equivalence relationship. With respect to #17475: I.m good with this. From: Juergen Boldt [mailto:juergen@omg.org] Sent: Friday, July 13, 2012 10:59 AM To: issues@omg.org; smm-rtf@omg.org Subject: issues 17472 - 17475 -- SMM RTF issues This is issue # 17472 From: Henk de Man To: Larry Hines Cc: Juergen Boldt , "issues@omg.org" , "smm-rtf@omg.org" , Henk de Man , Alain Picard , Pete Rivett , Arne Berre , "fred.a.cummins" X-Gm-Message-State: ALoCoQksqIh7GVwAF6Mewh9JrqjgeeMHdxauYBrS5eXbuTthMImmfhusrd7koZaMBsSx94Rdlj8F Larry, See below. On Sat, Jul 14, 2012 at 8:40 PM, Larry Hines wrote: With respect to #17472: These attributes seem better fit for Measure, the method of measurement. A measurement is a single data point obtained by applying a measure. The measure may have precision, confidence bounds and a distribution. [hdm] ÂWith respect to #17472: It is not correct to state that a measurement is a single data point. See e.g.ÂHubbard, Douglas W., How to Measure Anything, Finding the Value of âIntangiblesâ in Business, John Wiley & Sons, New Jersey, 2010. His main point is, that, in business measurements, just point measures aren't good measures. Often the point measure is not known or is not 100 % trustable. It is very important to be able to state measurement (so the result of measurement) as something that resides in a confidence interval, according to a certain level of confidence. ÂA measurement can be a point value. But it should also be possible to state a measurement as e.g. a confidence interval, with a confidence value. So-far about confidence interval and confidence. Now about distribution (stochastic distribution): A Monte Carlo measurement (i.e. outcome of Multi Carlo simulation or experiment) is typically expressed as a set of values that can be best described by a distribution curve. Stochastic enabling of measurement is important. But the distribution itself can not very well be defined as part of the measure, because the measure is re-usable in different contexts, and the distribution will often be specific to that context.   With respect to #17473: Each of these types seem to be categories of measures. Perhaps SMM should include some built-in measure categories such as Estimators, Simulators and Benchmarks. [hdm] ÂWith respect to #17473: It is not about the measure, but about the observation. It is about a specification the type of the observation. Note that the same measure can be used for for estimated measurement, as-is measurement, simulated measurement.  With respect to #17474: Measures can be equated. If there are multiple, equally good direct measures for a given characteristic then they are be directly associated by the measure equivalence relationship. [hdm]ÂWith respect to #17474.ÂThis is not about equivalence. ÂIt is like method overloading in Java ... When the same direct measure is applied in different contexts, slightly different or additional arguments (parameters) might be required in its operation. So, actually, more than one operation would be required. During discussions with Alain Picard this came out as the best solution.  With respect to #17475: Iâm good with this.   From: Juergen Boldt [mailto:juergen@omg.org] Sent: Friday, July 13, 2012 10:59 AM To: issues@omg.org; smm-rtf@omg.org Subject: issues 17472 - 17475 -- SMM RTF issues  This is issue # 17472ÂÂÂFrom: Henk de Man To: Henk de Man CC: Juergen Boldt , "issues@omg.org" , "smm-rtf@omg.org" , Alain Picard , Pete Rivett , Arne Berre , "fred.a.cummins" Subject: RE: issues 17473 - 17474 -- SMM RTF issues Thread-Topic: issues 17473 - 17474 -- SMM RTF issues Thread-Index: AQHNZDR5Qw8fbBSHbEKqTXMfnaVbIA== Date: Tue, 17 Jul 2012 15:54:39 +0000 Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.64.26.13] X-OriginalArrivalTime: 17 Jul 2012 15:54:41.0352 (UTC) FILETIME=[7A1EAC80:01CD6434] The proposal contained in 17473 and 17474 seems rather convoluted. First let me ask what are the semantics of the four observation types? What is an estimated observation, an actual observation, a simulated observation and a benchmark observation? I can imagine definitions for .actual. and even for .benchmark.. Actual would mean that the measure was actually applied. Benchmark would mean that the measure was actually applied for the purposes of obtaining a benchmark. Following this train of thought, I would guess that simulated is that the measure was simulated and not actually applied. Perhaps estimated means that the measure was estimated. This is why the examples you gave appear to me to be categories of measures. If asked about types of observations generally I would assert that observations could be typed as casual, natural, subjective, objective, direct, indirectly, controlled, uncontrolled, intrusive, nonintrusive, etc. Something that seems obvious to me is that the measure used in a subjective observation cannot be the same measure as one used in an objective measure. They may be or may not be assigned the same name, but they are different measures. The two measures would certainly have the same characteristic and the same domain. In 17474 especially you seem to be creating a relationship between measures. Why push the relationship into the definition of DirectMeasure? As a side note, I do think that observation is rather under specified. From: Henk de Man [mailto:hdman@cordys.com] Sent: Tuesday, July 17, 2012 8:01 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Henk de Man; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins Subject: Re: issues 17472 - 17475 -- SMM RTF issues Larry, See below. On Sat, Jul 14, 2012 at 8:40 PM, Larry Hines wrote: With respect to #17472: These attributes seem better fit for Measure, the method of measurement. A measurement is a single data point obtained by applying a measure. The measure may have precision, confidence bounds and a distribution. [hdm] With respect to #17472: It is not correct to state that a measurement is a single data point. See e.g. Hubbard, Douglas W., How to Measure Anything, Finding the Value of .Intangibles. in Business, John Wiley & Sons, New Jersey, 2010. His main point is, that, in business measurements, just point measures aren't good measures. Often the point measure is not known or is not 100 % trustable. It is very important to be able to state measurement (so the result of measurement) as something that resides in a confidence interval, according to a certain level of confidence. A measurement can be a point value. But it should also be possible to state a measurement as e.g. a confidence interval, with a confidence value. So-far about confidence interval and confidence. Now about distribution (stochastic distribution): A Monte Carlo measurement (i.e. outcome of Multi Carlo simulation or experiment) is typically expressed as a set of values that can be best described by a distribution curve. Stochastic enabling of measurement is important. But the distribution itself can not very well be defined as part of the measure, because the measure is re-usable in different contexts, and the distribution will often be specific to that context. With respect to #17473: Each of these types seem to be categories of measures. Perhaps SMM should include some built-in measure categories such as Estimators, Simulators and Benchmarks. [hdm] With respect to #17473: It is not about the measure, but about the observation. It is about a specification the type of the observation. Note that the same measure can be used for for estimated measurement, as-is measurement, simulated measurement. With respect to #17474: Measures can be equated. If there are multiple, equally good direct measures for a given characteristic then they are be directly associated by the measure equivalence relationship. [hdm] With respect to #17474. This is not about equivalence. It is like method overloading in Java ... When the same direct measure is applied in different contexts, slightly different or additional arguments (parameters) might be required in its operation. So, actually, more than one operation would be required. During discussions with Alain Picard this came out as the best solution. With respect to #17475: I.m good with this. From: Juergen Boldt [mailto:juergen@omg.org] Sent: Friday, July 13, 2012 10:59 AM To: issues@omg.org; smm-rtf@omg.org Subject: issues 17472 - 17475 -- SMM RTF issues This is issue # 17472 From: Henk de Man To: Larry Hines Cc: Juergen Boldt , "issues@omg.org" , "smm-rtf@omg.org" , Alain Picard , Pete Rivett , Arne Berre , "fred.a.cummins" , Henk de Man X-Gm-Message-State: ALoCoQmgCvkIoGt44DaI1yok9L88v3hDGMP9uDvvPPzFSn+StO344AXk79mlcH1mrtQcaHDKV2tb Larry, See below. In-lined comments. On Tue, Jul 17, 2012 at 5:54 PM, Larry Hines wrote: The proposal contained in 17473 and 17474 seems rather convoluted.  First let me ask what are the semantics of the four observation types? [hdm] The suggestion was not about four types, but just a string to allow any type. The four types were just examples. So, just something as simple as a string to enter the type of observation.  ÂÂWhat is an estimated observation [hdm] An observation which measurements have values that are just business analyst "guesstimates", or "values of planned or targeted performance". Note that we talk about measurements of characteristics of elements of business models / business systems. VDML applies SMM to the area of business design. , an actual observation [hdm] When parts of the business design have been implemented in real world, measurements can be created in the real world (actual performance), and these might be fed back to the model, per separate observation on the same thing.  , a simulated observation [hdm] Measures executed during a simulation game. They can be imported back into the model, per separate observation.  and a benchmark observation? [hdm] E.g. measurements that express industry benchmarks for similar situations, stored in the model per separate observation. ÂI can imagine definitions for âactualâ and even for âbenchmarkâ. Actual would mean that the measure was actually applied. Benchmark would mean that the measure was actually applied for the purposes of obtaining a benchmark. Following this train of thought, I would guess that simulated is that the measure was simulated and not actually applied. Perhaps estimated means that the measure was estimated. ÂThis is why the examples you gave appear to me to be categories of measures. [hdm] This is not about categories of measures. It is about measurements (to outcomes of measuring), Âstored per observation, but organized in a way that we can distinguish them (so per separate observation). Note that VDML integrates with SMM Observations. Btw: the VDML-SMM integration is quite significant.  If asked about types of observations generally I would assert that observations could be typed as casual, natural, subjective, objective, direct, indirectly, controlled, uncontrolled, intrusive, nonintrusive, etc.  [hdm] All good, but let's not try to fix a list in an enum, but rather something simple as a string (which was Alain Picards proposal). The same string can then be used for the "four" example types that I gave. Something that seems obvious to me is that the measure used in a subjective observation cannot be the same measure as one used in an objective measure.  [hdm] The measure is one, and a re-usable one in a library. The name is given, the characteristic (or trait) is given, the unit is given, and if defined, the operation or formula or functor or accumulator is given, etc. But different observations result in different measurements against the same measure. Here we hit on a triggering dilemma: In our business modeling context, there are very many dependencies between measures (thru aggregation thru binary or collective measures and/or thru re-scaling). So, if manual entry of guesstimates, querying of actuals, etc. would all require different measures, than this would lead to redundancy of almost complete libraries. Not just of direct measure, but also all aggregates that aggregate from them.. This would not be maintainable. What would be required is this: A direct measure has e.g. a query defined, which can be used for querying of actual performance (e.g. from a process instance database). The same measure can however be used, to just enter the value manually, as an estimate. But per different observation. So, in other words: dependent on the observation, the operation is used or not used. This is a special case of what was suggested per the other issue: allowing multiple operations on a direct measure. Note that any aggregated measure (rescaled, binary, collective) does not have this issue, but it just executes is formula or functor or accumulator. This will lead to a much better and maintainable situation: the aggregated measures can aggregate from underlying direct or aggregated measurements, regardless of how the leave measurements were created (estimates, or query result, etc.). Otherwise we would have to duplicate all these aggregated measures also.. !  ÂThey may be or may not be assigned the same name, but they are different measures. The two measures would certainly have the same characteristic and the same domain.  In 17474 especially you seem to be creating a relationship between measures. Why push the relationship into the definition of DirectMeasure? [hdm] 17474 is about allowing 0..* operations on a direct measure. For libraries of business measures to be maintainable, it is required to allow for this. Above I explained that we want to re-use the same measure for manual entry of guestimate as well as for actuals querying. Similar applies to variations of queries. E.g. exactly the same measure is applied in different parts of the business system. However, in one part of the business a different query service is used than in other part. We cannot duplicate the measure for this, because otherwise all rescaled, binary and collective measures, that are based on them, would have to be duplicated as well, which would kill the library approach, as the re-use would be gone.  As a side note, I do think that observation is rather under specified. [hdm] I am all in for building more on that, when we can keep the library pure and better re-usable.   From: Henk de Man [mailto:hdman@cordys.com] Sent: Tuesday, July 17, 2012 8:01 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Henk de Man; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins Subject: Re: issues 17472 - 17475 -- SMM RTF issues  Larry,  See below. On Sat, Jul 14, 2012 at 8:40 PM, Larry Hines wrote: With respect to #17472: These attributes seem better fit for Measure, the method of measurement. A measurement is a single data point obtained by applying a measure. The measure may have precision, confidence bounds and a distribution. [hdm] ÂWith respect to #17472: It is not correct to state that a measurement is a single data point. See e.g.ÂHubbard, Douglas W., How to Measure Anything, Finding the Value of âIntangiblesâ in Business, John Wiley & Sons, New Jersey, 2010. His main point is, that, in business measurements, just point measures aren't good measures. Often the point measure is not known or is not 100 % trustable. It is very important to be able to state measurement (so the result of measurement) as something that resides in a confidence interval, according to a certain level of confidence. ÂA measurement can be a point value. But it should also be possible to state a measurement as e.g. a confidence interval, with a confidence value. So-far about confidence interval and confidence. Now about distribution (stochastic distribution): A Monte Carlo measurement (i.e. outcome of Multi Carlo simulation or experiment) is typically expressed as a set of values that can be best described by a distribution curve. Stochastic enabling of measurement is important. But the distribution itself can not very well be defined as part of the measure, because the measure is re-usable in different contexts, and the distribution will often be specific to that context.   With respect to #17473: Each of these types seem to be categories of measures. Perhaps SMM should include some built-in measure categories such as Estimators, Simulators and Benchmarks. [hdm] ÂWith respect to #17473: It is not about the measure, but about the observation. It is about a specification the type of the observation. Note that the same measure can be used for for estimated measurement, as-is measurement, simulated measurement.  With respect to #17474: Measures can be equated. If there are multiple, equally good direct measures for a given characteristic then they are be directly associated by the measure equivalence relationship. [hdm]ÂWith respect to #17474.ÂThis is not about equivalence. ÂIt is like method overloading in Java ... When the same direct measure is applied in different contexts, slightly different or additional arguments (parameters) might be required in its operation. So, actually, more than one operation would be required. During discussions with Alain Picard this came out as the best solution.  With respect to #17475: Iâm good with this.   From: Juergen Boldt [mailto:juergen@omg.org] Sent: Friday, July 13, 2012 10:59 AM To: issues@omg.org; smm-rtf@omg.org Subject: issues 17472 - 17475 -- SMM RTF issues  This is issue # 17472ÂÂÂFrom: Henk de Man To: Larry Hines Cc: Juergen Boldt , "issues@omg.org" , "smm-rtf@omg.org" , Alain Picard , Pete Rivett , Arne Berre , "fred.a.cummins" , Henk de Man X-Gm-Message-State: ALoCoQmgCvkIoGt44DaI1yok9L88v3hDGMP9uDvvPPzFSn+StO344AXk79mlcH1mrtQcaHDKV2tb Larry, See below. In-lined comments. On Tue, Jul 17, 2012 at 5:54 PM, Larry Hines wrote: The proposal contained in 17473 and 17474 seems rather convoluted.  First let me ask what are the semantics of the four observation types? [hdm] The suggestion was not about four types, but just a string to allow any type. The four types were just examples. So, just something as simple as a string to enter the type of observation.  ÂÂWhat is an estimated observation [hdm] An observation which measurements have values that are just business analyst "guesstimates", or "values of planned or targeted performance". Note that we talk about measurements of characteristics of elements of business models / business systems. VDML applies SMM to the area of business design. , an actual observation [hdm] When parts of the business design have been implemented in real world, measurements can be created in the real world (actual performance), and these might be fed back to the model, per separate observation on the same thing.  , a simulated observation [hdm] Measures executed during a simulation game. They can be imported back into the model, per separate observation.  and a benchmark observation? [hdm] E.g. measurements that express industry benchmarks for similar situations, stored in the model per separate observation. ÂI can imagine definitions for âactualâ and even for âbenchmarkâ. Actual would mean that the measure was actually applied. Benchmark would mean that the measure was actually applied for the purposes of obtaining a benchmark. Following this train of thought, I would guess that simulated is that the measure was simulated and not actually applied. Perhaps estimated means that the measure was estimated. ÂThis is why the examples you gave appear to me to be categories of measures. [hdm] This is not about categories of measures. It is about measurements (to outcomes of measuring), Âstored per observation, but organized in a way that we can distinguish them (so per separate observation). Note that VDML integrates with SMM Observations. Btw: the VDML-SMM integration is quite significant.  If asked about types of observations generally I would assert that observations could be typed as casual, natural, subjective, objective, direct, indirectly, controlled, uncontrolled, intrusive, nonintrusive, etc.  [hdm] All good, but let's not try to fix a list in an enum, but rather something simple as a string (which was Alain Picards proposal). The same string can then be used for the "four" example types that I gave. Something that seems obvious to me is that the measure used in a subjective observation cannot be the same measure as one used in an objective measure.  [hdm] The measure is one, and a re-usable one in a library. The name is given, the characteristic (or trait) is given, the unit is given, and if defined, the operation or formula or functor or accumulator is given, etc. But different observations result in different measurements against the same measure. Here we hit on a triggering dilemma: In our business modeling context, there are very many dependencies between measures (thru aggregation thru binary or collective measures and/or thru re-scaling). So, if manual entry of guesstimates, querying of actuals, etc. would all require different measures, than this would lead to redundancy of almost complete libraries. Not just of direct measure, but also all aggregates that aggregate from them.. This would not be maintainable. What would be required is this: A direct measure has e.g. a query defined, which can be used for querying of actual performance (e.g. from a process instance database). The same measure can however be used, to just enter the value manually, as an estimate. But per different observation. So, in other words: dependent on the observation, the operation is used or not used. This is a special case of what was suggested per the other issue: allowing multiple operations on a direct measure. Note that any aggregated measure (rescaled, binary, collective) does not have this issue, but it just executes is formula or functor or accumulator. This will lead to a much better and maintainable situation: the aggregated measures can aggregate from underlying direct or aggregated measurements, regardless of how the leave measurements were created (estimates, or query result, etc.). Otherwise we would have to duplicate all these aggregated measures also.. !  ÂThey may be or may not be assigned the same name, but they are different measures. The two measures would certainly have the same characteristic and the same domain.  In 17474 especially you seem to be creating a relationship between measures. Why push the relationship into the definition of DirectMeasure? [hdm] 17474 is about allowing 0..* operations on a direct measure. For libraries of business measures to be maintainable, it is required to allow for this. Above I explained that we want to re-use the same measure for manual entry of guestimate as well as for actuals querying. Similar applies to variations of queries. E.g. exactly the same measure is applied in different parts of the business system. However, in one part of the business a different query service is used than in other part. We cannot duplicate the measure for this, because otherwise all rescaled, binary and collective measures, that are based on them, would have to be duplicated as well, which would kill the library approach, as the re-use would be gone.  As a side note, I do think that observation is rather under specified. [hdm] I am all in for building more on that, when we can keep the library pure and better re-usable.   From: Henk de Man [mailto:hdman@cordys.com] Sent: Tuesday, July 17, 2012 8:01 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Henk de Man; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins Subject: Re: issues 17472 - 17475 -- SMM RTF issues  Larry,  See below. On Sat, Jul 14, 2012 at 8:40 PM, Larry Hines wrote: With respect to #17472: These attributes seem better fit for Measure, the method of measurement. A measurement is a single data point obtained by applying a measure. The measure may have precision, confidence bounds and a distribution. [hdm] ÂWith respect to #17472: It is not correct to state that a measurement is a single data point. See e.g.ÂHubbard, Douglas W., How to Measure Anything, Finding the Value of âIntangiblesâ in Business, John Wiley & Sons, New Jersey, 2010. His main point is, that, in business measurements, just point measures aren't good measures. Often the point measure is not known or is not 100 % trustable. It is very important to be able to state measurement (so the result of measurement) as something that resides in a confidence interval, according to a certain level of confidence. ÂA measurement can be a point value. But it should also be possible to state a measurement as e.g. a confidence interval, with a confidence value. So-far about confidence interval and confidence. Now about distribution (stochastic distribution): A Monte Carlo measurement (i.e. outcome of Multi Carlo simulation or experiment) is typically expressed as a set of values that can be best described by a distribution curve. Stochastic enabling of measurement is important. But the distribution itself can not very well be defined as part of the measure, because the measure is re-usable in different contexts, and the distribution will often be specific to that context.   With respect to #17473: Each of these types seem to be categories of measures. Perhaps SMM should include some built-in measure categories such as Estimators, Simulators and Benchmarks. [hdm] ÂWith respect to #17473: It is not about the measure, but about the observation. It is about a specification the type of the observation. Note that the same measure can be used for for estimated measurement, as-is measurement, simulated measurement.  With respect to #17474: Measures can be equated. If there are multiple, equally good direct measures for a given characteristic then they are be directly associated by the measure equivalence relationship. [hdm]ÂWith respect to #17474.ÂThis is not about equivalence. ÂIt is like method overloading in Java ... When the same direct measure is applied in different contexts, slightly different or additional arguments (parameters) might be required in its operation. So, actually, more than one operation would be required. During discussions with Alain Picard this came out as the best solution.  With respect to #17475: Iâm good with this.   From: Juergen Boldt [mailto:juergen@omg.org] Sent: Friday, July 13, 2012 10:59 AM To: issues@omg.org; smm-rtf@omg.org Subject: issues 17472 - 17475 -- SMM RTF issues  This is issue # 17472ÂÂÂFrom: Henk de Man To: Henk de Man CC: Juergen Boldt , "issues@omg.org" , "smm-rtf@omg.org" , Alain Picard , Pete Rivett , Arne Berre , "fred.a.cummins" Subject: RE: issues 17473 - 17474 -- SMM RTF issues Thread-Topic: issues 17473 - 17474 -- SMM RTF issues Thread-Index: AQHNZDR5Qw8fbBSHbEKqTXMfnaVbIJcvbxOA///FyEA= Date: Wed, 18 Jul 2012 19:56:18 +0000 Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.64.26.13] X-OriginalArrivalTime: 18 Jul 2012 19:56:18.0742 (UTC) FILETIME=[65A69560:01CD651F] A measure is an evaluation process that assigns a comparable numeric or symbolic value to an entity in order to characterize a selected quantity or trait of the entity. Guesstimate is one evaluation process for determining small distances. Using a micrometer is a different measure. They characterize the same trait and can have the same unit, but the evaluation processes are not the same. You seem to be want an AbstractMeasure where the evaluation process is not specified and to which we can associate concrete measures that share characteristics, units and scope (or domain). Even though it would look very much like NamedMeasure, an AbstractMeasure would not have any measurements associated with it. From: Henk de Man [mailto:hdman@cordys.com] Sent: Wednesday, July 18, 2012 10:30 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins; Henk de Man Subject: Re: issues 17473 - 17474 -- SMM RTF issues Larry, See below. In-lined comments. On Tue, Jul 17, 2012 at 5:54 PM, Larry Hines wrote: The proposal contained in 17473 and 17474 seems rather convoluted. First let me ask what are the semantics of the four observation types? [hdm] The suggestion was not about four types, but just a string to allow any type. The four types were just examples. So, just something as simple as a string to enter the type of observation. What is an estimated observation [hdm] An observation which measurements have values that are just business analyst "guesstimates", or "values of planned or targeted performance". Note that we talk about measurements of characteristics of elements of business models / business systems. VDML applies SMM to the area of business design. , an actual observation [hdm] When parts of the business design have been implemented in real world, measurements can be created in the real world (actual performance), and these might be fed back to the model, per separate observation on the same thing. , a simulated observation [hdm] Measures executed during a simulation game. They can be imported back into the model, per separate observation. and a benchmark observation? [hdm] E.g. measurements that express industry benchmarks for similar situations, stored in the model per separate observation. I can imagine definitions for .actual. and even for .benchmark.. Actual would mean that the measure was actually applied. Benchmark would mean that the measure was actually applied for the purposes of obtaining a benchmark. Following this train of thought, I would guess that simulated is that the measure was simulated and not actually applied. Perhaps estimated means that the measure was estimated. This is why the examples you gave appear to me to be categories of measures. [hdm] This is not about categories of measures. It is about measurements (to outcomes of measuring), stored per observation, but organized in a way that we can distinguish them (so per separate observation). Note that VDML integrates with SMM Observations. Btw: the VDML-SMM integration is quite significant. If asked about types of observations generally I would assert that observations could be typed as casual, natural, subjective, objective, direct, indirectly, controlled, uncontrolled, intrusive, nonintrusive, etc. [hdm] All good, but let's not try to fix a list in an enum, but rather something simple as a string (which was Alain Picards proposal). The same string can then be used for the "four" example types that I gave. Something that seems obvious to me is that the measure used in a subjective observation cannot be the same measure as one used in an objective measure. [hdm] The measure is one, and a re-usable one in a library. The name is given, the characteristic (or trait) is given, the unit is given, and if defined, the operation or formula or functor or accumulator is given, etc. But different observations result in different measurements against the same measure. Here we hit on a triggering dilemma: In our business modeling context, there are very many dependencies between measures (thru aggregation thru binary or collective measures and/or thru re-scaling). So, if manual entry of guesstimates, querying of actuals, etc. would all require different measures, than this would lead to redundancy of almost complete libraries. Not just of direct measure, but also all aggregates that aggregate from them.. This would not be maintainable. What would be required is this: A direct measure has e.g. a query defined, which can be used for querying of actual performance (e.g. from a process instance database). The same measure can however be used, to just enter the value manually, as an estimate. But per different observation. So, in other words: dependent on the observation, the operation is used or not used. This is a special case of what was suggested per the other issue: allowing multiple operations on a direct measure. Note that any aggregated measure (rescaled, binary, collective) does not have this issue, but it just executes is formula or functor or accumulator. This will lead to a much better and maintainable situation: the aggregated measures can aggregate from underlying direct or aggregated measurements, regardless of how the leave measurements were created (estimates, or query result, etc.). Otherwise we would have to duplicate all these aggregated measures also.. ! They may be or may not be assigned the same name, but they are different measures. The two measures would certainly have the same characteristic and the same domain. In 17474 especially you seem to be creating a relationship between measures. Why push the relationship into the definition of DirectMeasure? [hdm] 17474 is about allowing 0..* operations on a direct measure. For libraries of business measures to be maintainable, it is required to allow for this. Above I explained that we want to re-use the same measure for manual entry of guestimate as well as for actuals querying. Similar applies to variations of queries. E.g. exactly the same measure is applied in different parts of the business system. However, in one part of the business a different query service is used than in other part. We cannot duplicate the measure for this, because otherwise all rescaled, binary and collective measures, that are based on them, would have to be duplicated as well, which would kill the library approach, as the re-use would be gone. As a side note, I do think that observation is rather under specified. [hdm] I am all in for building more on that, when we can keep the library pure and better re-usable. From: Henk de Man [mailto:hdman@cordys.com] Sent: Tuesday, July 17, 2012 8:01 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Henk de Man; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins Subject: Re: issues 17472 - 17475 -- SMM RTF issues Larry, See below. On Sat, Jul 14, 2012 at 8:40 PM, Larry Hines wrote: With respect to #17472: These attributes seem better fit for Measure, the method of measurement. A measurement is a single data point obtained by applying a measure. The measure may have precision, confidence bounds and a distribution. [hdm] With respect to #17472: It is not correct to state that a measurement is a single data point. See e.g. Hubbard, Douglas W., How to Measure Anything, Finding the Value of .Intangibles. in Business, John Wiley & Sons, New Jersey, 2010. His main point is, that, in business measurements, just point measures aren't good measures. Often the point measure is not known or is not 100 % trustable. It is very important to be able to state measurement (so the result of measurement) as something that resides in a confidence interval, according to a certain level of confidence. A measurement can be a point value. But it should also be possible to state a measurement as e.g. a confidence interval, with a confidence value. So-far about confidence interval and confidence. Now about distribution (stochastic distribution): A Monte Carlo measurement (i.e. outcome of Multi Carlo simulation or experiment) is typically expressed as a set of values that can be best described by a distribution curve. Stochastic enabling of measurement is important. But the distribution itself can not very well be defined as part of the measure, because the measure is re-usable in different contexts, and the distribution will often be specific to that context. With respect to #17473: Each of these types seem to be categories of measures. Perhaps SMM should include some built-in measure categories such as Estimators, Simulators and Benchmarks. [hdm] With respect to #17473: It is not about the measure, but about the observation. It is about a specification the type of the observation. Note that the same measure can be used for for estimated measurement, as-is measurement, simulated measurement. With respect to #17474: Measures can be equated. If there are multiple, equally good direct measures for a given characteristic then they are be directly associated by the measure equivalence relationship. [hdm] With respect to #17474. This is not about equivalence. It is like method overloading in Java ... When the same direct measure is applied in different contexts, slightly different or additional arguments (parameters) might be required in its operation. So, actually, more than one operation would be required. During discussions with Alain Picard this came out as the best solution. With respect to #17475: I.m good with this. From: Juergen Boldt [mailto:juergen@omg.org] Sent: Friday, July 13, 2012 10:59 AM To: issues@omg.org; smm-rtf@omg.org Subject: issues 17472 - 17475 -- SMM RTF issues This is issue # 17472 From: Henk de Man To: Larry Hines Cc: Juergen Boldt , "issues@omg.org" , "smm-rtf@omg.org" , Alain Picard , Pete Rivett , Arne Berre , "fred.a.cummins" , Henk de Man X-Gm-Message-State: ALoCoQmF6XKOm6evOzd7BhYvR43Gqn+7HJrOSDwD3OzIOi1Yb+xc72VbLATGgD47t/cDJHDxsNYW Larry, See below. On Wed, Jul 18, 2012 at 9:56 PM, Larry Hines wrote: A measure is an evaluation process that assigns a comparable numeric or symbolic value to an entity in order to characterize a selected quantity or trait of the entity. [hdm] We want to extend this a little bit. It should not only be possible to assign a single numeric or symbolic value, but it should, optionally and additionally, also be possible to 1) specify a confidence, saying e.g. "the measurement value lays, with confidence of 90% between the values x and y". 2) specify the outcome of measurement in terms of a distribution of values (so that we can have 1 measurement object instead of 10.000 in our model..). This is by no means conflicting with what you have, and if these things are optional extensions, they are not in the way for everybody, and they enables VDML to make the use of SMM we need.  Guesstimate is one evaluation process for determining small distances. Using a micrometer is a different measure. ÂÂÂThey characterize the same trait and can have the same unit, but the evaluation processes are not the same. [hdm] I did another attempt to explain the need. See attachment.  You seem to be want an AbstractMeasure where the evaluation process is not specified and to which we can associate concrete measures that share characteristics, units and scope (or domain). Even though it would look very much like NamedMeasure, an AbstractMeasure would not have any measurements associated with it. [hdm] No, we fully comply to and apply the core logic of SMM: applying measures, in observations, to create measurements, against the measures. We do not want measures without measurements. We really apply 100 % of SMM, intensively, in VDML. We only want a few refinements or relaxations, if you wish.   From: Henk de Man [mailto:hdman@cordys.com] Sent: Wednesday, July 18, 2012 10:30 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins; Henk de Man Subject: Re: issues 17473 - 17474 -- SMM RTF issues  Larry,  See below. In-lined comments. On Tue, Jul 17, 2012 at 5:54 PM, Larry Hines wrote: The proposal contained in 17473 and 17474 seems rather convoluted.  First let me ask what are the semantics of the four observation types? [hdm] The suggestion was not about four types, but just a string to allow any type. The four types were just examples. So, just something as simple as a string to enter the type of observation.  ÂÂWhat is an estimated observation [hdm] An observation which measurements have values that are just business analyst "guesstimates", or "values of planned or targeted performance". Note that we talk about measurements of characteristics of elements of business models / business systems. VDML applies SMM to the area of business design. , an actual observation [hdm] When parts of the business design have been implemented in real world, measurements can be created in the real world (actual performance), and these might be fed back to the model, per separate observation on the same thing.  , a simulated observation [hdm] Measures executed during a simulation game. They can be imported back into the model, per separate observation.  and a benchmark observation? [hdm] E.g. measurements that express industry benchmarks for similar situations, stored in the model per separate observation. ÂI can imagine definitions for âactualâ and even for âbenchmarkâ. Actual would mean that the measure was actually applied. Benchmark would mean that the measure was actually applied for the purposes of obtaining a benchmark. Following this train of thought, I would guess that simulated is that the measure was simulated and not actually applied. Perhaps estimated means that the measure was estimated. ÂThis is why the examples you gave appear to me to be categories of measures. [hdm] This is not about categories of measures. It is about measurements (to outcomes of measuring), Âstored per observation, but organized in a way that we can distinguish them (so per separate observation). Note that VDML integrates with SMM Observations. Btw: the VDML-SMM integration is quite significant.  If asked about types of observations generally I would assert that observations could be typed as casual, natural, subjective, objective, direct, indirectly, controlled, uncontrolled, intrusive, nonintrusive, etc.  [hdm] All good, but let's not try to fix a list in an enum, but rather something simple as a string (which was Alain Picards proposal). The same string can then be used for the "four" example types that I gave.  Something that seems obvious to me is that the measure used in a subjective observation cannot be the same measure as one used in an objective measure.  [hdm] The measure is one, and a re-usable one in a library. The name is given, the characteristic (or trait) is given, the unit is given, and if defined, the operation or formula or functor or accumulator is given, etc. But different observations result in different measurements against the same measure. Here we hit on a triggering dilemma: In our business modeling context, there are very many dependencies between measures (thru aggregation thru binary or collective measures and/or thru re-scaling). So, if manual entry of guesstimates, querying of actuals, etc. would all require different measures, than this would lead to redundancy of almost complete libraries. Not just of direct measure, but also all aggregates that aggregate from them.. This would not be maintainable. What would be required is this: A direct measure has e.g. a query defined, which can be used for querying of actual performance (e.g. from a process instance database). The same measure can however be used, to just enter the value manually, as an estimate. But per different observation. So, in other words: dependent on the observation, the operation is used or not used. This is a special case of what was suggested per the other issue: allowing multiple operations on a direct measure. Note that any aggregated measure (rescaled, binary, collective) does not have this issue, but it just executes is formula or functor or accumulator. This will lead to a much better and maintainable situation: the aggregated measures can aggregate from underlying direct or aggregated measurements, regardless of how the leave measurements were created (estimates, or query result, etc.). Otherwise we would have to duplicate all these aggregated measures also.. !  ÂThey may be or may not be assigned the same name, but they are different measures. The two measures would certainly have the same characteristic and the same domain.  In 17474 especially you seem to be creating a relationship between measures. Why push the relationship into the definition of DirectMeasure? [hdm] 17474 is about allowing 0..* operations on a direct measure. For libraries of business measures to be maintainable, it is required to allow for this. Above I explained that we want to re-use the same measure for manual entry of guestimate as well as for actuals querying. Similar applies to variations of queries. E.g. exactly the same measure is applied in different parts of the business system. However, in one part of the business a different query service is used than in other part. We cannot duplicate the measure for this, because otherwise all rescaled, binary and collective measures, that are based on them, would have to be duplicated as well, which would kill the library approach, as the re-use would be gone.  As a side note, I do think that observation is rather under specified. [hdm] I am all in for building more on that, when we can keep the library pure and better re-usable.   From: Henk de Man [mailto:hdman@cordys.com] Sent: Tuesday, July 17, 2012 8:01 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Henk de Man; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins Subject: Re: issues 17472 - 17475 -- SMM RTF issues  Larry,  See below. On Sat, Jul 14, 2012 at 8:40 PM, Larry Hines wrote: With respect to #17472: These attributes seem better fit for Measure, the method of measurement. A measurement is a single data point obtained by applying a measure. The measure may have precision, confidence bounds and a distribution. [hdm] ÂWith respect to #17472: It is not correct to state that a measurement is a single data point. See e.g.ÂHubbard, Douglas W., How to Measure Anything, Finding the Value of âIntangiblesâ in Business, John Wiley & Sons, New Jersey, 2010. His main point is, that, in business measurements, just point measures aren't good measures. Often the point measure is not known or is not 100 % trustable. It is very important to be able to state measurement (so the result of measurement) as something that resides in a confidence interval, according to a certain level of confidence. ÂA measurement can be a point value. But it should also be possible to state a measurement as e.g. a confidence interval, with a confidence value. So-far about confidence interval and confidence. Now about distribution (stochastic distribution): A Monte Carlo measurement (i.e. outcome of Multi Carlo simulation or experiment) is typically expressed as a set of values that can be best described by a distribution curve. Stochastic enabling of measurement is important. But the distribution itself can not very well be defined as part of the measure, because the measure is re-usable in different contexts, and the distribution will often be specific to that context.   With respect to #17473: Each of these types seem to be categories of measures. Perhaps SMM should include some built-in measure categories such as Estimators, Simulators and Benchmarks. [hdm] ÂWith respect to #17473: It is not about the measure, but about the observation. It is about a specification the type of the observation. Note that the same measure can be used for for estimated measurement, as-is measurement, simulated measurement.  With respect to #17474: Measures can be equated. If there are multiple, equally good direct measures for a given characteristic then they are be directly associated by the measure equivalence relationship. [hdm]ÂWith respect to #17474.ÂThis is not about equivalence. ÂIt is like method overloading in Java ... When the same direct measure is applied in different contexts, slightly different or additional arguments (parameters) might be required in its operation. So, actually, more than one operation would be required. During discussions with Alain Picard this came out as the best solution.  With respect to #17475: Iâm good with this.   From: Juergen Boldt [mailto:juergen@omg.org] Sent: Friday, July 13, 2012 10:59 AM To: issues@omg.org; smm-rtf@omg.org Subject: issues 17472 - 17475 -- SMM RTF issues  This is issue # 17472ÂÂÂFrom: Henk de Man To: Henk de Man CC: Juergen Boldt , "issues@omg.org" , "smm-rtf@omg.org" , Alain Picard , Pete Rivett , Arne Berre , "fred.a.cummins" Subject: RE: issues 17473 - 17474 -- SMM RTF issues Thread-Topic: issues 17473 - 17474 -- SMM RTF issues Thread-Index: AQHNZDR5Qw8fbBSHbEKqTXMfnaVbIJcvbxOA///FyECAAbwiAP//vgVA Date: Thu, 19 Jul 2012 15:23:43 +0000 Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.64.26.13] X-OriginalArrivalTime: 19 Jul 2012 15:23:43.0739 (UTC) FILETIME=[7BB8ECB0:01CD65C2] Henrk, Currently in SMM there are only two types of measurements, grades and dimensional. Dimensional can be used as base measurements for the collective measurements. Your interval measurements do not add, subtract, multiply, divide or accumulate in any simple algebraic manner. That is, an interval measurement cannot be a dimensional measurement available for use as a base measurement. There are even problems when we accumulate regular dimensional measurements (single data point) which have levels of confidence. How do the levels of confidence accumulate? In general, one might say that adding two base measurements with 90% confidence provides a measurement with 81% confidence. But different scenarios might achieve different confidence accumulation. Can we say we don.t define the accumulation of interval measurements or levels of confidence? That is, we extend SMM to include the extra information (in some manner), but we do not imply that the extra information accumulates in a SMM specified manner. There are still some headaches, but we don.t use any fundamental soundness. [I.m swamped today and will look at the attachment and your other notes tomorrow.] Larry From: Henk de Man [mailto:hdman@cordys.com] Sent: Thursday, July 19, 2012 9:31 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins; Henk de Man Subject: Re: issues 17473 - 17474 -- SMM RTF issues Larry, See below. On Wed, Jul 18, 2012 at 9:56 PM, Larry Hines wrote: A measure is an evaluation process that assigns a comparable numeric or symbolic value to an entity in order to characterize a selected quantity or trait of the entity. [hdm] We want to extend this a little bit. It should not only be possible to assign a single numeric or symbolic value, but it should, optionally and additionally, also be possible to 1) specify a confidence, saying e.g. "the measurement value lays, with confidence of 90% between the values x and y". 2) specify the outcome of measurement in terms of a distribution of values (so that we can have 1 measurement object instead of 10.000 in our model..). This is by no means conflicting with what you have, and if these things are optional extensions, they are not in the way for everybody, and they enables VDML to make the use of SMM we need. Guesstimate is one evaluation process for determining small distances. Using a micrometer is a different measure. They characterize the same trait and can have the same unit, but the evaluation processes are not the same. [hdm] I did another attempt to explain the need. See attachment. You seem to be want an AbstractMeasure where the evaluation process is not specified and to which we can associate concrete measures that share characteristics, units and scope (or domain). Even though it would look very much like NamedMeasure, an AbstractMeasure would not have any measurements associated with it. [hdm] No, we fully comply to and apply the core logic of SMM: applying measures, in observations, to create measurements, against the measures. We do not want measures without measurements. We really apply 100 % of SMM, intensively, in VDML. We only want a few refinements or relaxations, if you wish. From: Henk de Man [mailto:hdman@cordys.com] Sent: Wednesday, July 18, 2012 10:30 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins; Henk de Man Subject: Re: issues 17473 - 17474 -- SMM RTF issues Larry, See below. In-lined comments. On Tue, Jul 17, 2012 at 5:54 PM, Larry Hines wrote: The proposal contained in 17473 and 17474 seems rather convoluted. First let me ask what are the semantics of the four observation types? [hdm] The suggestion was not about four types, but just a string to allow any type. The four types were just examples. So, just something as simple as a string to enter the type of observation. What is an estimated observation [hdm] An observation which measurements have values that are just business analyst "guesstimates", or "values of planned or targeted performance". Note that we talk about measurements of characteristics of elements of business models / business systems. VDML applies SMM to the area of business design. , an actual observation [hdm] When parts of the business design have been implemented in real world, measurements can be created in the real world (actual performance), and these might be fed back to the model, per separate observation on the same thing. , a simulated observation [hdm] Measures executed during a simulation game. They can be imported back into the model, per separate observation. and a benchmark observation? [hdm] E.g. measurements that express industry benchmarks for similar situations, stored in the model per separate observation. I can imagine definitions for .actual. and even for .benchmark.. Actual would mean that the measure was actually applied. Benchmark would mean that the measure was actually applied for the purposes of obtaining a benchmark. Following this train of thought, I would guess that simulated is that the measure was simulated and not actually applied. Perhaps estimated means that the measure was estimated. This is why the examples you gave appear to me to be categories of measures. [hdm] This is not about categories of measures. It is about measurements (to outcomes of measuring), stored per observation, but organized in a way that we can distinguish them (so per separate observation). Note that VDML integrates with SMM Observations. Btw: the VDML-SMM integration is quite significant. If asked about types of observations generally I would assert that observations could be typed as casual, natural, subjective, objective, direct, indirectly, controlled, uncontrolled, intrusive, nonintrusive, etc. [hdm] All good, but let's not try to fix a list in an enum, but rather something simple as a string (which was Alain Picards proposal). The same string can then be used for the "four" example types that I gave. Something that seems obvious to me is that the measure used in a subjective observation cannot be the same measure as one used in an objective measure. [hdm] The measure is one, and a re-usable one in a library. The name is given, the characteristic (or trait) is given, the unit is given, and if defined, the operation or formula or functor or accumulator is given, etc. But different observations result in different measurements against the same measure. Here we hit on a triggering dilemma: In our business modeling context, there are very many dependencies between measures (thru aggregation thru binary or collective measures and/or thru re-scaling). So, if manual entry of guesstimates, querying of actuals, etc. would all require different measures, than this would lead to redundancy of almost complete libraries. Not just of direct measure, but also all aggregates that aggregate from them.. This would not be maintainable. What would be required is this: A direct measure has e.g. a query defined, which can be used for querying of actual performance (e.g. from a process instance database). The same measure can however be used, to just enter the value manually, as an estimate. But per different observation. So, in other words: dependent on the observation, the operation is used or not used. This is a special case of what was suggested per the other issue: allowing multiple operations on a direct measure. Note that any aggregated measure (rescaled, binary, collective) does not have this issue, but it just executes is formula or functor or accumulator. This will lead to a much better and maintainable situation: the aggregated measures can aggregate from underlying direct or aggregated measurements, regardless of how the leave measurements were created (estimates, or query result, etc.). Otherwise we would have to duplicate all these aggregated measures also.. ! They may be or may not be assigned the same name, but they are different measures. The two measures would certainly have the same characteristic and the same domain. In 17474 especially you seem to be creating a relationship between measures. Why push the relationship into the definition of DirectMeasure? [hdm] 17474 is about allowing 0..* operations on a direct measure. For libraries of business measures to be maintainable, it is required to allow for this. Above I explained that we want to re-use the same measure for manual entry of guestimate as well as for actuals querying. Similar applies to variations of queries. E.g. exactly the same measure is applied in different parts of the business system. However, in one part of the business a different query service is used than in other part. We cannot duplicate the measure for this, because otherwise all rescaled, binary and collective measures, that are based on them, would have to be duplicated as well, which would kill the library approach, as the re-use would be gone. As a side note, I do think that observation is rather under specified. [hdm] I am all in for building more on that, when we can keep the library pure and better re-usable. From: Henk de Man [mailto:hdman@cordys.com] Sent: Tuesday, July 17, 2012 8:01 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Henk de Man; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins Subject: Re: issues 17472 - 17475 -- SMM RTF issues Larry, See below. On Sat, Jul 14, 2012 at 8:40 PM, Larry Hines wrote: With respect to #17472: These attributes seem better fit for Measure, the method of measurement. A measurement is a single data point obtained by applying a measure. The measure may have precision, confidence bounds and a distribution. [hdm] With respect to #17472: It is not correct to state that a measurement is a single data point. See e.g. Hubbard, Douglas W., How to Measure Anything, Finding the Value of .Intangibles. in Business, John Wiley & Sons, New Jersey, 2010. His main point is, that, in business measurements, just point measures aren't good measures. Often the point measure is not known or is not 100 % trustable. It is very important to be able to state measurement (so the result of measurement) as something that resides in a confidence interval, according to a certain level of confidence. A measurement can be a point value. But it should also be possible to state a measurement as e.g. a confidence interval, with a confidence value. So-far about confidence interval and confidence. Now about distribution (stochastic distribution): A Monte Carlo measurement (i.e. outcome of Multi Carlo simulation or experiment) is typically expressed as a set of values that can be best described by a distribution curve. Stochastic enabling of measurement is important. But the distribution itself can not very well be defined as part of the measure, because the measure is re-usable in different contexts, and the distribution will often be specific to that context. With respect to #17473: Each of these types seem to be categories of measures. Perhaps SMM should include some built-in measure categories such as Estimators, Simulators and Benchmarks. [hdm] With respect to #17473: It is not about the measure, but about the observation. It is about a specification the type of the observation. Note that the same measure can be used for for estimated measurement, as-is measurement, simulated measurement. With respect to #17474: Measures can be equated. If there are multiple, equally good direct measures for a given characteristic then they are be directly associated by the measure equivalence relationship. [hdm] With respect to #17474. This is not about equivalence. It is like method overloading in Java ... When the same direct measure is applied in different contexts, slightly different or additional arguments (parameters) might be required in its operation. So, actually, more than one operation would be required. During discussions with Alain Picard this came out as the best solution. With respect to #17475: I.m good with this. From: Juergen Boldt [mailto:juergen@omg.org] Sent: Friday, July 13, 2012 10:59 AM To: issues@omg.org; smm-rtf@omg.org Subject: issues 17472 - 17475 -- SMM RTF issues This is issue # 17472 From: Henk de Man To: Larry Hines Cc: Juergen Boldt , "issues@omg.org" , "smm-rtf@omg.org" , Alain Picard , Pete Rivett , Arne Berre , "fred.a.cummins" , Henk de Man X-Gm-Message-State: ALoCoQlwog8R2yUZ27if2xlwd4oGZefOLnfIfbUZnmO4GWGjHhsXNytq1xkB+q2RUx5yr1OhAUqy Larry, In this e-mail we relate to two main-discussions: 1) can more flex wrt operation of a direct measure (in particular 0..* operations on a direct measure), which directly relates to a big library maintenance issue (see first attachment again). 2) the second item is about stochastic aspects of measurements, which is crucial in business design. When a busines architect would come to management with the statement that "this" is the point savings figure, he would be blown away immediately.." Good business measures have to say something about how good the measures are. This is a must. I came a lot into your direction on associated measurements, but there's still a pending issue that you might help resolving so that we can go that way. See my discussion of that in the second attachment (that's also in another e-mail). Meanwhile, we might be keen, in next iterations, on which discussion item relates to which issue number. Looking forward to your further responses (on all that need further discussion). But note: we can take our time for this. We will be good if we have consensus reached by the next OMG meeting in Jacksonville. So, let's carefully discuss further until then, hoping that we have consensus on all of them by then. Regards, Henk de Man On Thu, Jul 19, 2012 at 5:23 PM, Larry Hines wrote: Henrk,  Currently in SMM there are only two types of measurements, grades and dimensional. Dimensional can be used as base measurements for the collective measurements. Your interval measurements do not add, subtract, multiply, divide or accumulate in any simple algebraic manner. That is, an interval measurement cannot be a dimensional measurement available for use as a base measurement.  There are even problems when we accumulate regular dimensional measurements (single data point) which have levels of confidence. ÂHow do the levels of confidence accumulate? In general, one might say that adding two base measurements with 90% confidence provides a measurement with 81% confidence. But different scenarios might achieve different confidence accumulation.  Can we say we donât define the accumulation of interval measurements or levels of confidence? That is, we extend SMM to include the extra information (in some manner), but we do not imply that the extra information accumulates in a SMM specified manner. There are still some headaches, but we donât use any fundamental soundness.  [Iâm swamped today and will look at the attachment and your other notes tomorrow.]  Larry  From: Henk de Man [mailto:hdman@cordys.com] Sent: Thursday, July 19, 2012 9:31 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins; Henk de Man Subject: Re: issues 17473 - 17474 -- SMM RTF issues  Larry,  See below. On Wed, Jul 18, 2012 at 9:56 PM, Larry Hines wrote: A measure is an evaluation process that assigns a comparable numeric or symbolic value to an entity in order to characterize a selected quantity or trait of the entity. [hdm] We want to extend this a little bit. It should not only be possible to assign a single numeric or symbolic value, but it should, optionally and additionally, also be possible to 1) specify a confidence, saying e.g. "the measurement value lays, with confidence of 90% between the values x and y". 2) specify the outcome of measurement in terms of a distribution of values (so that we can have 1 measurement object instead of 10.000 in our model..). This is by no means conflicting with what you have, and if these things are optional extensions, they are not in the way for everybody, and they enables VDML to make the use of SMM we need.  Guesstimate is one evaluation process for determining small distances. Using a micrometer is a different measure. ÂÂÂThey characterize the same trait and can have the same unit, but the evaluation processes are not the same. [hdm] I did another attempt to explain the need. See attachment.  You seem to be want an AbstractMeasure where the evaluation process is not specified and to which we can associate concrete measures that share characteristics, units and scope (or domain). Even though it would look very much like NamedMeasure, an AbstractMeasure would not have any measurements associated with it. [hdm] No, we fully comply to and apply the core logic of SMM: applying measures, in observations, to create measurements, against the measures. We do not want measures without measurements. We really apply 100 % of SMM, intensively, in VDML. We only want a few refinements or relaxations, if you wish.   From: Henk de Man [mailto:hdman@cordys.com] Sent: Wednesday, July 18, 2012 10:30 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins; Henk de Man Subject: Re: issues 17473 - 17474 -- SMM RTF issues  Larry,  See below. In-lined comments. On Tue, Jul 17, 2012 at 5:54 PM, Larry Hines wrote: The proposal contained in 17473 and 17474 seems rather convoluted.  First let me ask what are the semantics of the four observation types? [hdm] The suggestion was not about four types, but just a string to allow any type. The four types were just examples. So, just something as simple as a string to enter the type of observation.  ÂÂWhat is an estimated observation [hdm] An observation which measurements have values that are just business analyst "guesstimates", or "values of planned or targeted performance". Note that we talk about measurements of characteristics of elements of business models / business systems. VDML applies SMM to the area of business design. , an actual observation [hdm] When parts of the business design have been implemented in real world, measurements can be created in the real world (actual performance), and these might be fed back to the model, per separate observation on the same thing.  , a simulated observation [hdm] Measures executed during a simulation game. They can be imported back into the model, per separate observation.  and a benchmark observation? [hdm] E.g. measurements that express industry benchmarks for similar situations, stored in the model per separate observation. ÂI can imagine definitions for âactualâ and even for âbenchmarkâ. Actual would mean that the measure was actually applied. Benchmark would mean that the measure was actually applied for the purposes of obtaining a benchmark. Following this train of thought, I would guess that simulated is that the measure was simulated and not actually applied. Perhaps estimated means that the measure was estimated. ÂThis is why the examples you gave appear to me to be categories of measures. [hdm] This is not about categories of measures. It is about measurements (to outcomes of measuring), Âstored per observation, but organized in a way that we can distinguish them (so per separate observation). Note that VDML integrates with SMM Observations. Btw: the VDML-SMM integration is quite significant.  If asked about types of observations generally I would assert that observations could be typed as casual, natural, subjective, objective, direct, indirectly, controlled, uncontrolled, intrusive, nonintrusive, etc.  [hdm] All good, but let's not try to fix a list in an enum, but rather something simple as a string (which was Alain Picards proposal). The same string can then be used for the "four" example types that I gave.  Something that seems obvious to me is that the measure used in a subjective observation cannot be the same measure as one used in an objective measure.  [hdm] The measure is one, and a re-usable one in a library. The name is given, the characteristic (or trait) is given, the unit is given, and if defined, the operation or formula or functor or accumulator is given, etc. But different observations result in different measurements against the same measure. Here we hit on a triggering dilemma: In our business modeling context, there are very many dependencies between measures (thru aggregation thru binary or collective measures and/or thru re-scaling). So, if manual entry of guesstimates, querying of actuals, etc. would all require different measures, than this would lead to redundancy of almost complete libraries. Not just of direct measure, but also all aggregates that aggregate from them.. This would not be maintainable. What would be required is this: A direct measure has e.g. a query defined, which can be used for querying of actual performance (e.g. from a process instance database). The same measure can however be used, to just enter the value manually, as an estimate. But per different observation. So, in other words: dependent on the observation, the operation is used or not used. This is a special case of what was suggested per the other issue: allowing multiple operations on a direct measure. Note that any aggregated measure (rescaled, binary, collective) does not have this issue, but it just executes is formula or functor or accumulator. This will lead to a much better and maintainable situation: the aggregated measures can aggregate from underlying direct or aggregated measurements, regardless of how the leave measurements were created (estimates, or query result, etc.). Otherwise we would have to duplicate all these aggregated measures also.. !  ÂThey may be or may not be assigned the same name, but they are different measures. The two measures would certainly have the same characteristic and the same domain.  In 17474 especially you seem to be creating a relationship between measures. Why push the relationship into the definition of DirectMeasure? [hdm] 17474 is about allowing 0..* operations on a direct measure. For libraries of business measures to be maintainable, it is required to allow for this. Above I explained that we want to re-use the same measure for manual entry of guestimate as well as for actuals querying. Similar applies to variations of queries. E.g. exactly the same measure is applied in different parts of the business system. However, in one part of the business a different query service is used than in other part. We cannot duplicate the measure for this, because otherwise all rescaled, binary and collective measures, that are based on them, would have to be duplicated as well, which would kill the library approach, as the re-use would be gone.  As a side note, I do think that observation is rather under specified. [hdm] I am all in for building more on that, when we can keep the library pure and better re-usable.   From: Henk de Man [mailto:hdman@cordys.com] Sent: Tuesday, July 17, 2012 8:01 AM To: Larry Hines Cc: Juergen Boldt; issues@omg.org; smm-rtf@omg.org; Henk de Man; Alain Picard; Pete Rivett; Arne Berre; fred.a.cummins Subject: Re: issues 17472 - 17475 -- SMM RTF issues  Larry,  See below. On Sat, Jul 14, 2012 at 8:40 PM, Larry Hines wrote: With respect to #17472: These attributes seem better fit for Measure, the method of measurement. A measurement is a single data point obtained by applying a measure. The measure may have precision, confidence bounds and a distribution. [hdm] ÂWith respect to #17472: It is not correct to state that a measurement is a single data point. See e.g.ÂHubbard, Douglas W., How to Measure Anything, Finding the Value of âIntangiblesâ in Business, John Wiley & Sons, New Jersey, 2010. His main point is, that, in business measurements, just point measures aren't good measures. Often the point measure is not known or is not 100 % trustable. It is very important to be able to state measurement (so the result of measurement) as something that resides in a confidence interval, according to a certain level of confidence. ÂA measurement can be a point value. But it should also be possible to state a measurement as e.g. a confidence interval, with a confidence value. So-far about confidence interval and confidence. Now about distribution (stochastic distribution): A Monte Carlo measurement (i.e. outcome of Multi Carlo simulation or experiment) is typically expressed as a set of values that can be best described by a distribution curve. Stochastic enabling of measurement is important. But the distribution itself can not very well be defined as part of the measure, because the measure is re-usable in different contexts, and the distribution will often be specific to that context.   With respect to #17473: Each of these types seem to be categories of measures. Perhaps SMM should include some built-in measure categories such as Estimators, Simulators and Benchmarks. [hdm] ÂWith respect to #17473: It is not about the measure, but about the observation. It is about a specification the type of the observation. Note that the same measure can be used for for estimated measurement, as-is measurement, simulated measurement.  With respect to #17474: Measures can be equated. If there are multiple, equally good direct measures for a given characteristic then they are be directly associated by the measure equivalence relationship. [hdm]ÂWith respect to #17474.ÂThis is not about equivalence. ÂIt is like method overloading in Java ... When the same direct measure is applied in different contexts, slightly different or additional arguments (parameters) might be required in its operation. So, actually, more than one operation would be required. During discussions with Alain Picard this came out as the best solution.  With respect to #17475: Iâm good with this.   From: Juergen Boldt [mailto:juergen@omg.org] Sent: Friday, July 13, 2012 10:59 AM To: issues@omg.org; smm-rtf@omg.org Subject: issues 17472 - 17475 -- SMM RTF issues  This is issue # 17472ÂÂÂFrom: Henk de Man To: Henk de Man CC: Juergen Boldt , "issues@omg.org" , "smm-rtf@omg.org" , Alain Picard , Pete Rivett , Arne Berre , "fred.a.cummins" Subject: Re: 17474 -- SMM RTF issues -- Multiple operation for Direct Measure Thread-Topic: Re: 17474 -- SMM RTF issues -- Multiple operation for Direct Measure Thread-Index: Ac18spgnI5NVfAt+QpSds2nvxFL46A== Date: Fri, 17 Aug 2012 19:57:57 +0000 Accept-Language: en-US X-MS-Has-Attach: X-MS-TNEF-Correlator: x-originating-ip: [10.64.26.6] It seems that we have a resolution for 17474. [hdm] After more discussion with Larry Hines (ref also the discussion that we had on informal libraries versus structured libraries, where e.g. n direct measures in Lib 2 might refer back to 1 named measure in Lib1), we think we can live with n direct measures on one characteristic, rather than n operations on 1 direct measure. However, Larry Hine.s suggestion was to use annotation as the reference. But that is not unique. The system cannot distinguish such a reference from any other custom annotation, so we would rather need to add a property to measure, e.g. .cross reference., or .source. (if generic enough, this property might even be moved to vdml element). [lhm] Adding .cross reference., or .source. or .external name. would be fine with me. I would, however, move the attribute up to Measure. Larry Hines, PhD Software Systems Developer, Sr. Principal Micro Focus larry.hines@microfocus.com 8310 Capital of Texas Highway, Suite 100 Austin, Texas, 78731, USA Telephone : 512-340-4740 This message has been scanned by MailController.