Issue 18814: Unbounded Lists In Device Definitions (gems-rtf) Source: Amergint Technologies (Mr. Rob Andzik, andzik(at)amergint.com) Nature: Uncategorized Issue Severity: Summary: It is not clear how to define the multiplicity of a list of Parameters or ParameterSets when the list is unbounded or can vary in length. For example, if the multiplicity is set to 5, it is assumed this means the list must contain 5 parameters. How does one specify "up to 5" or "unbounded"? Resolution: Revised Text: Actions taken: July 15, 2013: received issue Discussion: End of Annotations:===== ogle-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=mime-version:from:date:message-id:subject:to:content-type :x-gm-message-state; bh=nFy5C7uYMnRHTorsPYg0CEC54NO0zWVRtbD8DQAFWAw=; b=OYx/GCL1vaHEQJNl8jKkJNxP7KY+C2cDLBbJ3TAh6BHeaHaSUBwZxXh4P2lcZT1IXf oOg+3JSgDez/FvicrkImRYxkY4qVLsIDPXwIX/JgIcDmYdeyeINiEn+qFPTbgyR/KqTq dMXfQ1m8yAjerFl/ooDzoy7uLZpjTkSoSYGliJBy664hB4sYtYtmdxL0Bfmixty54sH+ Ct5yjupQ4rONNQ2U6LwAZ+7uVAYmq7qIWdbx9DC8LF1vyZQfX4PPOZlp/tnIDj1tyPnT JWtsMvvpqERCG7VF0DwkLD9yoTD+iQhseBSemCncDnVmXK/Qk09oBIiNw3kXx3Q86XQQ vzvA== X-Received: by 10.152.18.202 with SMTP id y10mr24532994lad.80.1373904389064; Mon, 15 Jul 2013 09:06:29 -0700 (PDT) From: Rob Andzik Date: Mon, 15 Jul 2013 10:06:09 -0600 Subject: GEMS Issues To: Juergen X-Gm-Message-State: ALoCoQmVuO5YGCt36s8ToKj9MnhEJYEUegjJLLjObUph9NmuX7+3w43u0jehxjKM680XYMBcsXUu X-Virus-Scanned: amavisd-new at omg.org Juergen, We have a couple more issues I want to capture and address in 1.3 (all minor) related to the GEMS Device Definition Files. This is issue # 18814 2) Unbounded Lists In Device Definitions It is not clear how to define the multiplicity of a list of Parameters or ParameterSets when the list is unbounded or can vary in length. For example, if the multiplicity is set to 5, it is assumed this means the list must contain 5 parameters. How does one specify "up to 5" or "unbounded"? ============================================================================= Thanks, -- Rob Rob Andzik 719-522-2813 AMERGINT Technologies www.amergint.com X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=mime-version:from:date:message-id:subject:to:content-type :x-gm-message-state; bh=nirUC3/r2AvbSpLjX169yBiWHUNhrOarkDUU7oe8IrE=; b=eeBDgjiW3grJ02lrg4f0Vp91IzWpdgH9m8VQvp2/YzD3DkN53qUOTLBCkxRY1Wp1+N 7VwmZja/FjPvZ8oifsdtv2yrz17hCGBV9zTD+z1J8PVA8A5hBWupwIV4vszVFa2+S2pd DzF68Qxfi0EMRb98jyi66Y14qFpJqymOtmyit34It+1XxS96IIuMuZh+NmwEY5zpq7Fu VV0qzsjzZgA8CZotw7TvaNluG17StKuFDOLMvYahLuq2/BxSnpyZ8G9bmH77m7FlPZyX XfuNA7spJCUJuSwXTrqrySAP5K5mE7fkmT17R5y90lRB/uKjvNgDT8RcNleBA3WNL62s BE/g== X-Received: by 10.152.87.101 with SMTP id w5mr7987768laz.48.1374268526395; Fri, 19 Jul 2013 14:15:26 -0700 (PDT) From: Rob Andzik Date: Fri, 19 Jul 2013 15:15:06 -0600 Subject: Issue 18814: Unbounded Lists In Device Definitions (DISCUSS) To: "gems-rtf@omg.org" X-Gm-Message-State: ALoCoQmrzzE9HLr7eYn1UCsbOZTrwXhoSFoyO3qgP8TM0qNLqMFNEEUoak/S66OsZj6cWlgrbrQT X-Virus-Scanned: amavisd-new at omg.org Issue 18814: Unbounded Lists In Device Definitions Initial recommendation is to state that a value of -1 implies unbounded list length. However that is not sufficient. How should we define the length of a parameter list? In some cases it is fixed length (supported) in other cases the length may vary. When it varies, it can still have a limit (e.g. no more than 5). X-Env-Sender: nthompson@rtlogic.com X-Msg-Ref: server-15.tower-220.messagelabs.com!1374501770!3257752!1 X-Originating-IP: [207.109.164.59] X-StarScan-Received: X-StarScan-Version: 6.9.9; banners=-,-,- X-VirusChecked: Checked From: "Thompson, Nigel" To: "andzik@amergint.com" , "gems-rtf@omg.org" Date: Mon, 22 Jul 2013 08:02:48 -0600 Subject: RE: Issue 18814: Unbounded Lists In Device Definitions (DISCUSS) Thread-Topic: Issue 18814: Unbounded Lists In Device Definitions (DISCUSS) Thread-Index: Ac6ExR2J3FFYLVkhQEi30tVwmYkLqwCHuNvw Accept-Language: en-US X-MS-Has-Attach: yes X-MS-TNEF-Correlator: acceptlanguage: en-US X-Mlf-Version: 7.4.2.7667 X-Mlf-UniqueId: o201307221402500039935 X-Virus-Scanned: amavisd-new at omg.org Why does the length need to be defined at all? One can easily enumerate child nodes and determine the number of sub-elements. I wondered why .multiplicity. is required for ParameterSet for the same reason. Nigel From: Rob Andzik [mailto:andzik@amergint.com] Sent: Friday, July 19, 2013 3:15 PM To: gems-rtf@omg.org Subject: Issue 18814: Unbounded Lists In Device Definitions (DISCUSS) Issue 18814: Unbounded Lists In Device Definitions Initial recommendation is to state that a value of -1 implies unbounded list length. However that is not sufficient. How should we define the length of a parameter list? In some cases it is fixed length (supported) in other cases the length may vary. When it varies, it can still have a limit (e.g. no more than 5). smime8.p7s smime8.p7s X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=references:from:mime-version:in-reply-to:date:message-id:subject:to :cc:content-type:x-gm-message-state; bh=52FPc9fyriWmvHOFflsOuKWkyCs0jYT+nS1Dijd9xr0=; b=jUbcmTzmPzT+7JeuBzsfvd9qaL43duVHgfcOvKo5beOhcHlni/KqCfB1d2kCRUHt1r RFWDUaOucLCW9iUuyTKvU8Y9mrWB133gFrW0QnluP1p2bEkuhU2pDNaNbt8FRwxtPrP0 XiAKY/CMXm/YW6f2Q25/UgDNNvUnvcrFyLlk80vTXIOVEmyCFSr73ZKB2aC8WfmS1YpW OkgUUQw6b023gfKpJvTM3OFTZPN4xI3pJ6Ab2gAuhibvyaHZeH4WaNNRlE5psRnwgqPR EpWePOquwBxbBV9EIDpWFh44mSMlck5j1mdDccEmsV12Raqej+m8ZWa82tX3A8ajIvw2 GPAQ== X-Received: by 10.152.87.15 with SMTP id t15mr12294207laz.85.1374505219164; Mon, 22 Jul 2013 08:00:19 -0700 (PDT) From: Rob Andzik Date: Mon, 22 Jul 2013 09:00:12 -0600 Subject: Re: Issue 18814: Unbounded Lists In Device Definitions (DISCUSS) To: "Thompson, Nigel" Cc: "gems-rtf@omg.org" X-Gm-Message-State: ALoCoQnLBjKU+/xjx/HInbV70ZhyECeeLcHss6R/9Lmn9kvkLkWo1j1wvqjZqyRrkAAMXR1rSGsf X-Virus-Scanned: amavisd-new at omg.org That's the question - do we need to define the length in a fixed field or can we let the field length vary? .When we first defined the spec we were trying to keep parsing as simple as possible. .Basically read the first N bytes and then based on the length field read more bytes. If we do not have a fixed length header than its a little harder, but only just a little. You also need to make a choice when to determine you got a bad message (eg after reading 20 bytes without a pipe character). Having a limit eliminates this question. . As for the multiplicity, the main reason is to specify that it is a list of ParameterSets (ie a list of structs) and possibly define the length of the list. . Rob Rob Andzik (mobile) On Jul 22, 2013, at 8:02 AM, "Thompson, Nigel" wrote: Why does the length need to be defined at all? One can easily enumerate child nodes and determine the number of sub-elements. I wondered why .multiplicity. is required for ParameterSet for the same reason. . Nigel . . From: Rob Andzik [mailto:andzik@amergint.com] Sent: Friday, July 19, 2013 3:15 PM To: gems-rtf@omg.org Subject: Issue 18814: Unbounded Lists In Device Definitions (DISCUSS) . Issue 18814: Unbounded Lists In Device Definitions . Initial recommendation is to state that a value of -1 implies unbounded list length. .However that is not sufficient. . How should we define the length of a parameter list? In some cases it is fixed length (supported) in other cases the length may vary. When it varies, it can still have a limit (e.g. no more than 5).. . . X-Google-DKIM-Signature: v=1; a=rsa-sha256; c=relaxed/relaxed; d=google.com; s=20120113; h=mime-version:in-reply-to:references:from:date:message-id:subject:to :cc:content-type:x-gm-message-state; bh=RsHjdwUFqDqiSoCuvzaqpdGamXDurqq9YIKgBMPAwHc=; b=UtzOmr0V1HmiTQv5XGF9p5dbg/szY42RiAJSgNd6KFWfxQ+oLoKIHwgXttjVMZJrY/ 8up7AGVkuNgnTj4ahI3khspqxA3LISNW/Qwulb8A7LDYozIM7uPzjTGj+vmTI51fO6gR opkB0hOrGzrR0wTnhHtpuGZVIXYHXiKlECsWXc/bFSqkomwfgo7jk94iNOdlF3sgDrha WXhVRoKOzxbJlEPvpcS+x9QlCkKBfXSZgkBhKkc2eUoxhl1hfTph6AR8HQHFlg1fO91x rDFps6ce8isCfE6p3ZhQvy4HX79AODB1pRHmR7z6eF3314fjbm75UrxqhQgfAq1u668o Pm2g== X-Received: by 10.112.11.50 with SMTP id n18mr12279268lbb.57.1374507211393; Mon, 22 Jul 2013 08:33:31 -0700 (PDT) From: Rob Andzik Date: Mon, 22 Jul 2013 09:33:11 -0600 Subject: Re: Issue 18814: Unbounded Lists In Device Definitions (DISCUSS) To: "Thompson, Nigel" Cc: "gems-rtf@omg.org" X-Gm-Message-State: ALoCoQlZataP/4uaIKErsTE6Gn1SX+c97L9hXzzD2wZm0m09B45qk+qAoNSl10xsuX733gb4rN1z X-Virus-Scanned: amavisd-new at omg.org Sorry -- I was answering this on my phone and had not seen your response to issue 18314. I was confused. Let me try again. Keep in mind this issue concerns the Device Definition File not the actual messages. ParameterSets are just a different type of Parameter. So it is possible for you to have a list of ParameterSets. This is already supported in both GEMS-ASCII and GEMS-XML. What we didn't do is give you a way to define such a list in the Device Definition file. Its missing the multiplicity attribute. As for the length of the list, we need some way to define that a list contains a specific number of entries. In the actual message it just works out. For example, the GetConfigMessage might contain only 5 entries in the list. A SetConfigMessage might return an error if you specify anything other than the expected 5 entries. In this example, I would hope that some external documentation describes that exactly 5 entries are expected. However, how do you specify that in the Device Definition File? -- Rob Rob Andzik 719-522-2813 AMERGINT Technologies www.amergint.com On Mon, Jul 22, 2013 at 9:00 AM, Rob Andzik wrote: That's the question - do we need to define the length in a fixed field or can we let the field length vary? When we first defined the spec we were trying to keep parsing as simple as possible. Basically read the first N bytes and then based on the length field read more bytes. If we do not have a fixed length header than its a little harder, but only just a little. You also need to make a choice when to determine you got a bad message (eg after reading 20 bytes without a pipe character). Having a limit eliminates this question. As for the multiplicity, the main reason is to specify that it is a list of ParameterSets (ie a list of structs) and possibly define the length of the list. Rob Rob Andzik (mobile) On Jul 22, 2013, at 8:02 AM, "Thompson, Nigel" wrote: Why does the length need to be defined at all? One can easily enumerate child nodes and determine the number of sub-elements. I wondered why .multiplicity. is required for ParameterSet for the same reason. Nigel From: Rob Andzik [mailto:andzik@amergint.com] Sent: Friday, July 19, 2013 3:15 PM To: gems-rtf@omg.org Subject: Issue 18814: Unbounded Lists In Device Definitions (DISCUSS) Issue 18814: Unbounded Lists In Device Definitions Initial recommendation is to state that a value of -1 implies unbounded list length. However that is not sufficient. How should we define the length of a parameter list? In some cases it is fixed length (supported) in other cases the length may vary. When it varies, it can still have a limit (e.g. no more than 5).