Folks, Attached are responses I received to my call of comment on the question of renumbering clauses in the VHDL LRM to conform with IEEE style rules. Given the number of people on the lists to which I sent the call, I would not characterize the response as strongly in favor or against. (1 for, 2 against, 2 informative.) I suspect it may be hard to mount a case against renumbering given this level of objection. Regards, PA -- Dr. Peter J. Ashenden peter@ashenden.com.au Ashenden Designs Pty. Ltd. www.ashenden.com.au PO Box 640 VoIP: sip://0871270078@sip.internode.on.net Stirling, SA 5152 Phone: +61 8 7127 0078 Australia Mobile: +61 414 70 9106 > -----Original Message----- > From: owner-vhdl-200x@server.eda.org > [mailto:owner-vhdl-200x@server.eda.org] On Behalf Of Peter Ashenden > Sent: Wednesday, 18 July 2007 15:06 PM > To: vhdl-200x@server.eda.org; 'Accellera VHDL TC'; > vhdl-lrm@lists.accellera.org > Subject: [vhdl-200x] Request for comment: renumbering VHDL LRM clauses > > > Dear colleagues, > > This is a request for comment on a proposal to renumber > clauses of the VHDL LRM to meet current IEEE style guidelines. > > The IEEE rules have evolved during the lifetime of the VHDL > standard, and the standard is now not conformant. IEEE would > like us to revise the document to conform with the current > style rules. That would result in significant changes to the > numbering of clauses and subclauses in the document. (A more > detailed description is below.) As a consequence, and > implementation that referred to LRM clause or subclause > numbers (eg, in error messages), would have to be revised, > and would have to maintain different references for different > versions of the language. > > The VHDL working group would appreciate comments from users > and implementers of the standard to determine the extent of > the effect of renumbering. If there is significant objection > to renumbering, the working group could make representations > to IEEE to maintain the current numbering. If there is not > significant objection, then the technical editor could > proceed to implement the renumbering. > > If you would like to comment, please reply to this message > (NOT reply-all), and I will collate responses. Please reply > by 5pm Friday 27-Jul US-PDT. Thanks. > > > Some further detail: > > The current numbering of LRM clauses and subclauses has been > maintained largely unchanged since the original VHDL-87 > standard. Explanatory information was added in the VHDL-93 > standard as Clause 0, in order not to change the numbering of > subsequent clauses. Additional subclauses were added in > VHDL-93 and -2000, but at the ends of clauses, so as not to > perturb exiting numbers. In VHDL-2006/D3.0, some subclauses > were moved and others inserted, but consequent renumbering is > minimal, especially given the scope of extensions. > > Since the earlier version of the LRM were published, IEEE has > evolved its style guidelines covering what goes in a standard > and where. Some or the guidelines stem from requirements on > ANSI and international standards. In particular, IEEE > standards are now required to have: > > - an Overview as Clause 1 containing the Scope and Purpose, > and NOT including > detailed discussions of the general technical content of > the standard > - if normative references are used, they constitute Clause > - if definitions are required, they constitute Clause 3 > > Implementing these changes would result in Clause numbers changing. > > The rules also require certain structure of clauses and > subclauses. If a clause contains subclauses, then there must > not be text preceeding the first subclause; that text should > itself be in a subclause. Subclauses must be similarly > structured. These rules avoid a reference to, say, Clause 5, > being ambiguous (interpretable as just referring to the > introductory text or referring to the clause and all its > subclauses). Application of these rules would result in > changes to subclause numbering within clauses. > > Furthermore, each clause or subclause can contain at most one > numbered list. If more are required, the clause must be > subdivided with each list in a separate subclause. This > ensures that a list item can be unambiguously cross > referenced by prefixing it with its containing subclause number. > > Regards, > > PA > > -- > Dr. Peter J. Ashenden peter@ashenden.com.au > Ashenden Designs Pty. Ltd. www.ashenden.com.au > PO Box 640 VoIP: > sip://0871270078@sip.internode.on.net > Stirling, SA 5152 Phone: +61 8 7127 0078 > Australia Mobile: +61 414 70 9106 > > > -- > This message has been scanned for viruses and > dangerous content by MailScanner, and is > believed to be clean. > > -- This message has been scanned for viruses and dangerous content by MailScanner, and is believed to be clean.
attached mail follows:
Peter, I understand your concerns. We had similar problems in the Verilog 1364 and 1800 standards, though 1364 goes back only to 1995, not to 1987. We did renumber clauses and subclauses, but also did some actions to reduce the scope of the problem. Maybe some of this can help you. First, sometimes it is possible to get IEEE permission for exceptions to some (not all) of the requirements, especially if it is based on a previous existing version. If you look through the standards guidelines, sometimes the possibility of exceptions is mentioned. Sometimes you can get an exception even if it is not explicitly mentioned. When we went from 1364-2001 to 1364-2005, we did have to add Clause 2 Normative References. We avoided Clause 3 as being optional. In order to minimize the impact of the clause renumbering by adding Clause 2, we moved Clause 5 to be Clause 11, and merged two others, such that Clauses 12+ remained with the same number, and so did 6-10, I think. Currently, we are merging 1364 and 1800 into a single new standard which will also be called 1800. Here there is no choice and major renumbering is necessary, but we have an Annex which has a high-level mapping of 1364 and 1800 clauses and sub-clauses into the new document. I hope IEEE will let us keep this in the final document. Hope this helps. Regards, Shalom > -----Original Message----- > From: Peter Ashenden [mailto:peter@ashenden.com.au] > Sent: Wednesday, July 18, 2007 8:36 AM > To: vhdl-200x@server.eda.org; 'Accellera VHDL TC'; > vhdl-lrm@lists.accellera.org > Subject: [Accellera:vhdl] Request for comment: renumbering > VHDL LRM clauses > > Dear colleagues, > > This is a request for comment on a proposal to renumber > clauses of the VHDL LRM to meet current IEEE style guidelines. > > The IEEE rules have evolved during the lifetime of the VHDL > standard, and the standard is now not conformant. IEEE would > like us to revise the document to conform with the current > style rules. That would result in significant changes to the > numbering of clauses and subclauses in the document. (A more > detailed description is below.) As a consequence, and > implementation that referred to LRM clause or subclause > numbers (eg, in error messages), would have to be revised, > and would have to maintain different references for different > versions of the language. > > The VHDL working group would appreciate comments from users > and implementers of the standard to determine the extent of > the effect of renumbering. If there is significant objection > to renumbering, the working group could make representations > to IEEE to maintain the current numbering. If there is not > significant objection, then the technical editor could > proceed to implement the renumbering. > > If you would like to comment, please reply to this message > (NOT reply-all), and I will collate responses. Please reply > by 5pm Friday 27-Jul US-PDT. > Thanks. > > > Some further detail: > > The current numbering of LRM clauses and subclauses has been > maintained largely unchanged since the original VHDL-87 > standard. Explanatory information was added in the VHDL-93 > standard as Clause 0, in order not to change the numbering of > subsequent clauses. Additional subclauses were added in > VHDL-93 and -2000, but at the ends of clauses, so as not to > perturb exiting numbers. In VHDL-2006/D3.0, some subclauses > were moved and others inserted, but consequent renumbering is > minimal, especially given the scope of extensions. > > Since the earlier version of the LRM were published, IEEE has > evolved its style guidelines covering what goes in a standard > and where. Some or the guidelines stem from requirements on > ANSI and international standards. In particular, IEEE > standards are now required to have: > > - an Overview as Clause 1 containing the Scope and Purpose, > and NOT including > detailed discussions of the general technical content of > the standard > - if normative references are used, they constitute Clause > - if definitions are required, they constitute Clause 3 > > Implementing these changes would result in Clause numbers changing. > > The rules also require certain structure of clauses and > subclauses. If a clause contains subclauses, then there must > not be text preceeding the first subclause; that text should > itself be in a subclause. Subclauses must be similarly > structured. These rules avoid a reference to, say, Clause 5, > being ambiguous (interpretable as just referring to the > introductory text or referring to the clause and all its > subclauses). Application of these rules would result in > changes to subclause numbering within clauses. > > Furthermore, each clause or subclause can contain at most one > numbered list. > If more are required, the clause must be subdivided with each > list in a separate subclause. This ensures that a list item > can be unambiguously cross referenced by prefixing it with > its containing subclause number. > > Regards, > > PA > > -- > Dr. Peter J. Ashenden peter@ashenden.com.au > Ashenden Designs Pty. Ltd. www.ashenden.com.au > PO Box 640 VoIP: > sip://0871270078@sip.internode.on.net > Stirling, SA 5152 Phone: +61 8 7127 0078 > Australia Mobile: +61 414 70 9106 >
attached mail follows:
Peter, If the new IEEE numbering is adopted, the old numbers should be included for cross referencing. Jake Karrfalt ASC Peter Ashenden wrote: >Dear colleagues, > >This is a request for comment on a proposal to renumber clauses of the VHDL >LRM to meet current IEEE style guidelines. > >The IEEE rules have evolved during the lifetime of the VHDL standard, and >the standard is now not conformant. IEEE would like us to revise the >document to conform with the current style rules. That would result in >significant changes to the numbering of clauses and subclauses in the >document. (A more detailed description is below.) As a consequence, and >implementation that referred to LRM clause or subclause numbers (eg, in >error messages), would have to be revised, and would have to maintain >different references for different versions of the language. > >The VHDL working group would appreciate comments from users and implementers >of the standard to determine the extent of the effect of renumbering. If >there is significant objection to renumbering, the working group could make >representations to IEEE to maintain the current numbering. If there is not >significant objection, then the technical editor could proceed to implement >the renumbering. > >If you would like to comment, please reply to this message (NOT reply-all), >and I will collate responses. Please reply by 5pm Friday 27-Jul US-PDT. >Thanks. > > >Some further detail: > >The current numbering of LRM clauses and subclauses has been maintained >largely unchanged since the original VHDL-87 standard. Explanatory >information was added in the VHDL-93 standard as Clause 0, in order not to >change the numbering of subsequent clauses. Additional subclauses were added >in VHDL-93 and -2000, but at the ends of clauses, so as not to perturb >exiting numbers. In VHDL-2006/D3.0, some subclauses were moved and others >inserted, but consequent renumbering is minimal, especially given the scope >of extensions. > >Since the earlier version of the LRM were published, IEEE has evolved its >style guidelines covering what goes in a standard and where. Some or the >guidelines stem from requirements on ANSI and international standards. In >particular, IEEE standards are now required to have: > >- an Overview as Clause 1 containing the Scope and Purpose, and NOT >including > detailed discussions of the general technical content of the standard >- if normative references are used, they constitute Clause >- if definitions are required, they constitute Clause 3 > >Implementing these changes would result in Clause numbers changing. > >The rules also require certain structure of clauses and subclauses. If a >clause contains subclauses, then there must not be text preceeding the first >subclause; that text should itself be in a subclause. Subclauses must be >similarly structured. These rules avoid a reference to, say, Clause 5, being >ambiguous (interpretable as just referring to the introductory text or >referring to the clause and all its subclauses). Application of these rules >would result in changes to subclause numbering within clauses. > >Furthermore, each clause or subclause can contain at most one numbered list. >If more are required, the clause must be subdivided with each list in a >separate subclause. This ensures that a list item can be unambiguously cross >referenced by prefixing it with its containing subclause number. > >Regards, > >PA > >-- >Dr. Peter J. Ashenden peter@ashenden.com.au >Ashenden Designs Pty. Ltd. www.ashenden.com.au >PO Box 640 VoIP: sip://0871270078@sip.internode.on.net >Stirling, SA 5152 Phone: +61 8 7127 0078 >Australia Mobile: +61 414 70 9106 > > > -- Jake Karrfalt CEO Alternative System Concepts, Inc. jake@ascinc.com ->> driving design innovation >> (603) 437 2234 office PO Box 128 (603) 437 2722 fax Windham, NH 03087 (603) 321 7150 cell http://www.ascinc.com -
attached mail follows:
Peter, Here's my personal opinion about this. (I am not speaking for Cadence here.) Insisting that each standard be of the same form does not benefit the users of those standards. It only benefits the IEEE, and even then the benefit is minimal: it enables REVCOM and NESCOM reviewers to approve a standard with less actual thought about the matter. The IEEE used to focus on providing a venue for the definition and publication of standards, and constrained the process only enough to ensure fairness in the development of standards and clarity in their documentation. More recently, the IEEE has been flogging standards as a source of income - e.g., the whole focus on corporate-sponsored standards, the pressure on working groups to pay the IEEE for consulting services to get a standard through the process, and the bizarre notion of holding conferences about standards development. Along with this trend, the IEEE seems to have decided to exert much more control over the structure of standards documents, apparently thinking that if standards are important, then standards about standards must also be important. I think the IEEE should go back to its previous approach - enabling standards development and dissemination (and making money in the process through document sales), and let standards developers go back to focusing on content over form. Regards, Erich | -----Original Message----- | From: Peter Ashenden [mailto:peter@ashenden.com.au] | Sent: Wednesday, July 18, 2007 1:36 AM | To: vhdl-200x@eda.org; 'Accellera VHDL TC'; | vhdl-lrm@lists.accellera.org | Subject: [Accellera:vhdl] Request for comment: renumbering | VHDL LRM clauses | | Dear colleagues, | | This is a request for comment on a proposal to renumber | clauses of the VHDL | LRM to meet current IEEE style guidelines. | | The IEEE rules have evolved during the lifetime of the VHDL | standard, and | the standard is now not conformant. IEEE would like us to revise the | document to conform with the current style rules. That would result in | significant changes to the numbering of clauses and subclauses in the | document. (A more detailed description is below.) As a | consequence, and | implementation that referred to LRM clause or subclause | numbers (eg, in | error messages), would have to be revised, and would have to maintain | different references for different versions of the language. | | The VHDL working group would appreciate comments from users | and implementers | of the standard to determine the extent of the effect of | renumbering. If | there is significant objection to renumbering, the working | group could make | representations to IEEE to maintain the current numbering. If | there is not | significant objection, then the technical editor could | proceed to implement | the renumbering. | | If you would like to comment, please reply to this message | (NOT reply-all), | and I will collate responses. Please reply by 5pm Friday | 27-Jul US-PDT. | Thanks. | | | Some further detail: | | The current numbering of LRM clauses and subclauses has been | maintained | largely unchanged since the original VHDL-87 standard. Explanatory | information was added in the VHDL-93 standard as Clause 0, in | order not to | change the numbering of subsequent clauses. Additional | subclauses were added | in VHDL-93 and -2000, but at the ends of clauses, so as not to perturb | exiting numbers. In VHDL-2006/D3.0, some subclauses were | moved and others | inserted, but consequent renumbering is minimal, especially | given the scope | of extensions. | | Since the earlier version of the LRM were published, IEEE has | evolved its | style guidelines covering what goes in a standard and where. | Some or the | guidelines stem from requirements on ANSI and international | standards. In | particular, IEEE standards are now required to have: | | - an Overview as Clause 1 containing the Scope and Purpose, and NOT | including | detailed discussions of the general technical content of | the standard | - if normative references are used, they constitute Clause | - if definitions are required, they constitute Clause 3 | | Implementing these changes would result in Clause numbers changing. | | The rules also require certain structure of clauses and | subclauses. If a | clause contains subclauses, then there must not be text | preceeding the first | subclause; that text should itself be in a subclause. | Subclauses must be | similarly structured. These rules avoid a reference to, say, | Clause 5, being | ambiguous (interpretable as just referring to the introductory text or | referring to the clause and all its subclauses). Application | of these rules | would result in changes to subclause numbering within clauses. | | Furthermore, each clause or subclause can contain at most one | numbered list. | If more are required, the clause must be subdivided with each | list in a | separate subclause. This ensures that a list item can be | unambiguously cross | referenced by prefixing it with its containing subclause number. | | Regards, | | PA | | -- | Dr. Peter J. Ashenden peter@ashenden.com.au | Ashenden Designs Pty. Ltd. www.ashenden.com.au | PO Box 640 VoIP: | sip://0871270078@sip.internode.on.net | Stirling, SA 5152 Phone: +61 8 7127 0078 | Australia Mobile: +61 414 70 9106 | |
attached mail follows:
I think the renumbers is something we just have to do. I don't agree with all the IEEE rules like a clause with subclauses can't contain text, but I can live with them. I also think that since the new number will not be close to the old numbers if we need or want to do some restructioning of clauses this would be the time to do it. -- John Peter Ashenden wrote: > Dear colleagues, > > This is a request for comment on a proposal to renumber clauses of the VHDL > LRM to meet current IEEE style guidelines. > > The IEEE rules have evolved during the lifetime of the VHDL standard, and > the standard is now not conformant. IEEE would like us to revise the > document to conform with the current style rules. That would result in > significant changes to the numbering of clauses and subclauses in the > document. (A more detailed description is below.) As a consequence, and > implementation that referred to LRM clause or subclause numbers (eg, in > error messages), would have to be revised, and would have to maintain > different references for different versions of the language. > > The VHDL working group would appreciate comments from users and implementers > of the standard to determine the extent of the effect of renumbering. If > there is significant objection to renumbering, the working group could make > representations to IEEE to maintain the current numbering. If there is not > significant objection, then the technical editor could proceed to implement > the renumbering. > > If you would like to comment, please reply to this message (NOT reply-all), > and I will collate responses. Please reply by 5pm Friday 27-Jul US-PDT. > Thanks. > > > Some further detail: > > The current numbering of LRM clauses and subclauses has been maintained > largely unchanged since the original VHDL-87 standard. Explanatory > information was added in the VHDL-93 standard as Clause 0, in order not to > change the numbering of subsequent clauses. Additional subclauses were added > in VHDL-93 and -2000, but at the ends of clauses, so as not to perturb > exiting numbers. In VHDL-2006/D3.0, some subclauses were moved and others > inserted, but consequent renumbering is minimal, especially given the scope > of extensions. > > Since the earlier version of the LRM were published, IEEE has evolved its > style guidelines covering what goes in a standard and where. Some or the > guidelines stem from requirements on ANSI and international standards. In > particular, IEEE standards are now required to have: > > - an Overview as Clause 1 containing the Scope and Purpose, and NOT > including > detailed discussions of the general technical content of the standard > - if normative references are used, they constitute Clause > - if definitions are required, they constitute Clause 3 > > Implementing these changes would result in Clause numbers changing. > > The rules also require certain structure of clauses and subclauses. If a > clause contains subclauses, then there must not be text preceeding the first > subclause; that text should itself be in a subclause. Subclauses must be > similarly structured. These rules avoid a reference to, say, Clause 5, being > ambiguous (interpretable as just referring to the introductory text or > referring to the clause and all its subclauses). Application of these rules > would result in changes to subclause numbering within clauses. > > Furthermore, each clause or subclause can contain at most one numbered list. > If more are required, the clause must be subdivided with each list in a > separate subclause. This ensures that a list item can be unambiguously cross > referenced by prefixing it with its containing subclause number. > > Regards, > > PA > > -- > Dr. Peter J. Ashenden peter@ashenden.com.au > Ashenden Designs Pty. Ltd. www.ashenden.com.au > PO Box 640 VoIP: sip://0871270078@sip.internode.on.net > Stirling, SA 5152 Phone: +61 8 7127 0078 > Australia Mobile: +61 414 70 9106 > -- -- mailto: johnr@model.com phone: (503)685-0864 -- http://www.model.com fax: (503)685-0921 --
attached mail follows:
Peter, Sounds like a bad idea at best. It maybe should be published with a new specification number if published in this form at all. The use of a new specification number would make it clear that it is not the same or just an expanded version of a preexisting specification. Change for the sake of change is fraught with headaches. Today it does not matter which version of the LRM I have I can talk with someone else about a particular section using a clause number. this would make the next version of the specification break this model. I would have to have both versions of the specification before me to synchronize. This sounds like a change just for the sake of change. the current document is difficult to navigate at times. Change section numbers because a new standard comes along introduces an additional possibility for errors in the new version. I cannot think of a good reason to make this type of change. Charlie ----- Original Message ----- From: "Peter Ashenden" <peter@ashenden.com.au> To: <vhdl-200x@server.eda.org>; "'Accellera VHDL TC'" <vhdl@lists.accellera.org>; <vhdl-lrm@lists.accellera.org> Sent: Tuesday, July 17, 2007 10:36 PM Subject: [vhdl-200x] Request for comment: renumbering VHDL LRM clauses | Dear colleagues, | | This is a request for comment on a proposal to renumber clauses of the VHDL | LRM to meet current IEEE style guidelines. | | The IEEE rules have evolved during the lifetime of the VHDL standard, and | the standard is now not conformant. IEEE would like us to revise the | document to conform with the current style rules. That would result in | significant changes to the numbering of clauses and subclauses in the | document. (A more detailed description is below.) As a consequence, and | implementation that referred to LRM clause or subclause numbers (eg, in | error messages), would have to be revised, and would have to maintain | different references for different versions of the language. | | The VHDL working group would appreciate comments from users and implementers | of the standard to determine the extent of the effect of renumbering. If | there is significant objection to renumbering, the working group could make | representations to IEEE to maintain the current numbering. If there is not | significant objection, then the technical editor could proceed to implement | the renumbering. | | If you would like to comment, please reply to this message (NOT reply-all), | and I will collate responses. Please reply by 5pm Friday 27-Jul US-PDT. | Thanks. | | | Some further detail: | | The current numbering of LRM clauses and subclauses has been maintained | largely unchanged since the original VHDL-87 standard. Explanatory | information was added in the VHDL-93 standard as Clause 0, in order not to | change the numbering of subsequent clauses. Additional subclauses were added | in VHDL-93 and -2000, but at the ends of clauses, so as not to perturb | exiting numbers. In VHDL-2006/D3.0, some subclauses were moved and others | inserted, but consequent renumbering is minimal, especially given the scope | of extensions. | | Since the earlier version of the LRM were published, IEEE has evolved its | style guidelines covering what goes in a standard and where. Some or the | guidelines stem from requirements on ANSI and international standards. In | particular, IEEE standards are now required to have: | | - an Overview as Clause 1 containing the Scope and Purpose, and NOT | including | detailed discussions of the general technical content of the standard | - if normative references are used, they constitute Clause | - if definitions are required, they constitute Clause 3 | | Implementing these changes would result in Clause numbers changing. | | The rules also require certain structure of clauses and subclauses. If a | clause contains subclauses, then there must not be text preceeding the first | subclause; that text should itself be in a subclause. Subclauses must be | similarly structured. These rules avoid a reference to, say, Clause 5, being | ambiguous (interpretable as just referring to the introductory text or | referring to the clause and all its subclauses). Application of these rules | would result in changes to subclause numbering within clauses. | | Furthermore, each clause or subclause can contain at most one numbered list. | If more are required, the clause must be subdivided with each list in a | separate subclause. This ensures that a list item can be unambiguously cross | referenced by prefixing it with its containing subclause number. | | Regards, | | PA | | -- | Dr. Peter J. Ashenden peter@ashenden.com.au | Ashenden Designs Pty. Ltd. www.ashenden.com.au | PO Box 640 VoIP: sip://0871270078@sip.internode.on.net | Stirling, SA 5152 Phone: +61 8 7127 0078 | Australia Mobile: +61 414 70 9106 | | | -- | This message has been scanned for viruses and | dangerous content by MailScanner, and is | believed to be clean. | |Received on Sun Jul 29 05:09:23 2007
This archive was generated by hypermail 2.1.8 : Sun Jul 29 2007 - 05:11:39 PDT