Re: [oasis-member-discuss] Re: Review of AIR = ASIS, possibly mandatorypolicy?

[Christopher B Ferris <chrisfer@us.ibm.com>]

Comments inlined below.

Generally, a big +1

Cheers,

Christopher Ferris
STSM, Emerging e-business Industry Architecture
email: chrisfer@us.ibm.com
blog: http://www.ibm.com/developerworks/blogs/dw_blog.jspa?blog=440
phone: +1 508 377 9295

"G. Ken Holman" <gkholman@CraneSoftwrights.com>
wrote on 02/23/2006 09:30:43 PM:

> At 2006-02-22 17:01 -0800, jon.bosak@sun.com wrote:
> >UBL TC,
> >
> >Forwarded as requested.
> >
> >Jon
> >------- Start of forwarded message -------
> >Date: Fri, 17 Feb 2006 16:11:17 -0800
> >From: James Bryce Clark <jamie.clark@oasis-open.org>
> >Subject: [chairs] Draft ASIS under review:  mandatory policy?
 Please review
> >  and  comment by 1 March
> >To: chairs@lists.oasis-open.org
> 
> Regarding:
> 
>    http://www.oasis-open.org/committees/download.
> php/16546/ArtifactStandardIdentificationSchemeForMetadata-1.0.1.pdf
> 
> Line 352 - this is an action item for OASIS staff, not an aspect of
> this standard ... the "SHALL" in this jumps out at me as
totally 
> unnecessary.  By the time most people have read this document,
this 
> action item should have long been acted on.  Unless, of course,
I'm 
> misunderstanding the sentence in which case it should be 
> rewritten.  But, if this updating is a one-time action, is it
on the 
> part of those who maintain the The OASIS Document Templates?  When
> will the decisions be incorporated into the templates (note in my
> postscript below that I've been working on the DocBook XML 
> templates)?  A lot of UBL 2.0 is already in development ... will
a 
> change in the templates while writing a new specification mean that
> work in progress would become non-conforming?

I raised roughly the same points in my feedback. There
needs to be a lot
more clarity around transition rules and grandfathering
of existing
work.

> 
> Lines 354-386 - I'm not sure what is meant by "consistent tabular
> form", especially since it looks like the RFC822-styled prompts
and 
> values (which is distinctly not "tabular" (I interpret this
as 
> systematic aligned rows and columns; the columns are not aligned in
> RFC822-styled mail headers).  I don't see guidelines as to how
this 
> "consistent tabular form" is to be rendered in various 
> renditions?  Should it be an HTML table?  If in a PDF table,
with 
> visible cell borders (perhaps, somewhere in the front 
> matter)?  Should it be also in HTML <meta> elements (I
think so; 
> though I don't know how I'll do this in DocBook)?
> 
> Line 408 - I think ECMA-6 is a poor choice for a 128-byte character
> set because of the two ambiguous "alternative graphic character
> allocations" (6.4.2) ... I suppose this is covered because the
> exclusive list is indicated, so I guess this isn't a problem.  Why
> use ECMA and not, say, ISO-646?  Actually, I guess it has the
same 
> issue.  Therefore, why bother citing anything and not just list
the 
> characters ... would this in fact cause problems with a mainframe
> tool that happens to be using EBCDIC?  The fact that it is the
> correct ECMA/ISO characters when it is being mounted in the 
> repository is fine regardless of how it was written, but given that
> this is a *naming* standard and not an *encoding* standard, then just
> listing the characters should be sufficient and any mention of 
> character set is probably unnecessary embellishment.

+1

> 
> Oh, I just found the reference to <meta> in HTML in section
6.4.1 ... 
> though it isn't an exhaustive list of the metadata.  If ASIS
is going 
> to the trouble of listing all of the metadata components, then 
> perhaps these should be mandatory XHTML metadata entries as 
> well.  Actually, not *just* XHTML, but for every rendition an
example 
> is needed of what it is to look like ... otherwise, how would I be
> able to modify the DocBook XML templates in a conformant fashion?

I raised this point as well.

> 
> Line 532 - Absent here are ZIP and TAR/GZ files (though I don't know
> what to do about them) ... I think they should be recognized ... 
> perhaps have a companion ".txt" file (though I hate the
idea of 
> having information in more than one file)?  Though perhaps having
a 
> companion ".txt" file for all binary file types will be
both 
> sufficient and consistent, but line 350 provides a list of required
> metadata elements "The following metadata MUST be associated
with 
> each Artifact...", so it would have to be documented that for
every 
> binary file (or file without a human-readable rendition) that one
can 
> expect to find a .txt file with the meta data.  That could add
a lotof files!
> 
> Line 596 - Shouldn't this be a reference to Section 7.1?
> 
> I think it is going to be difficult to get Joe and Jane 
> StandardsWriter to accurately divine what values for properties in
> these guidelines apply to documents they create for their committee
> ... could the guidelines require each TC to publish somewhere visibly
> in their work pages what the choices are for metadata properties 
> related to all documents for that particular committee?  That
way two 
> people in one TC don't end up making different decisions based on
> their respective interpretations of these guidelines.  I'm learning
> that one really has to spoon feed stuff like this to people who are
> made responsible for creating things ... they won't take the energy
> to have to interpret administrivia and distract them from their 
> technical writing.
> 
> So, I then took a look overall at ASIS and I realized that I really
> couldn't effectively distill what the important guidelines would be
> for, say, the UBL TC or my HISC subcommittee of UBL.
> 
> Consider for example the tree of file with the following directory
> at the apex:
> 
>    http://docs.oasis-open.org/ubl/cd-UBL-1.0/
> 
> ... the artifact in that directory has to be named "index.html"
in 
> order to be properly displayed by the server when the directory is
> referenced.  Does this mean that the directory has the artifact
> name?  Probably, but there are 244 files in that tree.  Are
*all* of 
> them (except the necessary exceptions for the directory "index.html"
> files) to be named with these ASIS guidelines?  If none of them
are 
> used outside of the context of UBL 1.0, why would the artifact naming
> guidelines apply?  I can see them applying to the apex directory
and 
> to any ZIP file that would be used out of context of the 
> directory.  That's it!  The difference is, when is a committee
> artifact found *only* in a given context and when is a committee 
> artifact allowed to live outside of that context?  The apex directory
> can be found in any context, so its name follows the artifact naming
> guidelines.  The ZIP or TAR/GZ packages of the directory can
be found 
> in any context, so its name would also follow the guidelines.  And,
> as above, I suppose also an associated text file with the metadata
> for those compressed packages.
> 
> This would greatly simplify a committee's work.  I had the 
> responsibility for creating the directory at the apex:
> 
>    http://docs.oasis-open.org/ubl/cd-UBL-1.0/fs/
> 
> There are 99 files just in my subtree.  BUT my subtree isn't
ever 
> found outside of the context of the UBL 1.0 deliverable, so none of
> the artifacts in that subtree would be out of context (or should be
> out of context, of course they might be if someone made copies them
> without copying the other files, but when they are in context on the
> UBL web site inside of the UBL 1.0 deliverable, they are 

I have tried, unsuccessfully apparently, to make exactly
this case in the TAB
before I retired my position. I maintained, consistently,
that it was folly
to try to imbue filenames with metadata, that all
we needed to do was to 
use the web as it was intended.

I recommended the use of RDDL, and produced a microformat
for the 
OASIS metadata that could be embedded in the RDDL
document which could
link to the actual artifact.

I also pointed out the folly of trying to force version
numbering
into the file names as it would conflict with use
of a version control
environment such as CVS or SVN, which I had been lobbying
OASIS to provide
its TCs throughout my tenure. We're now getting close
to the point 
where such facilities will become available to TCs
and IMO that will
likely result in conflicts with the ASIS draft as
currently written
because a document in the SVN store will not have
a filename reflecting
the version number unless you explictly create a new
entry in the
repository for each successive revision of an artifact,
which pretty
much defeats the whole purpose of a version control
system.

> correct).  The burden of going through all 99 files of my subtree
and 
> renaming each and every file to follow the artifact naming guidelines
> (examples, graphics, linked specifications, etc.) would deter me from
> going through the effort to produce everything that is needed by the
> readers.  I ended up with 99 files because that tree of hyperlinked
> documents and examples is what I felt the reader needed.  If
I had to 
> go to so much effort for all 99 files, I would have produced fewer
> files without as much information and the reader wouldn't have been
> as well served.
> 
> I believe OASIS has to make the process of writing specifications
> *easier* in order to help people with limited time involved in the
> already lengthy process of writing to produce something that can be


Bingo

> used.  Therefore, the burden should be focused to accomplish
the goal 
> and not so broad as to deter contributions.  As I said above,
I think 
> it is sufficient that the burden of identifying artifacts at the apex
> directory and any files that might be found outside of the context
of 
> the apex directory.
> 
> Perhaps it is easier than I thought and I was just confused by the
> lack of examples.  In particular, since I chair two subcommittees
and 
> one task group below the UBL umbrella, are these distinctions 
> irrelevant?  Are my work products just considered UBL work products?
> 
> . . . . . . . . . . Ken
> 
> p.s. regarding the templates, I put 33 hours into modifying Norm's
> former DocBook guidelines for OASIS standards into a new set of 
> guidelines to try and help writers using XML, by including all the
> sections matching the Word and OOo templates found in:
> 
>     http://docs.oasis-open.org/templates
> 
> You can find these new stylesheets and complete publishing package
details in:
> 
>     http://docs.oasis-open.org/templates/DocBook/spec-0.4/oasis-
> specification-0.4.html
> 
> In order to try and get my UBL colleagues to use XML to create the
> standards documents, I tried to make this environment as turnkey as
> possible with detailed examples of what to do.  Hopefully by
doing 
> so, members would be more encouraged about writing specification 
> documents since the environment would produce the desired 
> presentation without the writer having to think about it.  To
prove 
> to myself the environment is functional, I've since used the 
> environment to create the two work products announced in:
> 
>     http://lists.oasis-open.org/archives/ubl/200602/msg00062.html
>     http://lists.oasis-open.org/archives/ubl/200602/msg00069.html
> 
> So, your decisions impact on me by bringing my latest work to the
new 
> ASIS standard, and I don't see enough information in there to do a
> complete job.
> 
> 
> --
> Upcoming XSLT/XSL-FO hands-on courses: Washington,DC 2006-03-13/17
> World-wide on-site corporate, govt. & user group XML/XSL training.
> G. Ken Holman                
mailto:gkholman@CraneSoftwrights.com
> Crane Softwrights Ltd.          http://www.CraneSoftwrights.com/o/
> Box 266, Kars, Ontario CANADA K0A-2E0    +1(613)489-0999
(F:-0995)
> Male Cancer Awareness Aug'05  http://www.CraneSoftwrights.com/o/bc
> Legal business disclaimers:  http://www.CraneSoftwrights.com/legal
> 
> 
> ---------------------------------------------------------------------
> To unsubscribe from this mail list, you must leave the OASIS TC that
> generates this mail.  You may a link to this group and all your
TCs in OASIS
> at:
> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php
>