UTF-8 in ZIP

[robert_weir at us.ibm.com]

sc34wg1study-bounces at vse.cz wrote on 11/03/2010 08:22:52 AM:

> 
> sc34wg1study-bounces at vse.cz
> 
> Dear all,
> 
> > Also, it might be worth refreshing ourselves on how patent disclosures 
and
> > declarations work in ISO.  We don't interrogate vendors about patents 
in
> > study groups, or at least not with any expectations of a response. 
WG's
> > develop specifications with normative requirements and then we ask
> > whether _anyone_ is aware of patents that would be necessary to
> > implement these requirements.  If any are raised, this informationis 
sent to
> > ITTF who then contacts the rights owner to see if they will submit a 
patent
> > declaration using the "Patent Statement and Licensing Declaration" 
Form.
> 
> This is exactly right, as I understand it. Of course the earlier 
> such patent awareness is made plain, the better ...
> 

Better for whom?

My point, if it wasn't clear, is that unless and until we have alignment 
of the interests of vendors and other standardizers in this WG, then the 
differences in these interests are going to prevent progress.

My perception of the interests are:

1) Some want to ensure that document packages are and remain 100% 
compatible with existing ZIP-based tools and libraries.

2) Some want there to be a formal standard for a royalty-free archive 
format, suitable for use in document packages.

3) Some want to ensure that there is no fork of the existing ZIP 
Application Note, e.g., an independent standard that might diverge in the 
future and cause interoperability problems.

There are other interests, such as for adding higher-level defintitions 
around IRI path mappings, etc., but these are uncontroversial technical 
details.  The main conflict is expressed in the above three points.

It seems to me that pursuing #2 in the most direct way, e.g., creating a 
new standard that duplicates a subset of the details of the ZIP archive 
structure, will be opposed by interest #3, as well as interest #1.

So how can we satisfy all three interests?  I think that one way has 
already been discussed:

Create a profile standard of ZIP Application Note.  It references the ZIP 
Application Note via an RER.

Interest #1 is satisfied because the profile would only subset ZIP.  It 
would not introduce any new structures at the ZIP level, though it might 
add requirements at the payload level.

Interest #2 is satisfied, because the RER would contain a patent statement 
and PKWARE would be asked to review and approve this.  So it would lead to 
the desired greater clarity around IPR.  They would also get a formal 
standard in the form of the profile standard.

Interest #3 is satisfied because we would be profiling ZIP, not creating 
an independent fork of the core ZIP Application Note.

If anyone else can suggest an alternative way of resolving these three 
interests, then please speak up.  I'm all ears.

-Rob