Re: Of the XPointer Patent

["Thomas B. Passin" <tpassin@home.com>]

Now that I've read the wretched thing, I notice some aspects that haven't been
discussed in the postings I've read.  To me, it looks like it was thrown
together in a hurry by people who weren't sure exactly what they wanted to
accomplish.  I notice that the thing isn't signed - on the electronic copy on
the web site - by anyone with legal authority to do so - I'm sure Eve is not
such a person, she just signed the letter.  So we can't even tell from the
page on the web site whether this is approved officially by Sun or whether
it's a just a proposed draft.

- The W3C page that links to "Sun's declaration on patent 5,659,729 and
XPointer" says no more than that.  To me, it's like a courtesy to mention that
Sun has some claims to make.  The W3C page doesn't say anything about
licensing, or imputing validity to the patent.

-  '"XPointer Specification" shall mean the W3C XML Pointer Language
(XPointer) Version 1.0 Specification defining the XML Pointer Language,
referred to as XPointer, or any future version of the specification.'

The current draft is labeled "Version 1".  But it's a draft.  Is the current
draft version really "Version 1" and is it covered or not?  Hard to tell.

- '"Modification" shall mean programming code, including interfaces,
developed by or for You for use in an implementation of the syntax of the
Xpointer Specification or for use in combination with an implementation of
the XPointer Specification, that extends, enhances, alters, otherwise changes
or is derived from the XPointer Specification.'

The XPointer Specification does not discuss any programming code.  Presumably,
this paragraph is intended to mean "if you change or extend the ***syntax***
used in an implementation, that's a modification'.  That wouldn't necessarily
cover extension mechanisms if they used XPointer syntax in expressions.  It
seems clear from the wording in para 5 that, if you add some new syntax -
maybe a new expression that wasn't in the original XPointer - that if Sun
thinks any of their patents are being infringed by the new stuff, they can
still take action if they want to ***for that part only***.  In other words,
if ipart of t isn't XPointer, all bets are off for that part.

This would seem to rule out an xslscript-like approach at first glance,  but
maybe not, since the scripting portions just set up to eventually talk to xslt
(for example) in its own syntax.  So they're wouldn't really be modifying the
***XPointer*** syntax.

- More serious, I think, is the status of "modifications".  If you have a
"conforming" implementation, you are still covered by the "agreement" even if
you extend it, but you have to make your extensions known to the W3C.  Sounds
like other public licenses.  But what if there is a defect in the spec that
you have to fix or work around?  Now you ***can't*** be fully conformant.
Your modifications suddenly aren't covered any more.  I don't think this was
really the intent, but it seems to be written into the document.

Also, the "agreement" covers "a fully compliant implementation" or a subset of
one.  Where does that leave bugs and misunderstandings in the implementation?
Technically, an implementation that had a few problems in its syntax wouldn't
be in compliance.  But what implemenation doesn't have bugs?  So maybe this
thing can never apply.  What are the means by which one can say your code is
or is not in compliance?

- Here's an interesting wrinkle.  Sun wants you to agree not to go after them
for infringement of your patents that may relate to implementation of
XPointer.  That reciprocity is probably supposed to make this more palatable
if tested in a court, I imagine.  Suppose that in the future, Sun abuses your
patents and you want to go after them.  You issue a new, slightly
non-compliant version.  Suddenly, neither you nor Sun is covered by the
non-sue agreement.

However, Para. 5 clouds the issue.  Perhaps what I said only applies to the
non-compliant parts.  Para. 5 makes it hard to tell.

I don't agree with Paul, who said that it's the purpose of lawyers to make
things complicated.  I think the idea is to
1) Write language that has clearly understod meanings ***in a legally tested
sense*** so they can be defended successfully in court;  no matter how
obscure, if words and phrases have legally understood meanings, that's what
you are after.

2) In contrast to the above, to give the lawyer's side the broadest possible
powers and rights within those boundaries.

It's an interesting dance, isn't it?

To me, a non-lawyer, I don't think Sun's  lawyers have accomplished these
goals very well.  But you know what? Sun lets the world do most anything with
Java, at least in actual practice, if your name isn't "Microsoft".  I can't
believe they are going to restrict XPointer more than that, even if the patent
were to survive a test.  A lot of business have been built on Java.

Wow, this post is too long!!

Cheers,

Tom P