ISSUE-29: Complexity of XML Regex to Restricted Character set feature

Just a note to describe chairs' understanding of where we are
on this issue as of 2009-03-31.

- We may need to study the concern more. Any specific aspect of the
feature that may be encumbering implementation efforts? If so, are
they something that we can address or assuage?

- Is there any open-source code base that can be used as the basis
for implementors?

Some thoughts shared by TK on his implementation plan:
http://lists.w3.org/Archives/Member/member-exi-wg/2009Apr/0010.html

Note from 2009-04-08 informal conversation:
DP raised a point that the benefit of this feature may not be large
enough to overcome the implementation burden of the feature.
This amounted to his ACTION-494.

Discussed in 2009-04-15 telecon.
See http://www.w3.org/2009/04/15-exi-minutes.html

There are primarily three issues in dispute.

- Significance of benefit (both breadth and/or effectiveness)
- Implementation complexity (and economic factor relative to benefit)
- Interoperability concerns that may stem from implementation complexity
and lack of existing work that's open-sourced.

DP has taken on ACTION-494 to look into the magnitude of benefits to be brought by this feature.

There was an interesting point made by JS that may be worth considering
if we decide to retain this feature. JS suggested the section had better
be moved up to the main body of the spec, in order to make it clear that
it is something developers have to cope with.

Compactness benefits of the feature is presented by DP.
http://lists.w3.org/Archives/Member/member-exi-wg/2009Apr/0080.html

[caribou]: resolution to keep the RCS feature and continue discussion about complexity

JS to check charset subtraction handling in regex implementation
share experience that would indicate the implication of the operation.
(ACTION-502)

Siemens will get back to the group on their formal position on this issue.
The WG needs to either resolve this issue (preferred), or put off the decision by marking the feature as being at risk, the latter of which AgileDela expressed a flat opposition to.

Siemens' position (DP)
http://lists.w3.org/Archives/Member/member-exi-wg/2009Sep/0017.html

- Siemens sees use-cases that may profit from this feature.
- Siemens (fore)sees potential interoperability problems.
- Siemens fears future objections from outside the group.

"Siemens demands to at least mark this feature as at risk
for now and move on."

The WG agreed that the complexity concern has likely been addressed. There is the remaining concern of interoperability, which is to be indicated in the SOTD section of the CR specification by having a sentence to mark the capability as a "feature at risk".