On Sat, May 2, 2026 at 8:32 PM Gary Gregory <[email protected]> wrote:

>
> This whole discussion started _because_ there a few places in Commons we
> want to use this. And Commons felt like the right place as opposed to
> Xerces which is a parser implementation. Stuffing this in Xerces carries
> more risk because: (1) xerces is barely active and (2) like with log4j,
> some people are still confused about the separation between API and
> implementation. All of this is why I am backing Piotr's effort for putting
> this in a new Commons XML component.

So first, by putting it in Xerces I mean putting it in the Apache
Xerces project, not making it part of the xerces jar. There might be
some confusion about that. You certainly shouldn't have to depend on
Xerces to pull this into your classpath.

As to xerces being barely active, yes, this is a risk, and one I'm
trying to do something about. However, I don't think it's noticeably
different than commons in that respect. Both currently have exactly
one active developer. The primary difference I see between them is
that commons is less rigorous about code review and multiple approvals
than Xerces is so changes can be made by that one developer without
nagging a reviewer. Neither of these projects is what I would call
healthy, which is why I suspect it might be the better choice to
simply publish this library and keep it out of Apache entirely. Apache
only matters when there's an active community ready to work together
on a project. It's not a good fit for one-off, throw it over the wall
projects.

-- 
Elliotte Rusty Harold
[email protected]

---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]

Reply via email to