Re: Allow <info> as root element?

Hi,

On 30.11.22 00:48, Richard Hamilton wrote:
> I=E2=80=99ve run into the same thing. I have a separate file for our boo=
k info, so that I can move the copyright page to the back of the book for =
ebooks.
>
> I am lazy, however, and just ignore the errors when I=E2=80=99m editing =
the info page:-). I essentially do what Norm mentioned; that is, only pay =
attention to validation when the block is included in a book.

That's possible and probably all okay for your use case. But I think this =
is a
bit dangerous territory as it isn't suitable as a general practice. It cre=
ates
the wrong attitude. ;)

It might be totally okay for a single person if you are working on your ow=
n.
However, I don't think it would be okay for a group of writers. I can't sa=
y:
"Validation is king. Trust it. But ignore it here." Soon nobody trusts (or
wants) validation anymore.

IMHO, this is not a real technical issue (which could be theoretically sol=
ved
by adding <info> to the list of start elements) but I suppose it's more of=
 a
psychological factor. I fear, it would in a way undermine the trust of our
toolchain. Plus it creates some more burden on writers to dissect when to
ignore errors and when not.


> The other thing you could do, if you=E2=80=99re using xinclude, is to pu=
t the info block into a valid root element, so you can validate it, then u=
se xinclude to include just the info element when you build a deliverable.

Yes, technically possible, but still a restriction. Additionally, I'm forc=
ed to
add an XPointer expression into the <xi:include> element just to select th=
e
info. Sounds like a detour to me. :)


> I=E2=80=99m not sure how doable that is with assemblies.

That's why I wrote another post (see "State of topic oriented writing in
DocBook?") which reflects on this part. I hope it's not considered as a ra=
unchy
rant. ;)


=2D-
Gru=C3=9F/Regards
   Thomas Schraitle