Hi Benoît,
I appreciate your detailed explanation.
I agree with everything you've sad regarding general terms.
All your points are valid and well-explained.
I think this discussion drifted away from particular situation.
Specifically to git2cl there are still number of questions left:
-- Is our justification for regeneration of README.html is a merely fact that
we can do it? It is unnecessary as the outcome would be very close to original
(with the exception I'm going to explain below)
-- Does everybody here (including Paul Wise) realize that regeneration of
README.html with asciidoc introduce a disruptive and misleading change to the
document?
"Last updated" date in the footer *is* the date of content modification for
reader. Period.
Changing it would mislead regarding when document content was updated and this
is important because it put text into a different context and therefore adjust
meaning of a document. This is unacceptable.
Now why would you go through hassle of regeneration given you know the outcome
will be worse than original? Am I require to waste time in order to do trivial
packaging? Is there are no more important tasks to do than bother with one
file HTML reformatting? Where is the common sense?
Of course I can regenerate document, and patch it to keep original "Last
updated" date. (Obviously it will unnecessarily complicate packaging and
therefore future maintenance for whoever is going to be involved)
To me it is not worth troubles and it goes against common sense.
But as far as I know we don't have to do all this. (Is there a particular
policy paragraph applies?)
Alternatively, I can remove README.html from packaging - it won't hurt because
its content is redundant. (this option hasn't been discussed)
Perhaps removing of this file could be suggested to upstream first, so the
package can be updated later according to upstream decision.
I found interesting to note Paul Wise' reaction.
He doesn't bother himself explaining as we did not discuss any alternatives.
I've been told what to do and my disagreement immediately lead to punishment
("I withdraw my intention to sponsor"...) without any explanations.
But people disagree all the time, even the family members.
We can't be expected to be trained monkeys blindly doing as we told.
Mentoring goal is not obedience but understanding, isn't it?
Just one final note: in most cases formatting is not content.
But in many cases it is. Poetry might be a good example.
Formatting (or alternative CSS) might change perception.
What user does with the document in his browser is user's business.
If you do the change for someone else, you take responsibility for a change.
In that regards, the right thing to do might be to leave original document
alone, unless its regeneration fixes a *problem*.
What problem do we have in original git2cl README.html?
I'm not aware of any.
Kind regards,
Dmitry.
Attachment:
signature.asc
Description: This is a digitally signed message part.