So, I understand the %!include concept :
%!include(html): ''header.html'
%!include(html): ''footer.html'

But what about people who just wants easily switch between generating two valid 
documents, one with header and the other without ?
I think we probably need one option for that.
And the name "--no-headers" is particularly ambiguous since in txt2tags 
semantic, headers are obviously the three first lines of the document, with 
specific informations.

It's not only useful with the %!include option. Most of the website I create 
are with this --no-headers option, for the simple reason that the "real" 
headers are in my index.php file: my txt2tags files won't need the html headers 
in the target they generate because those targets will be included in the php 
file.

For example:
http://ifiction.free.fr/articles_inform7.php
is included into:
http://ifiction.free.fr/index.php?id=articles_inform7

As for people who want to remove the txt2tags "headers", they can simply remove 
them by hands, or use a postproc rule to delete them in the output.

I think this issue could be closed with a "won't fix".

When I talk of %include concept, it was a general name for txt2tags and 
non-txt2tags classical usage of header and footer.

I agree with you on the point that it's easy to remove the 3 lines of header 
manually.
But txt2tags should makes easy (not manual) most of the common uses.
So the question is : "Is it a common use, a need for users ?". Maybe not, but I 
would like to have one or two advices in more before close.

Finally, there is still the problem of the ambiguous name : --no-headers = no 
header and no footer (in a quite HTML centric point of view), and is not 
specifically related to the 3 txt2tags headers, but remove them because they 
are part of THE header. It's quite confusing, isn't it ?

Florent, to remove the headers from the source file (first three lines), ust 
leave the first line of the source empty. It will be a headless txt2tags source 
file. When converting this file, you'll get the HTML headers as in your first 
request.

But you're right, the --no-headers name is confusing since the footer is also 
removed. A best name would be --body-only.

I think now that templates are here, all this mess will vanish. In the template 
the user specify exactly wich tags he wants to use. Than maybe --no-headers 
could be better renamed to --no-template. Makes sense, since with no template, 
all that we have left is the document BODY.

But boy, everyday I regret more and more about the "first three lines" idea. 
It's simple when you actually *have* 3 lines of headings. Other than that, it's 
just confusing.

Title: Foo bar
Author: John

as the first lines would be a better choice.

About the first three lines : the other good soft, Pandoc, has it too !

The conversion is not so easy, because there is not two worlds, the one for the 
documents with header, and the other for the documents without header. You have 
a document with an header, and sometime you want it headless :

1) leave the first line empty
2) remove the 3 old first lines : you don't want them inside your doc ! -> you 
should save the header somewhere, it's work !
2bis) don't remove them : it's ugly AND you lose all the %%config area ! -> 
doesn't work at all !