Inserting attribution at a better point

Inserting attribution at a better point

[Karl Kleinpaste]

I was just reading/commenting in news.software.readers about the
problems of logically connecting attribution lines with a writers' own
text.  It seems to me that putting the attribution at the very top of
an article is not at all the right place; it should be put immediately
prior to the first unquoted text instead, to connect the writer with
his text better.

That is, the question revolves around the typical format...
________________

In article <123@foo.bar>, joe@schmo.net writes:
> In article <456@foo.bar>, jane@doe.org writes:

>> Jane's text goes here, which is appropriate because it's right
>> beside the attribution for her.

> Joe's text goes here, but it's now a significant distance from the
> attribution which identifies Joe.

New text continues hereafter.
________________

...versus a re-placement of the 1st attribution, farther into the
article, thus...
________________

In article <456@foo.bar>, jane@doe.org writes:
>> Jane's text goes here.

In article <123@foo.bar>, joe@schmo.net writes:
> Joe's text goes here, now closely connected to his attribution.

New text continues hereafter.
________________

Nth-level quotations make the problem of the 1st format all the worse,
because attributions get progressively farther from the text each
writer contributed to the resulting discussion.  Per-depth quotation
coloring helps, but does not solve this problem.  That's why I'd like
to see the 2nd format become possible in Gnus.

(I observe that I also take out the quotation `> ' from in front of
Jane's attribution line.  I'm not sure whether that is /really/
appropriate or not, but I confess that it's how I manually re-work
quotation format whenever I followup.)

How do others feel about this?

--karl

Re: Inserting attribution at a better point

[Justin Sheehy]

Karl Kleinpaste <karl@jprc.com> writes:

> ...versus a re-placement of the 1st attribution, farther into the
> article, thus...
> ________________
> 
> In article <456@foo.bar>, jane@doe.org writes:
> >> Jane's text goes here.
> 
> In article <123@foo.bar>, joe@schmo.net writes:
> > Joe's text goes here, now closely connected to his attribution.

This works for a trivial example like this, but would be of little
help in the common case of multiple attribution:

The outermost layer of attribution (joe) is the one that you are
actually replying to, and the messages that they replied to (jane's)
are included for context.  

If one places their response directly after the bit that is being
responded to, text ordering is important.  People's words are
interspersed for a reason.  

There are some people that seem to think that it is better to simply
include an entire message in one block and to place one's response
above or below the entire original.  Your approach might be useful for
these people, but only because these people's approach is weaker than
the more conventional way of structuring message responses.

> (I observe that I also take out the quotation `> ' from in front of
> Jane's attribution line.  

I don't like that part at all.  It confuses things, as that line was
part of joe's message and thus is quoted by you.

Even if you do rearrange the order of things, I would suggest not
removing any quotation markers.

In general, I don't think that this would be very useful at all.

-- 
Justin Sheehy

In a cloud bones of steel.
  

Re: Inserting attribution at a better point

[Kai Grossjohann]

>>>>> Karl Kleinpaste <karl@jprc.com> writes:

  > (I observe that I also take out the quotation `> ' from in front
  > of Jane's attribution line.

>>>>> Justin Sheehy <justin@linus.mitre.org> replies:

  > I don't like that part at all.  It confuses things, as that line
  > was part of joe's message and thus is quoted by you.

Please look at the above sample (which I think is pretty close to what
Karl usually does).  I find that it is quite clear who said what.
Only problem is what do I do when I want to include more than one
piece from a specific person.  The easiest thing to do is to use more
than one attribution line.

The above format is also similar to what Erik Naggum does.  I wonder
if he has some Lisp support for his quoting scheme?

Sample:

* Karl Kleinpaste:
| (I observe that I also take out the quotation `> ' from in front
| of Jane's attribution line.

* Justin Sheehy:
| I don't like that part at all.  It confuses things, as that line
| was part of joe's message and thus is quoted by you.


Maybe there's some additional spacing somewhere -- the above somehow
looks strange.  But I can't put my finger onto it.  I find Erik's
postings quite readable, too.

kai
-- 
You ate somebody? -- Just a leg. -- That's terrible! -- Not with mustard.
(Terry Pratchett: Interesting Times)

Re: Inserting attribution at a better point

[Lars Magne Ingebrigtsen]

Kai Grossjohann <grossjohann@charly.cs.uni-dortmund.de> writes:

> The above format is also similar to what Erik Naggum does.  I wonder
> if he has some Lisp support for his quoting scheme?

Yes, but it requires manual editing as well in the non-trivial cases. 

> I find Erik's postings quite readable, too.

That they are, but it's not a scheme I would like to see as a general
style because it requires consistency and actual work.  Which is
something one can't assume peope are willing to commit to articles
they write.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: Inserting attribution at a better point

[Michael Harnois]

Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

> That they are, but it's not a scheme I would like to see as a general
> style because it requires consistency and actual work.

As opposed to virtual work?

-- 
Michael D. Harnois, Redeemer Lutheran Church, Washburn, IA 
mharnois@sbt.net                      aa0bt@aa0bt.ampr.org 
Most lies succeed because no one goes through the work to 
figure out how to catch them. -- Paul Ekman

Re: Inserting attribution at a better point

[Kai Grossjohann]

>>>>> Kai Grossjohann <grossjohann@charly.cs.uni-dortmund.de> writes:

  > The above format is also similar to what Erik Naggum does.  I wonder
  > if he has some Lisp support for his quoting scheme?

>>>>> Lars Magne Ingebrigtsen <larsi@gnus.org> writes:

  > Yes, but it requires manual editing as well in the non-trivial
  > cases.

Okay, but WIBNI Gnus provided some support for this style, too?  (The
default style should maybe remain as it is.)  Or would that be
something for Supercite?

kai
-- 
You ate somebody? -- Just a leg. -- That's terrible! -- Not with mustard.
(Terry Pratchett: Interesting Times)

Re: Inserting attribution at a better point

[Justin Sheehy]

Kai Grossjohann <grossjohann@charly.cs.uni-dortmund.de> writes:

> Okay, but WIBNI Gnus provided some support for this style, too? 

Enh.  What would you propose that it do?

If it can't handle quoting messages that contain the more standard
quoting style already, it's not going to be a very useful sort of
automation.

> Or would that be something for Supercite?

Yes, that sounds more appropriate.  Supercite's main pupose seems to
be to provide users with strange nonstandard quoting styles, so this
would fit right in.

-- 
Justin Sheehy

In a cloud bones of steel.
  

Re: Inserting attribution at a better point

[Lars Magne Ingebrigtsen]

Justin Sheehy <justin@linus.mitre.org> writes:

> > Or would that be something for Supercite?
> 
> Yes, that sounds more appropriate.  Supercite's main pupose seems to
> be to provide users with strange nonstandard quoting styles, so this
> would fit right in.

Yup.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: Inserting attribution at a better point

[Kai Grossjohann]

>>>>> Kai Grossjohann <grossjohann@charly.cs.uni-dortmund.de> writes:

  > Okay, but WIBNI Gnus provided some support for this style, too? 

>>>>> Justin Sheehy <justin@linus.mitre.org> writes:

  > Enh.  What would you propose that it do?

I'd like a command `label this region with Lars'.  This should add an
attribution line stating that Lars said the following, and the region
should be massaged to a standard citation style, such as the one you
see above.

For the latter part, I have cooked up the following function:

(defun kai-message-unnest-citation (b e)
  "Retain one level of citation in the region."
  (interactive "r")
  (save-restriction
    (narrow-to-region b e)
    (goto-char (point-min))
    (while (re-search-forward "^ *\\([>:|] *\\)+" nil t)
      (replace-match "  > ")
      (forward-line 1))))

But the part that labels it with Lars' name is still missing.

kai
-- 
You ate somebody? -- Just a leg. -- That's terrible! -- Not with mustard.
(Terry Pratchett: Interesting Times)

Re: Inserting attribution at a better point

[Kai Grossjohann]

>>>>> Justin Sheehy <justin@linus.mitre.org> writes:

  > Enh.  What would you propose that it do?

Well, for simple cases, the message mode default currently generates:

,-----
| >>>>> John writes:
|                  
|   > >>>>> Paul writes:
|   >              
|   > > blurb blurb
|   >
|   > bla bla
`-----

This could maybe automatically changed to:

,-----
| >>>>> Paul writes:
|            
|   > blurb blurb
|            
| >>>>> John writes:
|            
|   > bla bla
`-----

As not all cases are simple, we need a semi-automatic way to do it.
One possibility would be to rearrange the attribution lines (`foo
writes') such that they are in the right order but keep the
nested indentation.  Then one can insert additional attribution lines
as appropriate.

kai
-- 
You ate somebody? -- Just a leg. -- That's terrible! -- Not with mustard.
(Terry Pratchett: Interesting Times)

Re: Inserting attribution at a better point

[Mats Lofdahl]

Kai Grossjohann <grossjohann@charly.cs.uni-dortmund.de> writes:

> >>>>> Kai Grossjohann <grossjohann@charly.cs.uni-dortmund.de> writes:
> 
>   > Okay, but WIBNI Gnus provided some support for this style, too? 
> 
> >>>>> Justin Sheehy <justin@linus.mitre.org> writes:
> 
>   > Enh.  What would you propose that it do?
> 
> I'd like a command `label this region with Lars'.  This should add an
> attribution line stating that Lars said the following, and the region
> should be massaged to a standard citation style, such as the one you
> see above.
> 
> For the latter part, I have cooked up the following function...
> 
> (defun kai-message-unnest-citation (b e)
>   "Retain one level of citation in the region."
> [...]

But why destroy the citations for everybody?  Wouldn't it be better to
post with the multiple level standard citation style and have gnus
optionally _present_ the stuff the way you want it on your screen?
And move the attributions around any way you want.  Surely it would be
as easy to write a article "washing" routine that did this?

-- 
 Mats Löfdahl

 "Good luck.  If you turn out to be a genius after all, remember that I
 never once for a moment believed in you." -- Orson Scott Card (1996)

Re: Inserting attribution at a better point

[Lars Magne Ingebrigtsen]

Mats Lofdahl <lofdahl@usa.net> writes:

> But why destroy the citations for everybody?  Wouldn't it be better to
> post with the multiple level standard citation style and have gnus
> optionally _present_ the stuff the way you want it on your screen?

Excellent idea.

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@gnus.org * Lars Magne Ingebrigtsen

Re: Inserting attribution at a better point

[Kai Grossjohann]

>>>>> Mats Lofdahl <lofdahl@usa.net> writes:

  > But why destroy the citations for everybody?  Wouldn't it be
  > better to post with the multiple level standard citation style and
  > have gnus optionally _present_ the stuff the way you want it on
  > your screen?  And move the attributions around any way you want.
  > Surely it would be as easy to write a article "washing" routine
  > that did this?

Well, it can't happen automatically.  And I did say `WIBNI if Gnus
provided support for this, too?'.  This does not imply to remove
support for the now standard citation style!

Just suppose some people prefer to cite similar to the way Erik Naggum
does, my suggestion was to have some support for this in Gnus.

kai
-- 
You ate somebody? -- Just a leg. -- That's terrible! -- Not with mustard.
(Terry Pratchett: Interesting Times)

Re: Inserting attribution at a better point

[Kai Grossjohann]

Mats Lofdahl <lofdahl@usa.net> writes:

  > Kai Grossjohann <grossjohann@charly.cs.uni-dortmund.de> writes:
  > > 
  > > Well, it can't happen automatically.
  > 
  > Why?  Isn't all the necessary information present if you use the
  > multiple level citation style?

Come to think about it, you're right.  If the citations are preserved
like they are in the above text, everything can happen automatically.

Sorry for the confusion.

kai
-- 
You ate somebody? -- Just a leg. -- That's terrible! -- Not with mustard.
(Terry Pratchett: Interesting Times)

Re: Inserting attribution at a better point

[Mats Lofdahl]

Kai Grossjohann <grossjohann@charly.cs.uni-dortmund.de> writes:

> Mats Lofdahl <lofdahl@usa.net> writes:
> 
> > But why destroy the citations for everybody?  Wouldn't it be
> > better to post with the multiple level standard citation style and
> > have gnus optionally _present_ the stuff the way you want it on
> > your screen?  And move the attributions around any way you want.
> > Surely it would be as easy to write a article "washing" routine
> > that did this?
> 
> Well, it can't happen automatically.

Why?  Isn't all the necessary information present if you use the
multiple level citation style?

> And I did say `WIBNI if Gnus provided support for this, too?'.  This
> does not imply to remove support for the now standard citation
> style!

My point is that if you follow up to a message with multiple level
citations and change them in the way you suggest, then that
information is destroyed.  So whomever wants to reply to _your_
message is forced to keep your citation style.  

_I_ don't think that is nice.

> Just suppose some people prefer to cite similar to the way Erik
> Naggum does, my suggestion was to have some support for this in
> Gnus.

If these people prefer that kind of citations, then why not make the
destruction local to their own screen?

-- 
 Mats Löfdahl

 "Good luck.  If you turn out to be a genius after all, remember that I
 never once for a moment believed in you." -- Orson Scott Card (1996)