Re: !MML

Re: !MML

[Sean Doran]

Kai wrote:

| I think if you want to go the route of marking up MIME stuff with
| normal text (as opposed to text properties), then the markup should be
| something which is unlikely to occur in the normal mail text.
|
| Parentheses and SGML-like tags are too likely to occur in normal mail
| text, IMHO.

I think the point was to have an actual language for describing
MIME messages that would be amenable to automation, rather than
to have something specifically designed to be especially user-friendly.

Something Lisp-like seemed pretty natural since Gnus developers
seem pretty fluent in Lisp.

Also, Lisp is not especially user-unfriendly, so if someone who knows Lisp
had some reason to go munging the MIME message at the description-language
level, she or he would have a relatively small learning curve.

Normally keystroke mechanisms would be used to do the manipulation.

| What about saying that a line beginning with a special string is such
| a MIME markup?  Let's say lines beginning with "@.:" or something.
| (Don't use "@#$%", that'll be too often used for four-letter words :-)

Well, I guess APL is another strong candidate...

	Sean. (with mad visions of higher-order functions in Gnus-to-Gnus MIME)

Re: !MML

[Kai.Grossjohann]

Sean Doran <smd@ebone.net> writes:

  > Also, Lisp is not especially user-unfriendly, so if someone who
  > knows Lisp had some reason to go munging the MIME message at the
  > description-language level, she or he would have a relatively
  > small learning curve.

The problem is that we don't want too many things to be wrongly
interpreted as MIME markup.  In fact, we want *nothing* to be wrongly
interpreted thusly.

Interpreting something like `part' preceded by an open parenthesis as
the beginning of MIME markup means that people who have Lisp code in
their messages will likely be in trouble.

Interpreting something like `part' preceded by a less-than character
as the beginning of MIME markup means that people who have SGML-like
text in their messages will likely be in trouble.

Note: in the original version of this message, I actually preceded the
word `part' with the characters mentioned above, and Gnus hang when
posting this message...

kai
-- 
Life is hard and then you die.

Re: !MML

[Lars Magne Ingebrigtsen]

Kai.Grossjohann@CS.Uni-Dortmund.DE writes:

> Interpreting something like `part' preceded by a less-than character
> as the beginning of MIME markup means that people who have SGML-like
> text in their messages will likely be in trouble.

Yes...  How about <#part>?  Is that likely to be something meaningful 
in some language?

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

Re: !MML

[Kai.Grossjohann]

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

  > Yes...  How about <#part>?  Is that likely to be something
  > meaningful in some language?

Good idea.  The end tag could be <#/part> or </#part>.  Hm.  I think I
like the former better.

kai
-- 
Life is hard and then you die.

Re: !MML

[Lars Magne Ingebrigtsen]

Kai.Grossjohann@CS.Uni-Dortmund.DE writes:

> Good idea.  The end tag could be <#/part> or </#part>.  Hm.  I think I
> like the former better.

I've changed 0.47 over to use <#part> and <#/part>.

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

Re: !MML

[Vladimir Volovich]

"LMI" == Lars Magne Ingebrigtsen writes:

 >> Interpreting something like `part' preceded by a less-than
 >> character as the beginning of MIME markup means that people who
 >> have SGML-like text in their messages will likely be in trouble.

 LMI> Yes...  How about <#part>?  Is that likely to be something
 LMI> meaningful in some language?

what about automagical highlighting of MML constructs while people
typeset a mail, so that all MIME things will be visible better and
also in case of accidental coincidence of some user's text with MML
construct it will be visible better? Is it possible?

	Best regards, -- Vladimir.

Re: !MML

[Lars Magne Ingebrigtsen]

Vladimir Volovich <vvv@vvv.vsu.ru> writes:

> what about automagical highlighting of MML constructs while people
> typeset a mail, so that all MIME things will be visible better and
> also in case of accidental coincidence of some user's text with MML
> construct it will be visible better? Is it possible?

Sure; I've added this to 0.49.

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

Re: !MML

[Yair Friedman]

Kai.Grossjohann@CS.Uni-Dortmund.DE writes:

> The problem is that we don't want too many things to be wrongly
> interpreted as MIME markup.  In fact, we want *nothing* to be wrongly
> interpreted thusly.
> 


Why not putting all of this information in a header?
X-Gnus-Structure: ((text/plain  inline (1 . 45)) (image/gif rms.gif (46 .58)))

This way there is no way you can mix the tags with data.

What do you think?  I hope it's not too late....

--
Yair.

Re: !MML

[Kai.Grossjohann]

Yair Friedman <yfriedma@johnbryce.co.il> writes:

  > Why not putting all of this information in a header?
  > X-Gnus-Structure: ((text/plain  inline (1 . 45)) (image/gif rms.gif (46 .58)))

That is a very nice idea, I think.  The information must be updated
after every change, of course.

kai
-- 
Life is hard and then you die.

Re: !MML

[Lars Magne Ingebrigtsen]

Kai.Grossjohann@CS.Uni-Dortmund.DE writes:

>   > Why not putting all of this information in a header?
>   > X-Gnus-Structure: ((text/plain  inline (1 . 45)) (image/gif rms.gif (46 .58)))
> 
> That is a very nice idea, I think.  The information must be updated
> after every change, of course.

One would have to use `after-change-hook' and check the buffer after
every keystroke to see what had changed.  It's certainly doable, but I 
don't think it's necessary.

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

Re: !MML

[Yair Friedman]

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

> Kai.Grossjohann@CS.Uni-Dortmund.DE writes:
> 
> >   > Why not putting all of this information in a header?
> >   > X-Gnus-Structure: ((text/plain  inline (1 . 45)) (image/gif rms.gif (46 .58)))
> > 
> > That is a very nice idea, I think.  The information must be updated
> > after every change, of course.
> 
> One would have to use `after-change-hook' and check the buffer after
> every keystroke to see what had changed.  It's certainly doable, but I 
> don't think it's necessary.

Maybe not just after *any* change.  Even undo groups typing into ~10
characters chunks.  Maybe even more liberal approach than that, like
just when auto-saving or after inserting a big external file/reference
whatever...
--
Yair Friedman.

Re: !MML

[Kai.Grossjohann]

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

  > One would have to use `after-change-hook' and check the buffer
  > after every keystroke to see what had changed.  It's certainly
  > doable, but I don't think it's necessary.

It should be sufficient to do it before disk i/o, i.e. autosave or
normal save.

kai
-- 
Life is hard and then you die.

Re: !MML

[Robert Bihlmeyer]

Hi,

>>>>> On 16 Nov 1998 17:12:36 +0100
>>>>> Kai.Grossjohann@CS.Uni-Dortmund.DE said:

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

 >> One would have to use `after-change-hook' and check the buffer
 >> after every keystroke to see what had changed. It's certainly
 >> doable, but I don't think it's necessary.

 Kai> It should be sufficient to do it before disk i/o, i.e. autosave
 Kai> or normal save.

Yes. The internal represention of the positions may be done as
markers. (auto)saving will convert these back to numeric positions.

	Robbe

-- 
Robert Bihlmeyer	reads: Deutsch, English, MIME, Latin-1, NO SPAM!
<robbe@orcus.priv.at>	<http://stud2.tuwien.ac.at/~e9426626/sig.html>

Re: !MML

[Sean Doran]

| > (multipart :type alternative
| > (part :type text/plain
| > This is plain text.)
| > (part :type image/gif
| > (external "~/rms.gif")))
| > 
| > Er.
|
| Double er.  The problem with this is that `This is plain text.'
| must not contain closed parentheses -- or its parens must be balanced, 

or you make it so that the closing token for a text/plain bodypart
or anything similar is a parenthesis on a line by itself (or beginning
a line with further Lispy MIME directives).

Or perhaps a dot on a line by itself?

(part :type text/plain
This is plain text.
.
) (part :type image/gif (external "~/rms.gif"))

I don't think that a special closing-token for such things
is totally alien, although it is a bit gross.

One might even have (part :type text/plain :closing-token "\rEOF\r"
Some random text
(of whatever nature)
.
Ending like a UNIX here-file.
EOF
)

	Sean.

Re: !MML

[Hrvoje Niksic]

Sean Doran <smd@ebone.net> writes:

> One might even have (part :type text/plain :closing-token "\rEOF\r"
> Some random text
> (of whatever nature)
> .
> Ending like a UNIX here-file.
> EOF
> )

If we absolutely must do it like this, then we might as well lose the
final paren:

(part :type text/plain :closing-token "BYE")
...some text...
BYE

But I still feel this is the wrong way to go.

-- 
Hrvoje Niksic <hniksic@srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
We must all confront and destroy our own inner geek.

Re: !MML

[Shenghuo ZHU]

Content-Type: multipart/mixed; boundary="==-=-="

--==-=-=


>>>>> "HH" == Hrvoje Niksic <hniksic@srce.hr> writes:

HH> If we absolutely must do it like this, then we might as well lose
HH> the final paren:

HH> (part :type text/plain :closing-token "BYE")
HH> ...some text...
HH> BYE

Using MML or !MML, we can take (message-unique-id) as boundary or
closing token, and put --==-=-=


 or (part) between them.

-- 
Shenghuo
--==-=-=--

Re: !MML

[Shenghuo ZHU]

>>>>> "ZSH" == Shenghuo ZHU <zsh@cs.rochester.edu> writes:

ZSH> Using MML or !MML, we can take (message-unique-id) as boundary or
ZSH> closing token, and put --==-=-=

Just put a &lt;part&gt; here. It makes me mad :-(

-- 
Shenghuo

Re: !MML

[Lars Magne Ingebrigtsen]

Shenghuo ZHU <zsh@cs.rochester.edu> writes:

> Just put a &lt;part&gt; here. It makes me mad :-(

:-)

This is a good place to mention that the way to quote the <part>
things in 0.47 is to write <!part> instead.

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

Re: !MML

[Hrvoje Niksic]

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

> Shenghuo ZHU <zsh@cs.rochester.edu> writes:
> 
> > Just put a &lt;part&gt; here. It makes me mad :-(
> 
> :-)
> 
> This is a good place to mention that the way to quote the <part>
> things in 0.47 is to write <!part> instead.

But then the recipient will see the exclamation, won't she?

-- 
Hrvoje Niksic <hniksic@srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
Don't get even, get odd!

Re: !MML

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@srce.hr> writes:

> > This is a good place to mention that the way to quote the <part>
> > things in 0.47 is to write <!part> instead.
> 
> But then the recipient will see the exclamation, won't she?

No, when I write <#part> (that's a <, #, !, p, a, r, t, >) you see
<#part>.  :-)

(The quoting mechanism just removes one bang, so to write <#!part>,
you write <#!!part>, and so on.)

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

!MML

[Lars Magne Ingebrigtsen]

It was suggested to me to use a more Lisp-like syntax instead of a
SGMLey syntax.  So how about:

(multipart :type alternative)
(part :type text/plain)
This is plain text.
(part :type image/gif)
(external "~/rms.gif")
(end multipart)

Hm...  Well, it would make parsing simpler.  :-)

-- 
(domestic pets only, the antidote for overdose, milk.)
  larsi@ifi.uio.no * Lars Magne Ingebrigtsen

Re: !MML

[Hrvoje Niksic]

Lars Magne Ingebrigtsen <larsi@ifi.uio.no> writes:

> It was suggested to me to use a more Lisp-like syntax instead of a
> SGMLey syntax.  So how about:
> 
> (multipart :type alternative)
> (part :type text/plain)
> This is plain text.
> (part :type image/gif)
> (external "~/rms.gif")
> (end multipart)

Would any of this be visible to the user?

-- 
Hrvoje Niksic <hniksic@srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
A professor is one who talks in someone else's sleep.

Re: !MML

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@srce.hr> writes:

> > (multipart :type alternative)
> > (part :type text/plain)
> > This is plain text.
> > (part :type image/gif)
> > (external "~/rms.gif")
> > (end multipart)
> 
> Would any of this be visible to the user?

All of it.  But almost nobody would ever do something like that.
They'd just do `C-c C-a rms.gif RET RET', which results in

(part :type image/gif) (external "~/rms.gif") (end part)

Eh.  I think I like the SGML'ey syntax better.  

(part :type image/gif (external "~/rms.gif"))

would be a possibility, though...  Yes, the `end' things are totally
unLispy.  So the multipart would be:

(multipart :type alternative
(part :type text/plain
This is plain text.)
(part :type image/gif
(external "~/rms.gif")))

Er.

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

Re: !MML

[Hrvoje Niksic]

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

> > > (part :type image/gif)
> > > (external "~/rms.gif")
> > > (end multipart)
> > 
> > Would any of this be visible to the user?
> 
> All of it.

This is disturbing to me, because I tend to mail/post lots of Lisp
code.

> would be a possibility, though...  Yes, the `end' things are totally
> unLispy.  So the multipart would be:
> 
> (multipart :type alternative
> (part :type text/plain
> This is plain text.)
> (part :type image/gif
> (external "~/rms.gif")))
> 
> Er.

Double er.  The problem with this is that `This is plain text.'
must not contain closed parentheses -- or its parens must be balanced, 
depending on how you implement it -- but both strike me as wrong.  Of
course, you could roll in your own parser, like in the case of SGML,
but it sounds yucky.

Again, perhaps it's just me and everyone else is ha-ha-happy about
these design ideas for multipart composition.

-- 
Hrvoje Niksic <hniksic@srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
Personifiers Unite!  You have nothing to lose but Mr. Dignity!

Re: !MML

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@srce.hr> writes:

> Double er.  The problem with this is that `This is plain text.'
> must not contain closed parentheses -- or its parens must be balanced, 
> depending on how you implement it -- but both strike me as wrong.

Yes.  I don't think Lisp is the right thing for free text.

> Again, perhaps it's just me and everyone else is ha-ha-happy about
> these design ideas for multipart composition.

The problem with any document language is that it makes it difficult
to talk about the language itself.  Trying to write a manual on LaTeX
(without \verbatim available) would be a nightmare, for instance.  So
any language we decide on has to be able to talk about itself.  The
MML language (which has only four commands, <part ...>, <multipart
...>, </part> and </multipart> makes it quite easy to achieve this --
one could, for instance, demand that users write <!part> to talk about 
<part>.

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

Re: !MML

[Bruce Stephens]

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

> would be a possibility, though...  Yes, the `end' things are totally
> unLispy.  So the multipart would be:
> 
> (multipart :type alternative
> (part :type text/plain
> This is plain text.)
> (part :type image/gif
> (external "~/rms.gif")))
> 
> Er.

Yes.  Er.

Aren't multipart messages kind of like messages with outlines?  i.e.,
they're buffers with a nested hierarchy of bits in them, and I ought
to be able to move around, and easily chop bits out and paste them
elsewhere.

Hmm, I'm probably talking about a possible user-interface: you'd still
need to have some kind of representation of things (and the right
representation is probably not MIME).

Re: !MML

[Lars Magne Ingebrigtsen]

Bruce Stephens <bruce@cenderis.demon.co.uk> writes:

> Aren't multipart messages kind of like messages with outlines?

I don't know; I've never used outline mode in Emacs.  Is it fun?

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

Re: !MML

[Per Abrahamsen]

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

> Bruce Stephens <bruce@cenderis.demon.co.uk> writes:
> 
> > Aren't multipart messages kind of like messages with outlines?
> 
> I don't know; I've never used outline mode in Emacs.  Is it fun?

I find `outline-minor-mode' very handy in large Emacs Lisp og LaTeX
files.  

You need to bind them to another prefix though, I use this:

	(setq outline-minor-mode-prefix "\C-l")

Re: !MML

[Kai.Grossjohann]

Lars Magne Ingebrigtsen <larsi@ifi.uio.no> writes:

  > It was suggested to me to use a more Lisp-like syntax instead of a
  > SGMLey syntax.

I think if you want to go the route of marking up MIME stuff with
normal text (as opposed to text properties), then the markup should be
something which is unlikely to occur in the normal mail text.

Parentheses and SGML-like tags are too likely to occur in normal mail
text, IMHO.

What about saying that a line beginning with a special string is such
a MIME markup?  Let's say lines beginning with "@.:" or something.
(Don't use "@#$%", that'll be too often used for four-letter words :-)

kai
-- 
Life is hard and then you die.

Re: !MML

[Lars Magne Ingebrigtsen]

Kai.Grossjohann@CS.Uni-Dortmund.DE writes:

> Parentheses and SGML-like tags are too likely to occur in normal mail
> text, IMHO.

Yes, but the only things that Message looks at are <part> and
<multipart>, so it's not that big a deal, I think.

(Beware when responding to this message, though -- Gnus will think
that that <part> thing up there is a part.  Er.  Perhaps the quoting
mechanism should auto-quote these things?  Or something?  This is
getting complicated.  :-)

> What about saying that a line beginning with a special string is such
> a MIME markup?  Let's say lines beginning with "@.:" or something.

That would be a possibility, but I'd like things to look all purty. 

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

Re: !MML

[Vladimir Volovich]

Content-Type: multipart/mixed; boundary="==-=-="

--==-=-=


"LMI" == Lars Magne Ingebrigtsen writes:

 LMI> (Beware when responding to this message, though -- Gnus will
 LMI> think that that --==-=-=


 thing up there is a part.  Er.  Perhaps
 LMI> the quoting mechanism should auto-quote these things?  Or
 LMI> something?  This is getting complicated.  :-)

Is it reasonable to require (and look for) that all MML constructs
must start from the beginning of a line. I.e. <!part> inside of a
line will not be even treated as a MML thing.

	Best regards, -- Vladimir.
--==-=-=--

Re: !MML

[Lars Magne Ingebrigtsen]

Vladimir Volovich <vvv@vvv.vsu.ru> writes:

> Is it reasonable to require (and look for) that all MML constructs
> must start from the beginning of a line. I.e. <!part> inside of a
> line will not be even treated as a MML thing.

Gnus doesn't display this very well (or at all), but RFC2045 (etc)
says that there is a difference between including a part <like, here>
and separating it

<like this>

That is, it should be possible to display graphics in the middle of a
paragraph, and it should also be possible to send something that is
really a collection of separate items.  In a way, MIME is just like
HTML, but with more combersome ways of delimiting structure, as well
as including many different documents in one document.

So, I think it makes sense to be able to <put a picture here>.  That
doesn't rule out reuiring the MML bits starting at the beginning of
lines, but I think it's purdier if people can just write them
anywhere. 

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

Re: !MML

[Hrvoje Niksic]

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

> (Beware when responding to this message, though -- Gnus will think
> that that <!munged-by-hn-=-part-=-> thing up there is a part.  Er.  Perhaps the quoting
> mechanism should auto-quote these things?  Or something?  This is
> getting complicated.  :-)

Yes.  :-)

-- 
Hrvoje Niksic <hniksic@srce.hr> | Student at FER Zagreb, Croatia
--------------------------------+--------------------------------
Try to use "ad nauseam" at least once per flame. It doesn't mean
anything; but it gives that polished feel to your postings.

Re: !MML

[Lars Magne Ingebrigtsen]

Hrvoje Niksic <hniksic@srce.hr> writes:

> > (Beware when responding to this message, though -- Gnus will think
> > that that <!munged-by-hn-=-part-=-> thing up there is a part.  Er.  Perhaps the quoting
> > mechanism should auto-quote these things?  Or something?  This is
> > getting complicated.  :-)
> 
> Yes.  :-)

I've made 0.48 quote parts.

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