*start*
01134 00024 USt
Date: 27 May 1981 9:53 am PDT (Wednesday)
From: Mitchell.PA
Subject: Editable Documents
To: Horning
cc: Mitchell

Jim,

I just thought I should record the main points of our conversation yesterday
about editable documents, InterPress, etc.  I am doing so at least partly because I
think it would be a worthwhile policy to use Laurel heavily in this task in order
to keep a complete record (e.g., for historical and patent reasons):

We should use the same base language (abstract machine?) for InterDocuments as
InterPress unless there is a really compelling reason to do otherwise.

More strongly, Interpress is just one son of the original Press, which was to
provide a common format for editable and printable documents; InterDocument is
the other.  A consequence of this is that there should exist documents which
satisfy both the InterPress and the InterDocument standards, i.e., which are both
editable and printable.

Given the above, our real job in developing InterDocument is to understand
very clearly just where it needs to be different than InterPress rather than
designing it from scratch.

JGM

*start*
00431 00024 USt
Date: 27 May 1981 10:17 am PDT (Wednesday)
From: Horning.pa
Subject: Re: Editable Documents
In-reply-to: Your message of 27 May 1981 9:53 am PDT (Wednesday)
To: Mitchell
cc: Horning

Jim,

That puts it very well.

What do you think about "InterDoc" rather than "InterDocument"?

The InterDoc concept corresponding to a "master" should be a "script"
(generalizes from manuscript, typescript, etc.).

Jim H.

*start*
01708 00024 USt
Date: 29 May 1981 7:39 am PDT (Friday)
From: Guibas.PA
Subject: Some Doc Issues
To: InterDoc^.PA
Reply-To: Guibas

I think the establishment of the InterDoc DL is a fine idea. Here are some
questions that the entire enterprise raises in my mind.

1. If an InterDoc document is processed by two different <editor-formatter>s (that
is, transformed into an InterPress file), what do we want to be able to guarantee
about the results? At what level do we expect the two documents (argh, it's so
difficult not to overload this word - may be we should think of a different term!)
to look the same? For example: There clearly must be a representation for "here is
a paragraph of justified text using the current margins". But do we expect both
InterDoc processors to compute the same line breaks? If so the the justification
algorithm has to become part of the standard (and the hyphenation, and the
pagination algorithms, too).

2. All <editor-formatter>s of any scope must rely on a model of document
structure, perhaps based on some kind of a hierarchy, like
book->chapter>-section->paragraph. (Think also of styles in BravoX, property lists
in STAR, macros in TEX, etc.) How can the standard reflect all these different
structures at once? Do we really expect to find a "universal" document structure
so that any other structure can be embedded in it? Or will we punt and flatten
out all structure? Is there any reasonable compromise between these extremes?

Another way to ask the question is this: Do we really expect that when a
document is transformed from a representation private to an <editor-formatter> to
InterDoc format and back, that there will be no information loss?

	Leo

*start*
01513 00024 USt
Date: 29 May 1981 1:13 pm EDT (Friday)
From: Lampson.PA
Subject: Re: Some Doc Issues
In-reply-to: Your message of 29 May 1981 7:39 am PDT (Friday)
To: Guibas
cc: InterDoc^.PA

Thoughts about your questions.

1.  I think it is unrealistic to expect two Interdoc processors to implement the
same line breaks  agreement at that level will be too difficult to achieve, and
also one of the reasons for handling a document rather than a master is the
expectation that fonts, margins etc may change.  On the othr hand, there are
some areas in which details of layout are critical (notably in tables), and we
must be careful to provide adequate control here.  I am not too clear on what that
might mean, however.

2. I don't think the structure will be "universal" at all.  There will be room for
reasonable men to disagree about whether the Interdoc structure is a good one,
and I am sure it will exclude some currently popular ways of doing business. 
This is a bit of a problem with Interpress too, but much less so.  For documents, I
see no hope of avoiding a normative standard which will seriously conflict with
some current methods.

I certainly expect that implementations of the standard must have the same
lossless property that is demanded of Interpress encodings.  If they don't, what is
the standard good for?

By the way, I think Interdoc is a fine working name, but I find it much less
satisfactory that Interpress, and I think we should search for a catchy
replacement.


Butler

*start*
02048 00024 USt
Date: 29 May 1981 11:39 am PDT (Friday)
From: Horning.pa
Subject: Re: Some Doc Issues
In-reply-to: Your message of 29 May 1981 7:39 am PDT (Friday)
To: Guibas
cc: InterDoc^

Leo,

First, a technical note about Laurel 6: In a group "conversation," when Deliver
asks you about a Reply-to: field, the appropriate answer is A (Answers to All);
ESC will cause only you to see most of the replies.

I'm pretty much in agreement with Butler's comments.

It seems essential that a "conforming editor" must be able to translate one of its
documents to a "script" in interchange format and back to internal format without
losing any information.  It also seems impractical to require that it be possible to
interchange documents between independently-designed editors without losing
information, or to require that different editors produce equivalent Interpress
masters from the same script.

I have no desire to standardize justification, line-breaking, hyphenation, etc.  It
seems to me that a script consists almost entirely of text and CONSTRAINTS on its
presentation THAT HAVE BEEN IMPOSED BY THE USER.  Default constraints
supplied by particular editors needn't be represented.  (This is the difference
between automatically-supplied line breaks and user-formatted tables.)

I think that the standard has to make provision for structures, but that it should
not attempt to fix the hierarchy or to interpret it itself.  I.e., at the interchange
level, there is no reason to use different mechanisms to indicate different 
constraints such as
	this range of text is to be bold face,
	this range of text is to be treated as a second-level heading,
	this range of text is to be justified,
	this range of text is to be treated as a paragraph, or
	this range of text is to all appear on the same page.
Perhaps another way of saying the same thing is that the standard should
specify how attributes are to be attached to text, and interpretations for a
standard set of attributes, but leave room for non-standard attributes.

Jim H.

*start*
00211 00024 USt
Date: 29 May 1981 1:17 pm PDT (Friday)
From: Warnock.PA
Subject: InterDoc
To: InterDoc^

Are figures, drawings, charts etc. part of editable documents? (Let's look to the
future folks).

*start*
00547 00024 USt
Date: 29 May 1981 1:29 pm PDT (Friday)
From: Horning.pa
Subject: Re: InterDoc
In-reply-to: Warnock's message of 29 May 1981 1:17 pm PDT (Friday)
To: Warnock
cc: InterDoc^

John,

I guess that it never entered my mind that they wouldn't be (in the standard),
although many editors don't currently support them, and they are likely to form
a relatively small fraction of the documents interchanged in this decade (if for
no other reason than the very different ways that different editors are likely to
treat them).

Jim H.

*start*
00347 00024 USt
Date: 29 May 1981 4:35 pm EDT (Friday)
From: Lampson.PA
Subject: Re: InterDoc
In-reply-to: Warnock's message of 29 May 1981 1:17 pm PDT (Friday)
To: Warnock
cc: InterDoc^

Of course figures, drawings, charts etc. are part of editable documents; surely
that is not in the future.  You'll have to be more adventurous, John.

*start*
01564 00024 USt
Date: 31 May 1981 12:27 pm PDT (Sunday)
From: Mitchell.PA
Subject: Re: Some Doc Issues
In-reply-to: Lampson's message of 29 May 1981 1:13 pm EDT (Friday)
To: Lampson
cc: Guibas, InterDoc^

This is a comment on your response to Leo's last question, vis. "Do we really
expect that when a document is transformed from a representation private to an
<editor-formatter> to InterDoc format and back, that there will be no information
loss?"

I think there is an important difference between InterPress and InterDoc when
one considers going from the interchange encoding to a private encoding and
back again.  In the InterPress case, the master is not edited in this process, while
in the InterDoc case, the conversion will most often be done just so that someone
can edit the InterDoc "script".

Imagine that one has a single-font, text-only, Interdoc-compatible editor, E, and
edits some text of script S.  Assume that E converts S into a private encoding
P(S), then lets a user edit some text, and finally converts P(S') to S'.  I would
certainly expect that the parts of S that E cannot edit, such as illustrations,
should survive the two conversions intact and unchanged.  I am not so sure as
to what can be said about the edited portions of S'.  For example if one edited a
few characters in a Helvetica14 paragraph, would only those characters appear in
E's single font in S', would the entire paragraph be in E's font, or would none of
the paragraph appear in E's font (because it is a paragraph property rather than
a textual one)?

JGM