[sugar] Annotation (was Re: Viewing PDFs from Browse)

Samuel Klein sj
Mon Oct 13 18:51:17 EDT 2008

On Mon, Oct 13, 2008 at 3:28 PM, Michael Stone <michael at laptop.org> wrote:
> On Mon, Oct 13, 2008 at 11:02:03AM -0400, Samuel Klein wrote:
>> Annotation is different from editing and original creation.
> Unless you are being intentionally vague (as I often am), please be more
> precise. I argue that since annotation is clearly comprised of both
> editing and original creation of a new work based on or in reference to
> some older works, it cannot be _wholly_ different from both editing and
> original creation.

Let me add an exception: Inline references, footnotes and endnotes
intended for inclusion in an original work is a type of edit.   Other
commentary and (private) annotation which are not so intended are not.

The intent, and relationship to other works, of a set of annotation
(associated with a collection of original work; explicitly review of,
commentary on, or links to/from that work) and the average size of a
chunk of meaningful annotation (a handful or links or sentences) is
different from that of original works.

Editing an original work -- modification in place with intent to
produce a new standalone version of that work -- differs from, though
is often related to, annotation of that work.

One can treat annotation as a document diff -- wikis did this for many
years before the 'page namespaces' became omnipresent.  This treatment
is generally deprecated (save for traditional inline footnotes and

One can treat a document diff as a type of annotation -- but these are
two distinct modes with different ideal methods for display and
review, and different expectations for ownership and the importance of
sequence preservation, and different priorities for how tightly they
are bound to the original; so the way they are displayed, aggregated
and updated should in practice be different.

>> A user can in fact annotate a file or document they do not have
>> locally.
> I don't believe you -- or perhaps I simply reject your usage of the word
> "have". Example, please?

Memo: Hamlet, I.iii.  Polonius' monologue.  Pol. warns against the
credit crunch, hundreds of years in advance.  [sj, 08/10/13]

Memo: <aaaaaaaaa.gif>, attached.  Commentary on section 2 or 3 of the
article [[uncylopedia:AAAAAAAAA!]] (I don't have it handy); consider
for addition next to Louis Armstrong.

>> I find that use case more likely than the alternative; I
>> have never succeeded in pushing a book report upstream into a
>> publisher's next edition.
> a) You've never contributed errata to an author?
> b) I imagine, perhaps incorrectly, that remote annotation works best
> with a cheap well-functioning network. Do you agree?

a) Errata, and a proofreader's annotation of a work, are an
interesting middle ground of edits which are recorded and shared as
annotation.  I still find the use case of commentary and review far
more widespread.

a')  A letter to the author is out of scope here, neither edit nor
normal annotation, but of interest to a broader discussion of

b) I do not know what you mean by 'works best'.  Most remote
annotation occurs offline and is never aggregated, but works just fine
for the production of notes for one's personal journal.  We would be
better off if there were a ready way for noters to smoothly aggregate
and share such notes, but smooth doesn't mean instantaneous; it would
be enough for annotations to aggregate smoothly with an asynchronous
network with effective latencies in months.

>> A set of annotations is in general much smaller on disk,
> Why would you assume that annotations occur in the same medium as the
> original work? For example, I could easily imagine calling a video-taped
> lecture about a book or collection of photos an "annotation".

And I could call 'annotations' works of art that expect physical
interaction and weigh thousand of tons.  But I will not; annotations
are normally associated with brevity, and we can produce useful tools
for handling annotations that expect this.

>> much more frequently versioned,
> Evidence or reasoning please?

primary sources, art, and other referenced works tend not to be
edited, but only annotated.  Complete or mostly-complete works tend to
be edited infrequently, but annotated at length.  New or incomplete
works tend to be edited frequently, with only limited annotation
(there is less to comment on).  I was thinking of interactions with
complete works, but perhaps "much" was too much.

>> and more likely to need sharing (rather than
>> independent discovery of copies of the same work) than originals.
> This also makes little sense to me.

Again, thinking of complete or reference works, you and I don't need
to have access to the same original of Hamlet to discuss it
intelligently.  I do however need to share my annotations with you to
discuss /them/ as they are not already available through a standard
library or web search.

>> A set of annotations made to one version of a document are also
>> relevant to other versions,
> They are sometimes relevant; it seems to me to depend on the
> rate-of-churn of the annotated work. Another problem is that it's often
> unclear whether you want to annotate a work or the context in which
> the work occurred.

I mean variations on a Work, not revisions in a revision history.
My comments on the 1623 First Folio Hamlet generally apply to your
2003 Dover; and your notes on your King James to my American Standard
-- even if the way they apply is in the discovery of meaningful
differences between the versions.

>> Conversely defining annotations such that they exist separately >> from a
>> document helps make different sets of annotations overlay with >> one another.
> Perhaps, but it also tends to curtail their visibility.

I'm not sure how you mean this.  In the sense that not etching my
comments on Hamlet into the original folio curtails their visibility?
to the contrary, I would say that keeping them separate helps
disseminate them more broadly.

>> To use your image-drawing example, if there is an original image that
>> I can expect others to have seen, and I am drawing on top of it, I
>> would indeed hope that the drawing is stored in such a way that the
>> original can be deciphered
> _Why_ do you prefer immutable history to destructive updates?

Firstly, because the original Mona Lisa is beautiful.
Secondly, because when I am annotating I am not intending to make new
versions of the original, destructive or otherwise.

> I can do
> rather things with liberal application of scissors, paint, and glue
> which are just as valuable (though they are non-invertible) as the
> things I can do with my version control system. (The point being that
> destructive updates are fine to me, and if history is needed, then
> they can be applied to a copy.)

I'm not against glue, but you've lost me here.  Can you give an
exapmle of something valuable that you can do with scissors and glue
that you cannot do with your version control system?  Can you give an
example of something you do with either glue or version control that
you cannot capture as a separate annotation?

>> -- so that if you and I draw different
>> things, I can tell that it was on the same image,
> It's not necessary for the original content to be _decipherable_; it's
> enough for the original content to be _recognizable_ should you
> encounter it again (or want to go looking for it).

As long as 'looking for it' and 'recognition' can be done
automatically on a large scale, I do not mean to privilege perfectly
deciphering the original.

> So what? If your system is too hard to build or to run then I probably
< wouldn't have been able to <BOOM DE YADDA> in the first
> place, in which case we'd both be poorer.

That sounds like a darn good argument for not trying to build that
system, whatever it is.   Your system?  My?  Any system, really.

(I thought we were talking about how to handle annotations in Read,
Browse and Sugar, and whether they should be tightly bound to having
the annotated work in your Journal offline -- I don't know where other
systems came into it)

>> After passing through a world with millions of works and terabytes of
>> data, the small part of this that I hope to preserve as my own gloss
>> and memory of it, the part I might keep in a real journal, is
>> effectively my set of annotations, assessments, margin notes :
>> these should be built deeply into [Sugar], and the usability and
>> sharability of them refined.

Sugar is in brackets here since this is my opinion in general for the
interface I would like to  use every day.   This discussion was
inspired by the sugar list, and the discussion was about in the future
recording interactions with and notes about books being read in the
sugar Journal.

>> See also [[annotation]] on the wiki, for the above and for discussion
> Prior work in this area is worth citing; to that end, I have created an
> appropriate section at the bottom of your [[Annotation]] page.

Thanks!  Some more cites have been added.  Ian, Josh, Rufus : take a
look at that page, and read the original thread if you like here:

I don't have Geof Glass's email handy, but he might be interested in
where this conversation is going

-------------- next part --------------
A non-text attachment was scrubbed...
Name: aaaaaaaaa.gif
Type: image/gif
Size: 96146 bytes
Desc: not available
Url : http://lists.laptop.org/pipermail/sugar/attachments/20081013/036e769c/attachment-0001.gif 

More information about the Sugar-devel mailing list