Anno., SGML, granularity

Marc Andreessen (
Wed, 11 Aug 93 20:48:49 -0500

Terry Allen writes:
> Sorry, folks, I wasn't thinking when I rejected out of hand
> the notion of offsets for storing the position of annotations
> outside of a file. But this seems like a rat's nest anyway.
> Let me suggest instead that the positioning of annotations be
> made at the granularity provided by the document's IDs.
> That means you wouldn't be able to insert an annotation at
> any point in a paragraph (do we really need to be able to
> do that? is anyone using annotations this way for good
> reason?), and recalcitrant authors could make the granularity
> of possible annotations very coarse by omitting IDs. But
> annotations would clump naturally, instead of being spread
> through the document like a mist, and that might be useful.
> And someone who wanted to encourage annotations could just
> provide lots of IDs.
> More generally, on annotations: how can this possibly work
> at a large scale? If I have a document that the world can
> annotate, and it's controversial, I could generate hundreds
> of thousands of annotations (think of the number of calls to
> the White House when something big is in the news). No one
> can sensibly sort through such a heap. How do Webbers really
> envision annotations working in practice?

At the risk of flooding the net simply by sending a new msg to
www-talk tonight...

The way you (we) want to think of annotations is this: they're virtual
document modifications that may exist on any of several levels: fully
public but not physically attached to the document, fully public and
physically attached to the document [ecccch, I hope nobody actually
implements these], limited to a given subgroup of people, limited to
those who have expressed interest in the opinions of the annotator,
personal, etc.

This means the document author should not a priori be able to
"restrict" or "enhance" the capability of people to add annotations
that are not physically part of her document. That is, you at
O'Reilly should not be able to affect how I at NCSA can add an
annotation to your document that can only be viewed by others at NCSA
-- and vice versa. As long as the annotation isn't actually,
physically affecting your document, you shouldn't care what I do with
the annotation. (Concomitantly, browsers should provide user
interfaces sufficiently well-designed to make it clear what's really
part of the document and what's an annotation added by some yoyo at

With some of the annotation systems we are talking about, we will be
dealing with the problems of scale (thousands of annotations) you
mention: those systems are the public systems [actually, "public"
annotations maintained by the same server storing the document itself
and "global" annotations distributed through a global Usenet-style
network of communicating servers, in my current taxonomy].

There are and will be other annotation systems, or levels, (e.g. the
ones that exist today -- personal and workgroup) where the scale issue
will not be a problem, as very few people (and therefore relatively
few annotations) will be involved and you will a priori (there's that
term again) care about the opinions of those leaving the annotations
(else you could not (a) be yourself or (b) have that group's
annotation server on your list of annotation servers).

Further, mechanisms like kill files, and in the future more
sophisticated mechanisms (c.f. the collaborative filtering work going
on at Xerox PARC -- there was an interesting paper in CACM a while ago
-- basically, you say to the system one way or another "I respect
Bill's opinion but not Joe's" and the system gives you Bill's
annotations but not Joe's), will let you weed out the good stuff from
the bad stuff.

So I don't think we have to worry about "too many annotations".
You'll always just be able to disregard/ignore the ones you don't want
to see, since they will practically never be physically attached to
the documents themselves and the act of fetching and displaying them
will always be a value-add your browser will perform only if you want
it too (I myself would probably never have my browser display any
fully public annotations, etc.)

If we implement a system where a user can place an annotation in the
middle of this paragraph but not in the middle of *this* paragraph, or
at the end of this section but only at the beginning of *this*
section, etc. -- which is what will happen if annotation locations are
tracked solely by the preexisting names/anchors/id's in the document
-- users will be seriously confused and turned off. I have nightmares
about 300 email messages a day from neophyte users saying "You have a
bug in Mosaic -- I can't place annotations where I want, except in
these three cases...."

OK, www-talk mail robot, do your worst...