tools-yak
[Top] [All Lists]

[tools-yak@collab] Re: "Purple numbers" in a different form

To: tools-yak@xxxxxxxxxxxxxxxxxxx
From: Eric Armstrong <Eric.Armstrong@xxxxxxx>
Date: Tue, 03 Feb 2004 22:18:29 -0800
Message-id: <40208EB5.7070107@sun.com>
Ok. I'll buy that, for now. But I sidetracked the
original discussion.    (01)

Do self-referencing links really have to be numbered?    (02)

John Sechrest wrote:    (03)

> In the NLS system, It was possible to address any object.
> Purple numbers are a "Poor man's reference" that substitute for that.
> 
> If you want to point at a concept, you really would like to get to 
> the specific sentence, or even the specific word of the sentence.
> 
> So purple numbers are a gross approximation of that.
> 
> If you move to the next higher level (subsection instead of 
> paragraph), then you reduce what you can point at even more.
> 
> Right now, the process of pointing at that
> paragraph is too complex/cumbersome for us that only the 
> dedicated do it. IE, we are not cross referencing as much as
> we might like to think we would.
> 
> However, the expresibility is there when we need it.
> 
> I would rather say that we need more access, not less.
> 
> and my desire would be to have the viewspec which would drive
> that.
> 
> So I would say yes, we need purple numbers at every paragraph.
> 
> 
> What you really want is to have a viewspec that says 
> "don't show me the purple numbers" Or "show me purple numbers
> on the heading boundries"
> 
> So really I would wish that we outline what a purplewiki
> viewspec would look like and see how that changes the 
> conversation about purplenumbers.
> 
> 
> 
> 
> 
> Eric Armstrong <Eric.Armstrong@Sun.COM> writes:
> 
>  % It definitely requires a use case analysis. For example,
>  % if I'm linking to a section heading, I'll create a
>  % link like <a href="theDoc#32767">The Heading</a>
>  % 
>  % In that use case, the ID is of no value to the end
>  % user, since it's hidden anyway. What they'll be
>  % matching is the text in the link to the heading they
>  % arrive at.
>  % 
>  % The numbers are arguably of some value when manually
>  % verifying links, because you can keep the two
>  % documents side by side and compare numbers -- if they're
>  % in reasonable sequence in the target doc, which isn't
>  % a given. This is a relatively rare use case, though.
>  % 
>  % Then there is the case where I right click and drop the
>  % link in an email. The link still looks like
>  % http://someServer/theDoc#32767, no matter whether it
>  % appears as "link", "327667", or a little circle on the
>  % target page. So in that case, we can choose anything
>  % we want for aesthetic reasons.
>  % 
>  % The only other use case I can think of where the number
>  % would be useful is when you want to copy or type in the
>  % number, rather than the link.
>  % 
>  % I can see two versions of that one:
>  %    a) I say, see paragraph 32767 in some document.
>  %       --I can't see that happening very much.
>  % 
>  %    b) I create a list of document links with "#" at the
>  %       end, and then type in numbers at the end because I
>  %       find that easier than copying and pasting each
>  %       link separately.
>  %       --while valid, it's pretty rare, so I don't see it
>  %         as compelling
>  % 
>  % Those are the cases that have come to mind, anyway.
>  % None of them to point to a displayed number as being
>  % a necessary part of the equation.
>  % 
>  % While I'm at it, I might as well pick up an ax and threaten
>  % another sacred cow. To wit: Are purple numbers really
>  % necessary on every paragraph?
>  % 
>  % There are probably better use cases to justify that one.
>  % I'm sure they guys who have been using purple numbers
>  % regularly will have found good reasons for it.
>  % 
>  % It will be an interesting if they don't though. My
>  % experience so far suggests that when I want to link
>  % to something, it's invariably to a section heading.
>  % 
>  % (Lists are interesting, in that regard. As are tables.
>  % A link to the whole enchilada seems more useful than
>  % a link to an individual item or cell -- but I won't
>  % be surprised if my intuition is disconfirmed by others.)
>  % 
>  % And so far, I've really found very *few* uses for
>  % transclusion at the paragraph level. In a content
>  % management system, I may want to include sections,
>  % but I'm really not sure about paragraphs.
>  % 
>  % When I quote something, I generally want a sentence,
>  % or I'll pick bits and pieces of multiple paragraphs
>  % to make a quote. (Actually, the need to do that so
>  % has made me more alert to leaving quotable bits in
>  % my own writing.)
>  % 
>  % And when I build up a document using conditional
>  % text, it's not at the paragraph level at all --
>  % it's either a phrase within a sentence, or entire
>  % sections.
>  % 
>  % That's generally true in my experience, anyway.
>  % Which raises the question of whether purple numbers
>  % would be more valuable if used sparingly, to add
>  % the minimal additional clutter.
>  % 
>  % Looking forward to thoughts.
>  % 
>  % 
>  % 
>  % John Sechrest wrote:
>  % 
>  % >  Yes, I understand. But I wonder about the steps of how it is
>  % >  actually done by the user. having to click on the link in order
>  % >  to know that it is the link you want to start at....
>  % > 
>  % >  But I guess much of the value of viewspecs was bound into the
>  % >  chorded keyboard. You could with one key and one click drive
>  % >  thru a document much easier than we can with our current stuff.
>  % > 
>  % > 
>  % >       
>  % > 
>  % > Eric Armstrong <Eric.Armstrong@Sun.COM> writes:
>  % > 
>  % >  % John Sechrest wrote:
>  % >  % 
>  % >  % >  What is the mechanism by which you name the things you want to 
>  % >  % >  see if you have a (purplelinkhere) instead of the actual name
>  % >  % >  of the paragraph?
>  % >  % > 
>  % >  % It would be a link like <a href="...#327676">link</a>
>  % >  % 
>  % >  % So the number would be there, and you could read it in
>  % >  % the browser window when you hover over the link or
>  % >  % copy it. It just wouldn't appear as a number in the
>  % >  % page.
>  % >  % 
>  % >  % That's useful for a large content management systems
>  % >  % based on something like Nodal, where the idea is that
>  % >  % everything exists as an independent unit, so the numbers
>  % >  % get large.
>  % >  % 
>  % >  % It's an idea I've moved away from, but thought I'd
>  % >  % raise the question about numbers anyway.
>  % >  % 
>  % >  % -- 
>  % >  % This message is archived at:
>  % >  % 
>  % >  % 
>http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=tools-yak&i=40202EAB.8050901@sun.com
>  % > 
>  % > -----
>  % > John Sechrest          .         Helping people use
>  % >                         .           computers and the Internet
>  % >                           .            more effectively
>  % >                              .                      
>  % >                                  .       Internet: sechrest@peak.org
>  % >                                       .   
>  % >                                               . 
>http://www.peak.org/~sechrest
>  % > 
>  % 
>  % -- 
>  % This message is archived at:
>  % 
>  % 
>http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=tools-yak&i=40208608.5090506@sun.com
> 
> -----
> John Sechrest          .         Helping people use
>                         .           computers and the Internet
>                           .            more effectively
>                              .                      
>                                  .       Internet: sechrest@peak.org
>                                       .   
>                                               . http://www.peak.org/~sechrest
>     (04)

-- 
This message is archived at:    (05)

http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=tools-yak&i=40208EB5.7070107@sun.com    (06)
<Prev in Thread] Current Thread [Next in Thread>