Re: [Pan-users] About colors of read and watched messages...

[Bobby D. Bryant]

On 2002.12.13 04:10:19 -0600 Maurizio Colucci wrote:

[me:]

> > The second is "read" vs. "unread" vs. "new", i.e. how much previous
> > attention the article has received.
>
> I apologize, but in this sentence you seem to mix different concepts
> once again: the "new" concept is separate from read-unread.

Technically, yes.  But in practice, no.  With the default 
color/boldness scheme the unwatched messages show up _visually_ as 
light-medium-dark, and for text newsgroups that precisely maps onto the 
degree of attention a message has had.  "Light" means you either read 
it or took the trouble to mark it as read.  "Dark" means you have never 
had a chance to see it before at all.  And "medium" means "something in 
between" -- you _have_ had a chance to see it, but you neither read it, 
marked it read, marked it ignored, nor deleted it.  It's still in an 
ambivalent intermediate state between "new" and "tended".  Cognitively 
it works perfectly, at least for busy text newsgroups.

The only difference I would like to have on that part of the encoding 
would be to make the unread color slightly lighter, since I have to 
look more closely to distinguish the bolding from the ordinary unread 
color.

> (Furthermore, but this is another matter, I would completely eliminate
>  the "new" feature because I can see no use for it... after all, new 
> unread
>
> messages and old unread messages are on the same level, since I have
> to read them both. Why a message downloaded now would be more
> important than one downloaded one hour ago?)

It is more important due to incomplete triage.  If I didn't read it 
last time I fetched articles then that means that it's not of pressing 
importance but I still might want to read it later if time allows, but 
right now I want to see whether any of the new messages are more 
important and should be read first.  Coding the new ones means that I 
don't have to waste time going over all the previous unread ones making 
the read-now-or-postpone decision all over again.

It would be unbearable to try to follow a busy text newsgroup without 
the three-level coding that Pan now offers for unwatched threads.  And 
Pan really needs to provide the same support for watched threads.  As 
it is, the "watch" feature is almost useless to me.


> > The current system uses a color scheme to differentiate between
> > watched and not-watched; I suggest continuing that.
>
> Why?

Some threads merit priority attention, and color coding is the easiest 
way to find them.  We're primates with that straight-ahead focus of 
visual attention.


> > The current system also uses a color or value (darkness) or font
> > (boldness) scheme to differentiate between "read" vs. "unread" vs.
> > "new", and I also suggest continuing that.
>
> You still don't say why this solution would be better than the one I
> proposed in reply to Duncan and at the end of this message.

Hopefully I've explained it better above.


> If it's just because it minimizes changes to existing code?
> I can't say that I agree with such an argument.

No, that has nothing to do with it.  Any comments I have made about 
simplicity are just to forestall any notion that it would be a big, 
complex system.


> This would make me sick...
> 1) Suppose you see a bunch on black messages, and between them you
> see, isolated, one "blue" message.
> Now, how do you tell which color is it? Is it dark blue, medium blue,
> or light blue?

Forgive me, but I'm starting to have trouble thinking you are serious 
about this.  For the non-watched threads this is how I've interpreted 
it since the day Pan introduced the bolding for new messages, and I've 
never had the slighest trouble doing exactly what you describe.


> In absence of blue neighbors, this would be at least annoying.
>
> 2) Suppose there are two near messages: one is medium blue, the other
> one is  medium red. The meaning is, according to your definition, that
> one is old,  unread and watched, and the other one is old, read and
> unwatched.

No, "medium" means "neither new nor read", regardless of the color.  
The color only indicates "watched" vs. "unwatched", and can be encoded 
with a color since it is a completely orthogonal issue to what the 
darkness encodes.


> So they are on the same level (in the sense that you should read them
> both).
> But the difference doesn't leap to the eye! THe is NOTHING to bind
> them.
> (except of course, the fact that they have the same intensity of
> color, medium... but this is crazy. Try to imagine).

Yeah, I've been imagining it for several days.  And the more I imagine 
it, the more I want it.  Then I could start using "watched" 
systematically.


> These are only two situations, but I guess there will be many others.
>
> > In fact the boldness for "new" is already orthogonal to watched vs.
> > unwatched
>
> And this is fine, because I must stress that they are independent
> concept...

See top of this message.

[snip rest]

later,

Bobby Bryant
Austin, Texas