[Athen] Best practise for threads in discussions.

[Pratik Patel]

As usual I have lots of opinions about this, having used all sorts of
methods.  Sometimes I prefer a combination of  techniques, depending on the
context.  I agree with both Patrick and Sean.   Can I be any more
schizophrenic? I'll use this e-mail to respond to points made by both
Patrick and Sean.


-----Original Message-----
Patrick Burke (PB) wrote:

Overall, after much experience & consideration, I give it a big Who 
Cares. ... But then I'm a long-time proponent of top-posting (when 
you open a new message, the newest content is right there, & if you 
need context you can read further down...). So it may not be wise to 
listen to me. ...

PP: I tend to be an efficiency nut and the top-down method doesn't work well
in a long thread.  One of the main reasons why I love Microsoft Outlook so
much is that it allows one to sort and group messages many different ways. I
find myself grouping messages by threads -- something that Outlook calls
"conversations."  Within that thread group, messages are sorted by date and
time received, allowing me to go through them one after another.  I
recognize that most people use the top-down method to respond; and, as a
screen reader user, it's often far more efficient to go through each message
one after another, just reading the top part and ignoring the rest.  I
recognize that this is certainly an adaptation for the current practice.
But the reality is that getting people to change -- including myself -- on
how they post message is going to be quite difficult.


PB: I suppose the "Re7: ..." method is better than "Re: re: re: re: ...", 
if I have to choose a method of showing this.

If we're talking about an online forum situation, some sites use 
nested lists very effectively to show subthreads. It gets a little 
wordy with all the "nesting level 2" indicators (via Jaws in my 
case). But this method does show very clearly which messages are 
originals & which are follow-ups.

PP: I agree that some forums do employ the list nesting structure
effectively. The down side is the unnecessary bit of chatter.  Instead, I
would highly prefer the Gmail method of indicating messages.  I think Gmail
does a fantastic job of grouping and, in turn, using heading navigation to
allow one to jump from response to response.  Within messages, the Gmail
interface often ends up hiding headers, original messages, responses to
original, etc.  It all depends on what level of nesting we're talking about.
It allows one to unhide a particular message or hidden section on demand.
It is very very efficient.  If screen readers were more reliable navigating
through Gmail, I'd abandon Outlook in a heartbeat.

Sean Keegan (SK) wrote:

One method that has gained some use on discussion lists is similar to 
what I have done in this message.  That is, put either the question or 
main point above and then respond to the message below.  If there are 
several points, then separate each point with your response.  Delete the 
remainder of the actual message.

PP: I do like the method that SEAN points out.  In long e-mails like this
one, I often end up using it myself.  However I am not fond of the ">" char
at all.  I have always found the > symbol quite annoying when used by
e-mails.  It is often not included in the low-punctuation verbosity levels
by default.  Yes I can customize punctuation levels to have the symbol read;
but I'm lazy and, more important, having "greater than" said each time I
encounter a new line is simply annoying.

SK: Some people simply use the ">" symbol to designate the main topic in the

previous message that is being responded to, but others put some 
information such as "My reply" to separate who is writing which 
message.  If the conversation has gone back and forth and the poster 
wishes to refocus to the original message, then I have also seen the 
convention where the phrase "Original Post" or "Original Message" is 
sometimes used instead of a person's name.

PP: Instead, I prefer to use a format similar to what I've done here.  The
first time the poster's name is encountered, I take it in whole and put
his/her initials in parens.  Each time she/he is quoted subsequently, I
preface the  quote by using the initials.  What I've found people to do is
quote someone at the beginning without identifying the person, put their
reply immediately following the quote, interweave another quote from the
original poster without identifying, and then respond to it.  It is the most
ineffective method.  Sometimes the conversation could be clear form the
context and other times, it takes a little while to understand the context.
I am forced to go back to the original messages from the thread to pick up
context and return to the response to gain a better understanding.

SK: Does this help someone using a screen-reader?  I have heard answers all 
over the place, from "yes" to "no" to, as Patrick pointed out, "who 
cares?"  What I do like about separating the main points and replying to 
those individually is when checking e-mail via a portable device (e.g., 
iPhone, Blackberry, etc.) - I get the main points of the message and do 
not have to sift through a bunch of replies to figure out what was the 
original post.  Also, it helps to focus the issues in question and can 
make it easier to understand to what part of a message a person is 
responding.

PP: I do find the integrated threads quite useful when retrieving e-mail
messages on a mobile device.  I agree that the method is most successful
when used properly.  This again brings me back to Gmail.  In situations like
these, a good user interface can make a huge difference.


Warm regards,

Pratik