On 2012-07-24 01:41, Mark Rejhon wrote:
On Mon, Jul 23, 2012 at 5:34 PM, Gunnar Hellström
<[email protected] <mailto:[email protected]>> wrote:
On 2012-07-23 21:17, Mark Rejhon wrote:
Example 1: I suggest that this could be better demonstrated
by not
cutting at the word boundaries "He", "llo, m", "y Juliet!"
maybe, or
something like that. Experience and/or cynicism say that
implementers
are quite likely to look at the examples, ignore the text, and
misunderstand what's going on if the examples provide convenient
semantics not required by the protocol.
[Comment]
I don't like this change. Are you sure?
In some earlier messages, I mentioned that word transmission is
**greatly preferable* *to broken-word transmission.
Also, if an implementer misunderstands, this detail is a more
harmless misunderstanding than broken-word transmission.
There are other examples in the spec.
Comments welcome from people other than Kevin and Gunnar -- I
need more comments because I have comments that they prefer this
Introduction, so I need to reconcile conflicting advice about the
Introductory example. XEP-0301 permits you to transmit real-time
text any way you want: character-at-a-time, word-at-a-time, word
bursts, original typing intervals, time-smoothed, etc. The
Introductory Example is unable to demonstrate all of the possible
methods. IMHO, I chose the 'safest' introductory example.
Again, word transmission is greatly preferable over broken-word
transmission. (There's been arguments in some accessibility
organizations in some countries, some say they prefer keypress
intervals, some prefer word transmission instead of keypresses,
etc.) I am talking to a guy from a telco in UK, and he informed
me of a political debate.
Can at least a few more "outsiders" comment on this change,
please? Thanks :-)
I have also noticed occasional theoretical discussions about word
transmission instead of real-time text. But that just introduces
delays. Long words can take long time to type on small devices,
and many times you have benefit of seeing the word created in
real-time so that you can keep your mind in sync with the sender.
We discussed this before. If a word takes longer than usual to type,
you just simply transmit whatever the word is so far. It will cause
occasional broken words for those times that people take a long time
to compose a word. But that's OK.
Even if it could be mentioned somewhere, the spec is about
real-time text, and the first example, showing the very basic
features shall also show a realistic example of transmitting
real-time text. Not word-by-word.
I disagree. There are implementers demanding word transmission.
Your edit is rejected unless there's lots of demand (i.e. 5 people
publicly agreeing with you with no further disagreements in the
mailing list)
Word-by-word also have the risk of delaying the last typed word
from being transmitted. It must have some inactivity timeout and
transmit whatever is typed if the user just stops typing at the
end of a word without any space or punctuation mark. In order to
not interfere with slow typing, a timeout should likely be in the
order of 7 seconds. That is an unfortunate extra delay in these
circumstances.
That is not a problem, if you have a time limit for word composition
(i.e. 1 second transmission interval, and reset the transmission
interval everytime you hit the spacebar. So that words come out very
quickly, with no delays more than 1 second)
Please accept the proposals for the first example being a
real-time text example.
I need comments from several people. There are people (some of Darren
Sturman's colleagues) who would disagree with you.
Thank you,
Mark Rejhon
The word-by word issue is a theoretical one that has been up for
discussion in the real-time text area occasionally through the years. (
especially in UK, I do not know why. ) I am not aware of any
implementation or investigation that has shown that it has any benefit
over Natural Typing or time smoothing. Even if you say that there is
demand for word-by word, it should not be allowed to confuse the readers
in the first example of a real-time text standard.
I propose that you follow the advice and do real-time text in the first
example.
Please note that no explanations around example one says that it is
intentionally made on word boundaries. It would be more logical to
mention it in the text among implementation notes in chapter 6, and let
the specification start with explaining the main implementation.
And, in fact word-by-word is mentioned in 6.1.4. It is not under a
heading that would make you find it easily. You may change the heading
or divide the section in two if you want it to be more visible.
Gunnar
