cocoon-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Geoff Howard <>
Subject Re: Whiteboard Tool (was: [RT] FirstFriday - monthly virtual Hackathon (wiki pages))
Date Wed, 22 Oct 2003 22:38:32 GMT
Stefano Mazzocchi wrote:

> On Tuesday, Oct 21, 2003, at 23:19 Europe/Rome, Geoff Howard wrote:
>> Stefano Mazzocchi wrote:
>>> doable? digest this first
>>> and come back to me. This is the list of algorithms that they 
>>> implemented in SubEthaEdit.
>> Been reading it.  Not yet digested... :)
> :-)
>> I just finished the section describing the star architecture of 
>> Jupiter which drastically simplified things.  I also wonder if a sort 
>> of "row level locking" (well, actually line-level-locking) wouldn't do 
>> even more.
> I had the experience of writing on the same line.

Ok, you can make the granularity finer - how about the same word?  The 
point I'm trying to make is that the paper we're digesting is driving at 
a general pure solution for the most general case.  But I get the 
feeling that imposing 2-way instead of N-way communication (with the 
"star" architecture as the Jupiter people did) and possibly implementing 
a locking algorithm (at the word or line level, whatever) would simplify 
things to the point that it'd be doable.  We wouldn't earn PhDs for it 
but we'd probably finish in time to actually use it!

It's the difference between finding a geometric solution to tri-secting 
an angle vs. measuring the angle and dividing by three.   The first has 
stumped everyone that has tried for about 2,000 years, the second can be 
done by a 12 year old in a few minutes.

>>   I have the disadvantage of not having ever seen/used SubEthaEdit -- 
>> is it really that useful to be able to have different people editing 
>> the same line simultaneously?
> Yes it is, it is amazing, it's like you never used an editor before.

That's what I'm worried about! ;)  I wouldn't know how to be productive 
if we both started typing different words at the same place, even if the 
algorithm sorted out what order the letters should go in.

>>> Scary to death, if I have to be honest.
>>> But *very* intellectually stimulating.
>>>> We might have to put every letter inside a div tag so we can insert 
>>>> them in strange locations while typing somewhere else, but what the 
>>>> hell, if it works... :))
>>>> Anyone in it?
>> I think that would be monumentally harder and not any more useful than 
>> an existing editor which already has a Memento type of architecture 
>> (which designMode and friends may but don't expose do they?)
> memento-type?

I switched names with concepts.  I meant the Command pattern from GoF, 
where an operation is encapsulated and was mapping this in my mind to 
the operations O that are referred to throughout the paper.  Having now 
finished the paper though I think this is not a direct link because the 
current algorithms only exist for primitive string operations (insert a 
string at position x), not more complex composite operations (move 
section a to position y) which I'm assuming would be the Commands 
implemented by editors (I'm also making the big assumption they do some 
kind of encapsulation like this...).

>>> Count me in, but this is going to require *massive* thinking, 
>>> expecially on how to reduce the problems of latency over the network.
>> The last thing I need is something else to play with, but I'm 
>> interested.  I don't suppose SubEthaEdit is coming out with support 
>> for other platforms?
> No, the state this explicitly.



View raw message