Re: SubEthaEdit style networked editing

Re: SubEthaEdit style networked editing

[David Reitter]

> Thank you David! It is a shame this has not been finished... I  
> *really*
> hope these guys do a good job. IMHO Emacs really needs support for
> collaborative editing.

Suppose you could go and help them. It's a collaborative thing :)

Re: SubEthaEdit style networked editing

[Shug Boabby]

> Suppose you could go and help them. It's a collaborative thing :)

heh, yeah. however, having someone who doesn't have the necessary
skillset to help out can be more of a hindrance than a help... my lisp
is basic at best and i know almost nothing about network protocols and
reverse engineering.

SubEthaEdit style networked editing

[Shug Boabby]

Hi everyone,

I have used a program named SubEthaEdit in the past for collaborative
editing of a text file. One computer is designated as the server for
the file, then other users can connect and edit the file... everyone
sees everyone else's edits in realtime in their editor (in different
colours).

This is the most incredible collaborative editing tool I have ever
seen, with only two problems... it is only for the Mac and, most
importantly, it is not Emacs.

Does anyone know of any Emacs packages that accomplish the same thing?
Or has Emacs hit a dead-end in its ability to be extended?

Re: SubEthaEdit style networked editing

[David Kastrup]

"Shug Boabby" <Shug.Boabby@gmail.com> writes:

> I have used a program named SubEthaEdit in the past for
> collaborative editing of a text file. One computer is designated as
> the server for the file, then other users can connect and edit the
> file... everyone sees everyone else's edits in realtime in their
> editor (in different colours).
>
> This is the most incredible collaborative editing tool I have ever
> seen, with only two problems... it is only for the Mac and, most
> importantly, it is not Emacs.
>
> Does anyone know of any Emacs packages that accomplish the same
> thing?  Or has Emacs hit a dead-end in its ability to be extended?

Have you taken a look at
M-x make-frame-on-display RET
?

It is also in the menu bar.  However, for actual colloboration, I
think it better to decouple the working versions via CVS or a similar
version control system: that way people retain consistent copies.

-- 
David Kastrup, Kriemhildstr. 15, 44793 Bochum

Re: SubEthaEdit style networked editing

[Shug Boabby]

> Have you taken a look at
> M-x make-frame-on-display RET

I have heard about such a technique being used for collaborative
editing, but I fail to see how it achieves what I want... perhaps it
has not been explained to me correctly. My understanding is that this
would require all users to use X forwarding and be using the Emacs
installed on a single machine. This is not what I mean... I want to be
able to use my Emacs, with my settings, on my machine, editing the file
on my machine.

> However, for actual colloboration, I  think it better to decouple the
> working versions via CVS or a similar version control system: that
> way people retain consistent copies.

absolutely, i use both SVN and CVS... but these tools are for version
control and syncing versions, not real time collaborative editing of
the same file at the same time. They only realistically allow one
person to edit a file at a time.

The situation I am finding myself in is when 2 -> 5 people wish to edit
the same document at the same time, and see everyone else's changes
appear in realtime. Is such a setup even possible in Emacs?

Re: SubEthaEdit style networked editing

[david.reitter]

> The situation I am finding myself in is when 2 -> 5 people wish to edit
> the same document at the same time, and see everyone else's changes
> appear in realtime. Is such a setup even possible in Emacs?

Someone seems to be working on implementing connectivity to SubEthEmacs
and presumably to other instances of Emacs:

http://www.emacswiki.org/cgi-bin/wiki/SubEthaEmacs

Implementing connectivity from Emacs to Emacs, would be entirely
possible. A very simple way, for example, would be to use Unix `diff'
to find the patches applied in the last minute, then transfer and apply
on the other end. Real 'live' editing shouldn't be too hard either -
for example alongside current "Undo" functionality, which records the
latest changes.

Needless to say, none of this is usable yet. Somebody needs to do more
work on it!

Re: SubEthaEdit style networked editing

[Shug Boabby]

Thank you David! It is a shame this has not been finished... I *really*
hope these guys do a good job. IMHO Emacs really needs support for
collaborative editing.

Re: SubEthaEdit style networked editing

[Mathias Dahl]

"Shug Boabby" <Shug.Boabby@gmail.com> writes:

> The situation I am finding myself in is when 2 -> 5 people wish to edit
> the same document at the same time, and see everyone else's changes
> appear in realtime. Is such a setup even possible in Emacs?

This is a bit off topic and I am not trolling: what are the things you
need to edit with other people at the same time? I have always
wondered, probably because I seldom or never has seen the use myself.

Re: SubEthaEdit style networked editing

[Shug Boabby]

Mathias asked:
> what are the things you need to edit with other people at the same time?
> I have always wondered, probably because I seldom or never has seen
> the use myself.

Text (technical) documents... primarily short LaTeX reports that I am
rapidly working on with 2 -> 5 other members of my team. We are all
sitting in the same room working on a 5 -> 7 page document and giving
it a heavy edit after getting the core content down.

It is not always plausible (especially during the final editing stages)
to designate a set of paragraphs to individual authors, and using
something like svn/cvs is just not responsive enough... i want to see
their edits *now*... not when i ask to show me updates (which i have to
merge with what i have just done in the last 5 minutes, which gets mega
complicated if we even went near each other).

Collaborative editing of plain text is a very different experience than
with code. With code it is structured and everyone knows which bits
they are working on... and if you both end up committing something to
svn at the same time, merge usually has no problems if you worked on
different functions. But with plain text, you could have two people
editing the same paragraph... you also have vocal (or aim, if remote)
communication and feedback... but it's mostly in the form of nods of
approval or disapproval at what they just did on the screen.

The alternative to collaborative editing would be to have everyone
huddle round the same screen and designate one person to do the edits,
but in practice this is difficult and a lot slower. You might know what
you want to write, but you don't know how it looks until it is written.

I'm really surprised that collaborative editing isn't a bigger deal in
the Free Software Community... you could have hackathons on a single
file to rewrite something. So long as everyone has a good idea what
they are doing up front (i.e. template, specification or badly written
file as a base), it is an incredibly fast way to produce work! No more
waiting a day/hour to hear back before you can go change that
algorithm...

Re: SubEthaEdit style networked editing

[Shug Boabby]

Mathias Dahl wrote:
> This is a bit off topic and I am not trolling: what are the things you
> need to edit with other people at the same time? I have always
> wondered, probably because I seldom or never has seen the use myself.

Sorry, my other reply was very long. to answer your question more
directly, I believe that the situation where collaborative editing is
most useful is when you find more than one person huddled around a
computer screen, with a designated typist.

With collaborative editing... you can all be adding your 2 cents, and
you see what other people have added, graphically (the background of
each character is colour coded for each user). and the typist is now in
a better position to contribute... without the stress.

If you don't believe me... go grab "Gobby" and use it the next time you
find yourself looking over someone's shoulder (or with them over yours).

Re: SubEthaEdit style networked editing

[Mathias Dahl]

"Shug Boabby" <Shug.Boabby@gmail.com> writes:

> Mathias Dahl wrote:
>> This is a bit off topic and I am not trolling: what are the things you
>> need to edit with other people at the same time? I have always
>> wondered, probably because I seldom or never has seen the use myself.
>
> Sorry, my other reply was very long. to answer your question more
> directly, I believe that the situation where collaborative editing is
> most useful is when you find more than one person huddled around a
> computer screen, with a designated typist.

I liked it anyway. I am always interested in seing how other people
use their programs and how they work with stuff.

Re: SubEthaEdit style networked editing

[Stefan Monnier]

> editing of a text file. One computer is designated as the server for
> the file, then other users can connect and edit the file... everyone
> sees everyone else's edits in realtime in their editor (in different
> colours).

I'm sure this can be done in an Emacs package, and IIRC someone has done
something along these lines several years ago (but maybe it was for XEmacs
and used some patches to the C code; can't remember well).

Basically you just need to add an after-change-functions hook that sends the
changed text&location to the other processes and have each one of those
processes listen to those messages and apply the change (after adding some
color if needed).

The main difficulty might be in ensuring that the changes are seen in the
same order by every process or to somehow correctly handle the case where
changes are received in a different order.  But it shouldn't be that hard,
really, as long as you can assume a reliable and fast enough communication
between the different machines.


        Stefan

Re: SubEthaEdit style networked editing

[Shug Boabby]

Stefan wrote:
> The main difficulty might be in ensuring that the changes are seen in the
> same order by every process or to somehow correctly handle the case where
> changes are received in a different order.  But it shouldn't be that hard,
> really, as long as you can assume a reliable and fast enough communication
> between the different machines.

Stephan, that sounds promising! I was worried that Emacs would not be
able to handle it. I am by no means proficient enough in lisp or
networked programming to do this, but it is a feature that is becoming
increasingly more popular. In fact, I'd go so far as to say that once
the support is there, the Emacs devel team will find themselves having
"hackathons" on particular files to completely brainstorm new ideas. It
speeds up collaborative editing by several orders of magnitude and the
only downside is in trying to remember all the changes for the
Changelog! :-)

I think in this day and age, it is fair to assume that there is a
speedy connection between all collaborators.

Re: SubEthaEdit style networked editing

[Kevin Rodgers]

Shug Boabby wrote:
> I think in this day and age, it is fair to assume that there is a
> speedy connection between all collaborators.

But not that they are actually working at the same time.

-- 
Kevin Rodgers

Re: SubEthaEdit style networked editing

[Shug Boabby]

Kevin wrote:
> But not that they are actually working at the same time.

of course we can assume this... it would be pointless otherwise! the
whole point of collaborative editing *NOT VERSION CONTROL* is that two
(or more) people can work on the same file *at the same time*.

if you think this has anything to do with version control (two people
editing the same file at different times, saving revisions and storing
version info), then you have missed the point of the collaborative
editing concept. at the end of a collaborative editing session, one
person would commit the file to the version control system with a
Changelog.

Re: SubEthaEdit style networked editing

[Shug Boabby]

Shug wrote:
> Kevin wrote:
> > But not that they are actually working at the same time.
> of course we can assume this... it would be pointless otherwise!

unless, Kevin, you mean at an atomic level :-)

Re: SubEthaEdit style networked editing

[Kevin Rodgers]

Shug Boabby wrote:
> Kevin wrote:
>>But not that they are actually working at the same time.
> 
> of course we can assume this... it would be pointless otherwise! the
> whole point of collaborative editing *NOT VERSION CONTROL* is that two
> (or more) people can work on the same file *at the same time*.

That may very well be the definition of "collaborative editing", but the
Emacs developers (the antecedent of "they") collaborate (i.e. work 
together) asynchronously from distant time zones.

> if you think this has anything to do with version control (two people
> editing the same file at different times, saving revisions and storing
> version info), then you have missed the point of the collaborative
> editing concept. at the end of a collaborative editing session, one
> person would commit the file to the version control system with a
> Changelog.

Sounds spiffy!

-- 
Kevin Rodgers