Twitter + SVN hook script == instant history?

[rstackhouse]

I am always thinking about how the folks at home can follow along if
they can't make it to a meeting whether that by by web-cam, video,
blog, Twitter or carrier pigeon.

I've tinkered with SVN hook scripts in the past. I've also toyed with
updating Twitter using Ruby (you can go to http://twitter.com/TAMUUWeb
and play spot the script if you want ;) ).

So I started thinking about combining the two. What if after every
pair got done with their turn at the wheel, they saved the state of
their work to SVN (I'd be willing to go the git or bzr route too I've
just never done hook scripts with either of those). What if after that
SVN commit, a script on the SVN server took the commit message and
flung it out to Twitter? Result, instant CodeDojo timeline. We could
even cache the stuff for posterity's sake on a webpage, since tweets
disappear into the ether after 3 months.

What do y'all think? Stroke of genius or gonzo 3 in the morning idea?

[Mike Abney]

I like the idea for project. It makes it easy to catch up if you
weren't in the room. I am less sure of the value in the context of the
CodeDojo. Or maybe I am just not on the same page with how the Dojo
would work. So in the version you've got below, you would have an SCM
repository, and only one pair working at a time. The value of the
Twitter stream there would be that the pair would commit each time
they ping-ponged or each time a pair member switched out for someone
new. To be a valuable way to follow the Dojo, the commit message would
then need to contain something along the lines of a play-by-play of
what had happened. That sounds like a lot to remember/write....
Perhaps something like that could be done by someone else as the pair
was working. Then, when the time came to commit, the "reporter" could
somehow update their local copy, commit that change to a second repo
with their record of what happened, and *that* gets posted to Twitter.
The only problem there is the 140 character limit. Perhaps posting to
a "liveblog" would work better? A link could be provided that would
take the "follower" to the specific changeset/version.


Mike Abney

http://practicallyagile.com
http://improvingpodcasts.com
http://agilebcs.org

[Robert Stackhouse]

I was thinking that each time a pair left the podium, they or the moderator would commit the code. Not necessarily every time the driver switched.

I would like to go with Twitter to start. I think that could also encourage people to write terse meaningful commit messages with the 140 limit.

--
Robert Stackhouse

http://robertstackhouse.com
http://robertstackhouse.com/friendfeed
http://agilebcs.org
http://bcsbloggers.org
http://uweb.tamu.edu

[Mike Abney]

I'm good with this as a mechanism. (I think better flow might be achieved if one member of the pair leaves at a time rather than the pair leaving at once, but I'm more than willing to experiment.) I'm also good with having it involve training/practice to create concise, cogent commit messages. However, the premise you started with was for this to work as a way to follow along from home or after-the-fact for those who could not be there. For that to work, I think it might require more context around what happened. E.G.: "We tried X, but that didn't work because of Y, as discovered when the test became difficult to write."

Mike Abney

http://practicallyagile.com

http://agilebcs.org

[Robert Stackhouse]

Somebody in the peanut gallery can run the play by play if need be. I've done it for @TAMUUWeb in the past, and am willing to volunteer my services for the Dojo as well if I'm not the one driving.

[dpx (Ben)]

I've seen this on a few projects and in fact I follow on in my RSS
reader for that specific reason. Let me see if I can dig up the code
they use to produce this. I don't know the code but the guy who does
this is Bryan Bishop from UT. http://twitter.com/kanzure and http://heybryan.org.
Someone shoot him an email and see if he will send you the hooks he
uses.

- Ben

On Aug 20, 10:14 am, Robert Stackhouse <robertstackho...@gmail.com>
wrote:

> Somebody in the peanut gallery can run the play by play if need be. I've
> done it for @TAMUUWeb in the past, and am willing to volunteer my services
> for the Dojo as well if I'm not the one driving.
>
>
>
> On Thu, Aug 20, 2009 at 9:54 AM, Mike Abney <m...@mikeabney.com> wrote:
> > I'm good with this as a mechanism. (I think better flow might be achieved
> > if one member of the pair leaves at a time rather than the pair leaving at
> > once, but I'm more than willing to experiment.) I'm also good with having it
> > involve training/practice to create concise, cogent commit messages.
> > However, the premise you started with was for this to work as a way to
> > follow along from home or after-the-fact for those who could not be there.
> > For that to work, I think it might require more context around what
> > happened. E.G.: "We tried X, but that didn't work because of Y, as
> > discovered when the test became difficult to write."
>
> > Mike Abney
>
> >http://practicallyagile.com
> >http://agilebcs.org
>
> > On Aug 20, 2009, at 9:24 AM, Robert Stackhouse wrote:
>
> > I was thinking that each time a pair left the podium, they or the moderator
> > would commit the code. Not necessarily every time the driver switched.
> > I would like to go with Twitter to start. I think that could also encourage
> > people to write terse meaningful commit messages with the 140 limit.
>

> >> > updating Twitter using Ruby (you can go tohttp://twitter.com/TAMUUWeb