3: Dave Rolsky, author of DateTime and tons of other modules

Hi.  I am good how about you?

What could go wrong with technology?  Hey, we're programmers, we
know it always works, it's perfect, so...

Very psychedelic.

<span class="transcript-timeindex">1:13</span> OK, sure.  So,
back in '95/'96, I think, I was in grad school for music
composition, that's what my degree is in, and I was, at the time... I
had recently met Matt Mackall, who people might recognize as the
author of Mercurial, and he's also done Linux development, so I
knew him back, way back then, actually through a local vegetarian
group, and I wanted to write a web app for the composers at my
school to list all of our pieces, so that performers could find
what we had written and may be sort, and you know, "I want to find
something that includes clarinet" or something like
that, <span class="transcript-timeindex">2:00</span> and so I
had done a little programming when I was a kid -- I got my first
computer when I was young, 5 or 6, so I asked Matt, "what should I
learn to do this?  What's the right language?"  So he pointed me
at Perl. <span class="transcript-timeindex">2:15</span> I think
I got a Perl book sometime when I was in grad school.  This app
never materialized.  It was a good idea, but it didn't happen,
like many projects, right?  So this should be familiar to
everybody <span class="transcript-timeindex">2:24</span>
watching this: "Projects You Never Did".  That got me started
looking at Perl, which is very ironic, since Matt is totally a
Python guy these days, anyway.

<span class="transcript-timeindex">2:37</span> Exactly,
exactly.  When I graduated from the masters program, I was kind of
burned out on the whole academic thing, and so I decided to not
get a Ph.D., so I end up getting a job in support at a local
insurance company in St. Paul.  I lived in Minneapolis, Minnesota.
<span class="transcript-timeindex">3:00</span> And there was
kind of downtime.  When things weren't crashing, there was free
time, because we weren't like a support line for Dell, it was just
for the company, so if things there were working, we had time.  So
I kind of taught myself more programming there, and I was actually
working on another project that I never finished, which was a web
UI for MajorDomo, if anybody remembers MajorDomo.

Oh wow.  Oh my God.  OK.  So I learned more of Perl there, and
then eventually after about 15 months of that job, I got my first
programming job, this was during the first dot-com boom, when
anybody with a, I'll be polite, a butt, could get a seat, because
that's all you needed was a butt to put in the seat to be a
programmer, <span class="transcript-timeindex">3:50</span> so I
had no idea what the heck I was doing, but I managed to get a
programming job.  And it was a Perl place, and so that was kind of
where it all started.  And actually, that job I met Ken Williams,
who is the person who introduced me to Mason, and we eventually
wrote the Mason book together, and he introduced me to more of the
Perl community
stuff, <span class="transcript-timeindex">4:10</span> pushed me
to put stuff on CPAN, told me about YAPC, things like that.  So
that's where it really got started.

<span class="transcript-timeindex">4:33</span> Yeah, I've
done... I think the first thing I uploaded was something that's
kind of a predecessor to Log::Dispatch.  It was like a logging
module I had written for the company I was working at, and I got
permission to release it, and it was really terrible, but I
uploaded that, and
then <span class="transcript-timeindex">4:49</span> I think I
wrote, I actually wrote a small bug tracking system for our
support team, and I had done my own templating system, I think for
that, I don't know, this was so long ago.  And then Ken was like
"You should look at Mason".  So then I started using Mason, and
then I started contributing to
Mason, <span class="transcript-timeindex">5:05</span> So, I
mean, for some of the things I worked on its projects that other
people have started, probably most notably Mason and Moose, where
I got excited about the project, and, you
know, <span class="transcript-timeindex">5:19</span> if it
doesn't work the way I want it, then patches welcome, and I am
kind of a sucker for that, so I end up doing the patches, and then
working on docs, and all of these other
things, <span class="transcript-timeindex">5:29</span> and then
for other stuff, I mean I have lots of stuff I have uploaded to
CPAN that I don't think anybody has ever used, including me.
There is stuff that is up there that is just experimental, or
that's long
abandoned. <span class="transcript-timeindex">5:39</span> so
even though I have tons of CPAN modules, I suspect maybe, I don't
know, 25% of them, 35% of them see any use, but some of them are
used pretty heavily.  So it's kinda...

<span class="transcript-timeindex">5:59</span> (laugh) I sure
wouldn't trust everything!  I'm embarrassed by my old code.  I
look at Log::Dispatch, which is one of the earliest ones, and I
know lots of people use it, and I still use it, but whenever I
have to hack on it, I'm like "Ugh, not this again".

Someday I'll <span class="transcript-timeindex">6:13</span>
<span class="transcript-garbled">[garbled]</span> to use Moo or
something.

Sure! Sure.

<span class="transcript-timeindex">6:43</span> Yeah.  It really
wouldn't be hard.  Log::Dispatch, in particular, is really not
much code.  It's a pretty simple module.  So the big things I want
to do for that are just kind of clean it up, possibly use Moo,
<span class="transcript-timeindex">6:54</span> and also to
split it up into separate distros.  Right now its kind of an ugly
distro where it ships all these output modules, like for Apache
logs <span class="transcript-timeindex">7:04</span> whose using
mod_perl anymore?  I'm not, so I don't need it.  It has a syslog
output and four different ways to send email, and these should all
be separate
modules.  <span class="transcript-timeindex">7:12</span> Cause
right now you install it and it doesn't have any of the prereqs
declared, 'cause otherwise I'd need to declare tons of prereqs,
and it would be a
mess.  <span class="transcript-timeindex">7:21</span> So
splitting up into different distros would be a nice thing to do
too.  <span class="transcript-timeindex">7:26</span> and we
could <span class="transcript-garbled">[garbled]</span>.

<span class="transcript-timeindex">7:30</span> It would just
make it easier to manage, I think.

<span class="transcript-timeindex">8:04</span> Yeah,
definitely, and it's reasonably amenable to that since each
subclass, each output class is pretty small, doesn't do much, you
could convert them one at a time.  I think that's definitely a
reasonable project.

<span class="transcript-timeindex">8:21</span> For paid work,
or free for software stuff?

<span class="transcript-timeindex">8:28</span> Not too much.
The last couple of years have been, I just have a lot of stuff
going on in my
life. <span class="transcript-timeindex">8:38</span> Some
health problems for me and my wife, which have been distracting,
and also I am doing a lot of activism related to animal advocacy
<span class="transcript-timeindex">8:47</span>, so last year,
the group I work with, we did a big veg-fest in the Twin Cities,
and I was kind of the lead organizer for that, so I spent a lot of
time on that.  And now we're doing it
again <span class="transcript-timeindex">8:58</span>, and I'm
still the lead organizer.  So, you know, things like that can be
distracting.  It's kind of like putting on a YAPC, basically, in
terms of amount of work, to give people a sense of
perspective. <span class="transcript-timeindex">9:08</span>
Probably the biggest thing I've done recently, let's see, well,
there's that module, I don't know if people have seen, called
Courriel, which is a terrible pronunciation
<span class="transcript-timeindex">9:16</span>, which is the
French word, official French word for email, that France uses.
But I just picked it because the Email:: name space was already
taken, and it's kind of an attempt to improve on the
email <span class="transcript-timeindex">9:27</span>
Email::Mime modules, which I don't like the API of much, and just
try to give it a more modern, easy to use
API <span class="transcript-timeindex">9:35</span>.  So, we're
using that at work, and it seems to work pretty well.  I think
other people are using it, so that's something.  There's
definitely work to be done on
that.  <span class="transcript-timeindex">9:41</span> I've been
working on some tools related to calling Postgres command line
programs.  It's the Pg::CLI distro, which actually
is <span class="transcript-timeindex">8:52</span> turned out to
be something we need a lot at work as well.  We're trying to move
to Postgres and for managing
databases <span class="transcript-timeindex">9:59</span>.  And
then I am working on something that I hope, I'm working on it
slowly, very slowly, that I hope will eventually replaces Moose's
built-in type system, but also be usable for Moo, and really be
available standalone.  It's called Specio, or actually I think it
<span class="transcript-timeindex">10:15</span> totally I
should say Specio, but I don't know how to pronounce Latin.  It's
kind of a standalone type
system.  <span class="transcript-timeindex">10:24</span> I
think those some of the most recent things.  I don't even
remember.  What have I done on CPAN?  Who knows?

Not too much.  I hate to admit it.  You know what programming
there is?  Working on Wordpress, and tweaking Wordpress templates,
and writing PHP code, which just - it's unpleasant but necessary.
Wordpress is <span class="transcript-timeindex">10:53</span>, uh,
usable.  It works pretty well for what we're doing, which is a
small CMS with a little bit of blogging.  It's a shame there's no
comparable Perl project that
would <span class="transcript-timeindex">11:07</span> really do
what we want, but writing a Perl project just to run our web site
is probably out of the scope of putting on a
veg-fest.  <span class="transcript-garbled">[garbled]</span>
<span class="transcript-timeindex">11:19</span> There's sometimes
little programs to do data analysis and stuff like that.  I'm
definitely one of the more geeky activists I know.  I try to use
my programming wherever I can.

<span class="transcript-timeindex">11:43</span> That's definitely
what I put on CPAN, I think.  It's like libraries for other people
to use in
their <span class="transcript-garbled">programming</span>.

<span class="transcript-timeindex">12:20</span> So I work for a
company called MaxMind.  It's a pretty small company, I can't
remember how many people right now, 15, 18.  If people have heard
of the GeoIP database, we produce that.  And we also have a
transaction fraud detection service that's built on top of that.
<span class="transcript-timeindex">12:40</span> and GeoIP is
available both as downloadable databases and web services.  So
those are kind of our two main things: the GeoIP stuff and then the
fraud detection.  <span class="transcript-timeindex">12:49</span>
And so I work for them.  We have, I think, 5 developers working on
our various products.  Yeah, that's what I do full time.  It's
pretty good.  I like working there.  We get to release free
software.  We use tons of free software internally.  So yeah, I
enjoy it.

<span class="transcript-timeindex">13:25</span> There's a couple
free downloads, which are kind of a subset of the full data, and
then there's more stuff you can pay for to subscribe to, and then
there's even more stuff if you want to subscribe to the web
service.  So there's kind of three tiers, basically.

<span class="transcript-timeindex">13:46</span> Oh yeah.  I don't
know what percentage.  The vast majority of our code is in Perl.
We have a little bit of C code and C++ code for some of the things
we needed to make fast, but yeah, the vast majority is.  All of
the back end stuff is basically, all the web server stuff, most of
the database generation, is in Perl.

<span class="transcript-timeindex">14:15</span> Pretty much the
whole dev team is remote.  Even the people who are at the office,
there's only a couple of them, so they still have to effectively
be remote.  The company is located in Waltham, Massachusetts.  But
yeah, the whole dev team is remote.  One person who is on my team
besides me people might know is Olaf Alders, who is one of the
guys behind metacpan, so we hired him recently.  He's been great
to work with.  It's definitely a real Perl shop.

<span class="transcript-timeindex">14:53</span> Moo?  Did you say
Moo or Moose?

<span class="transcript-timeindex">15:13</span> Good question.
The class I'm doing is Intro to Moose, and it's basically for
people who already know perl -- it's not an Intro to Perl-type
class -- and hopefully, at least, have some basic familiarity with
object orientation principles, because I'm not going to really
cover that too much either.  But really, its if you want to learn
how Moose works.  You've seen it in other people's code, you've
heard about it, you want to start using it for your stuff, I will
teach you all the Moose principles, all the features it has.  It's
a very interactive class, so what I do is I give a lecture
section, that will last 20-40 minutes, and then there is a series
of exercises that the students do, and the exercises are written.
You write some code, and I've written tests for your code.  And
you've run my test suite against your code, and once your test
suite passes, you know you've done the exercise right.  It tries
to give you a lot of feedback in the test about what's not done
yet.  <span class="transcript-timeindex">16:21</span> I've given
the class many times before.  Students really seem to like that
format of how the exercises work.  I think it's good, because it
forces you -- it's one thing to listen to me talk, but it's much
more useful to be forced to actually use the things that I'm
talking about.

<span class="transcript-timeindex">16:41</span> I'll also mention
Steven Little's class -- maybe you'll interview him too -- he's
doing an advanced Moose class the next day, which will go more
into depth on some of the Moose features, and also cover maybe a
bit more of how to use Moose well, how to use certain features
appropriately, how to mix features together, whereas my class is
more of like here's the basics of what you can do, and not get too
far into here's how to do it really optimally.  I think the
classes will compliment each other nicely, so if people want to
take both I think
that <span class="transcript-garbled">[garbled]</span>.

<span class="transcript-timeindex">17:25</span> Call to action?

<span class="transcript-timeindex">17:34</span> If they want to
get involved in Perl, I think the Perl community really spends a
lot of time on IRC, so, if you join one of the channels on
irc.perl.org, that's a good way to get involved.  If you're
interested in Moose, the #moose channel is quite active, and you
can get help with Moose, both Moose and Moo are discussed on that
channel, so that's a great way, if you want to get involved.
Anything Moose related, that's the place to
go.  <span class="transcript-timeindex">18:06</span> If people
want to help with some of my modules, there are a lot of bug
reports, sadly, for them in rt.cpan.org, so that's always a good
place to start if you want to fix some bugs.  I'm more than happy
to look at patches.  I will warn people, sometimes I'm kind of slow
to respond.  I tend to have a lot of stuff going on.  I felt
guilty about it, but it's not that I'm ignoring you.  And feel
free to poke me.  If you send a path and I haven't looked at it in
a couple weeks, if you nag me politely, (politely, that's the
key), if you nag me politely, I will respond, because then I'll
feel guilty.  <span class="transcript-timeindex">18:40</span> So,
absolutely feel free to do that.  You can also private message me
on IRC.  I'm on IRC as autarch.

OK.  Sounds good.  And email is fine too, but if you have a bug
report, send it to rt.cpan.org.

<span class="transcript-timeindex">19:01</span> That's always,
<span class="transcript-garbled">nothing to encourage
action</span> than tests and patches, yes.

<span class="transcript-timeindex">19:16</span> Alright.  Thank
you.

<span class="transcript-timeindex">19:19</span> So are we all
done?