Tip:
Highlight text to annotate it
X
DiBona: So...we're here doing an interview.
Liu: We are.
We are here to get nuggets of insight
from Chris DiBona.
DiBona: Little chunklets. Liu: Just little ones.
Because we can't handle anything more.
But first we have to ask about the Chrome comic.
>> DiBona: I should point out
when I heard they were doing comic, I was like...really?
A comic? Are you sure?
They were like, yeah. I was like, okay.
And he, uh, made my goatee very smooth, you know?
Liu: It's very, uh...
DiBona: Sweet... Liu: Yeah.
DiBona: That the internet it.
There were a lot of great parodies, actually,
of that Chrome comics.
>> Liu: So how's--the feedback's been very positive from Chrome.
DiBona: I think people seem to like it, yeah.
There's been already a couple of really interesting products
that have taken the Chrome code
and done different things with it
which I thought was really very positive.
There's already a ton of people out there patching.
sending patches in to Chrome to fix things.
In fact, we think there's more patchers now
than there are employees.
Of course, they don't have the same kind of output.
I mean, we're just talking individual bug fixes,
but it's pretty impressive that, you know,
we're able to look at that many people that fast.
Liu: Yeah, so...
DiBona: I mean, not all patches
are appropriate for us to accept but, yeah.
Liu: So obviously this was a very, you know,
secretive project for a while,
and then you kinda like just put it out there.
So again, companies get flack sometimes
for like throwing code over the wall...
DiBona: Sure, sure. So it's funny, you know.
We actually have these great internal debates
about when it's okay to release some source code.
Is it okay to just release it and forget about?
Is it okay to release it and not have a big deal around it?
Or should they all be heavy, you know, sorta lots of people
working on building a community and all the rest?
I think the answer is you should do all of those things, right?
If you look at most open source projects,
99% of them is one person working on
a code base in the public, and they get no help.
They get no bug fixes. They barely get the users.
And then you have small sets of those,
they try a couple of extra developers,
couple extra bug fixers.
And thus, a community, right?
And then you have others that are,
frankly, incredibly unusual
like Linux Chrome, like Samba, like Mozilla...
which are incredibly large, with thousands of developers,
thousands of people working on them.
And those are not the norm.
Liu: What was the choice--
like what went into the decision to...
hey, we're gonna make this browser.
Why don't we just open up a project on code
and then develop a new open for the entire way?
DiBona: So with Chrome, you know, ask Glen,
as he can probably answer it best.
But there are times when they release projects
that are, um...you know...
you know, being developed in-house
in a very focused manner, you know, for years,
you know, in the case of Chrome and Android both.
I think they're both more than two years old, those efforts.
And so, you know,
they want to basically to tie performance,
tie performance reviews, tie employees to the product.
And...and that's not...
something that's incompatible
with open sourcing, by any measure.
But there is a level of...
"we just wanna do this work ourselves for a while" goin' on.
We're actually-- we're not really under control.
Control takes a lot of work. Man: Right.
I firmly believe the only way to establish control
in open source is not through the use of restrictive licensing
and not through any sort of tricks around that.
I firmly believe that the way you have control
and the way you have leadership in open source
is by showing intelligent
technical leadership on a project.
So as long as the Chrome team at Google, you know,
makes the right decisions technically for the product,
they're gonna be in charge of Chrome
and Chromium, right?
Chromium meaning the open source...
Man: Open source.
DiBona:...version of Chrome that Chrome is built from.
So...with that said,
if people decide that Google isn't giving that leadership,
then they will fork and they'll create something else.
You know, probably call it Technicium or something.
Whatever, you know?
Come up with something I want to name it after.
And, um, the thing is that's okay.
That's a healthy thing.
That's something we wanna actually encourage as a company
so when we release things like Chrome,
we're doing it to make the internet better.
We're not doing it to build market share for Google.
We're doing it to make the internet a free and fair place,
a place that can really exist in a world
where people are used to responsiveness and functionality
in desktop applications.
So...if people take our technology
and advance that strategy, great.
We're very happy. Same with Android, you know?
It's like, you know, if people don't like Android
and they wanna change it
and make it better for their customers,
well, that's good for the idea of a free and fair
and open cellular phone operating system
and cellular phone marketplace as well.
Which is good for Google, right?
So we feel that we compete quite well
in free and fair markets like the internet
and like free and open cell phones.
So that's why we're releasing this technology.
So if somebody says that they can do that
even better than we can... that serves us.
It does well in my eyes.
So we try to encourage anyone.
But, yes, there's some things
we strategically decided not to release.
GFS is one of them.
And so what we do instead in those cases
is we try to ensure that people are publishing.
So, for instance, Jeff Dean and Sanjay Ghemawat
have done papers on GFS,
MapReduce, uh, BigTable.
You know, Luiz Barroso did a paper on, you know,
drive lifetimes.
And if they're read by technical people,
they can come up with these technologies fairly rapidly.
And, in fact, that's happened in the open source world
with Hadu and with HBase
and with some of the other things that we've created.
So I'm actually really happy about this, you know?
I think it's really important that actually
GFS has an open source competitor.
Because it both keeps us honest,
but also, if it gets better than what we have,
we can just adopt it.
Man: Exactly.
DiBona: So we don't feel competitive with Hadu.
But the other side of this is is there's some software
that if we release, we make the internet worse, right?
And that would be like if we released, uh,
the information of how we do web ranking,
page ranking for the web search.
Then web spammers would get hold of it,
and that would be a very big problem.
And not just for Google,
but for Yahoo and Microsoft as well.
Well, I firmly believe, and quality numbers bear out,
that Google gives the best search on the internet.
And part of that has to do
with how we look at these people who abuse the internet
and change how those results are displayed to end users.
So if we were to let that information out,
it would make things very, very bad for not just us
but for Yahoo and Microsoft too.