Havoc's Blog

Aaron,
I don’t really get your post on some level. I feel I’m missing
something. The whole argument I’ve been repeatedly making on xdg list
is that GNOME will not adopt a “D‑Conf” unless it’s actually
better than GConf, and of course Waldo says the
same with respect to KConfig.

For the record, I don’t think the current round of D‑Conf discussion
will go anywhere, because a number of the people involved are known to
be more about sending mail than writing software, and some other
people involved seem very reluctant to listen to the GConf/KConfig
experiences enough to genuinely improve on the existing systems. This
means that the remaining people who might do something useful are
probably confused or drowned out.

If people go off and write something called D‑Conf that doesn’t take
into account the KConfig and GConf lessons, then we won’t use
it. Pretty simple. If they want to try to get it right though, then
let them. What’s to be afraid of?

For my part, I’m just documenting
here
the “lessons of GConf” in the hope that sooner or later someone will
use this information wisely.

BTW, Aaron you say:

there are two good kinds of standards: ones that
document successes that already exist and were, really, already
“standard” before being canonized as such, and ones that are such
great technologies that people run to adopt them.

I agree wholeheartedly and that is how freedesktop.org should work
(and how it has worked in successful cases such as fontconfig and
Cairo and EWMH). Anybody can host their experimental project on
freedesktop.org, and discuss their own crazy ideas on the mailing
lists, but there is no standards committee. The purpose of
freedesktop.org is to allow people to see what’s already de facto, and
give people a place to try to develop great technologies with input
from the desktop projects. Some (most?) of the attempts will fail
and that’s fine.

At least, that’s my view of things. Other people are free to think
whatever they want. I’m not paranoid about it because the major
desktop projects pretty obviously have a veto on any proposed “common”
system or standard in any area, so what people think doesn’t really
matter, only what the core desktop developers think will make a
difference.

I guess the main conclusion is, don’t confuse “people who post to a
mailing list” with “the actual developers” — surely anyone who’s
worked on an open source project understands this…

(This post was originally found at http://log.ometer.com/2005–04.html#10.2)

In an earlier post
I mentioned the book On
Intelligence, the author Jeff Hawkins has now started a company called Numenta.

This is what I wanted to do when I was in school — take a
reverse-engineering pragmatic approach to figuring out how to copy
human intelligence. I couldn’t figure out how to do it in an academic
context though. Most people seemed to be doing research where they
presupposed questionable-sounding theories (the cognitive science
“mind as a bunch of processing modules” or AI “mind as Prolog”/“mind
as logic engine” nonsense). After presupposing everything interesting
they would design experiments to test little nitpicky aspects of it.
The day-to-day life of a grad student looked sort of horrible to be
honest.

So I figured software development is a lot more fun day-to-day, and
even if working on intelligence is interesting, it’d be preferable to
come back to it in some other way later in life.

I’m pretty interested in this Numenta thing then. My guess is that
we’ve decided AI is more star-trek-impossible than it is because of
all the people who’ve approached it in the wrong way. If Numenta gets
even a basic, highly limited version of their concept working, a whole
class of applications that were computationally intractable will
become feasible.

The reason I think Hawkins is more likely to be right than some of the
past train-wreck AI attempts is that his theories have a hope of
explaining actual conversations, creativity, culture, etc. I studied
anthropology/linguistics/pscyhology and it always struck me that the
psychology theories of mind had no chance of explaining the world as
documented by anthropology.

My favorite reading in school was a monster called The
improvisational performance of “culture” in realtime discursive
practice, notable for its awe-inspiringly complex and pretentious
writing style (starting with the title). Once parsed it turns out to
contain a theory of why and how people talk. It spells out the theory
by explaining a single conversation and why that conversation happened
as it did.

If you take this one conversation as an example, the general direction
of Hawkins’s AI thinking has a hope in hell of replicating the
conversation, and the general direction of a lot of other AI thinking
has no hope.

The other thing I like about Hawkins is that he doesn’t seem to be
religious. A lot of people studying intelligence seem to have gone
into it with an emotional investment in the value of
math/logic/computers.

Of course, I never knew that much about all this, and anything
I did know is now almost a decade out of date… plus I’ve forgotten
most of it. So don’t invest in Numenta on my advice. But I’m hoping it
will turn out to be interesting.

(This post was originally found at http://log.ometer.com/2005–04.html#10)

I’ve been trying to get “D‑Conf” discussions on the rails; some people
don’t seem to understand that this is about creating a certain user
experience for admins, programmers, and end users. The requirements
flow from that, not from a bunch of technology modules or libraries
that are available.

I was recently reading a report on VoIP technology consisting of a
discussion between various industry participants. Most of them kept
trying to predict and talk about the future by extrapolating trends,
guessing at upcoming legal changes, finding out about business deals
in the works, and combining technical buzzwords: in this context they
would tend to think “Skype got lots of users because it was P2P”
(which reminded me of this
“architecture astronauts” article).

One of the participants in the conversation had it right, though. He
repeatedly argued that Skype had its userbase because it let you make
free calls, it worked most of the time, and the calls had adequate
quality. In other words it was a useful product that solved people’s
problems, and shockingly, that resulted in people using it.

Check out this thing
(disclaimer: I’ve never tried it, maybe it rocks, but I’m going to
criticize their web site anyway). Their marketing slogan in the page
banner is “Blending: VoIP, IM, Presence, and Social Networking”; since
this web page has a “Geek Zone” link presumably the main page is
supposed to be the non-geek end user zone. How isolated from reality
do you have to be if you think “Blending: VoIP, IM, Presence, and
Social Networking” will make someone buy a product. Exercise: call
someone outside the tech industry and ask them if they would like to
blend VoIP, IM, Presence, and Social Networking today and how much
they would be willing to pay to do so.

My guess is that someone got too caught up in BS about “convergence.”
Maybe “convergence” is such a popular idea because anyone can become
a futurist this way without thinking too much. You just take a couple
of technologies and WHAT IF WE COMBINED THEM???? IT WOULD BE
AWESOME!!!!

To be fair, if that web site is really marketing to VCs or other tech
industry people, maybe their slogan is a good one.

And I don’t mean to say that combining technologies always sucks. Lots
of people use IM and phone calls in conjunction for example, and
surely you could create IM/phone integration that was useful. But just
sticking both technologies in one product doesn’t have any magic
benefit. You have to get down to the details of exactly how they
really work together to create an improved user experience. What does
the software do. It doesn’t matter which market it’s in or
which technologies it uses.

(This post was originally found at http://log.ometer.com/2005–04.html#9)

Take down
the photo, it burns…

(This post was originally found at http://log.ometer.com/2005–04.html#7)

Ross,
maybe just fixing this bug
would be better than a tray icon; the main intended application of the
“urgency hint” is exactly the “you have new IM” type of thing.

Plus you get to use my horrible “make a GtkWidget flash in a
theme-friendly way” hack and terrify your neighbors 😉

(This post was originally found at http://log.ometer.com/2005–04.html#5)

People are always using my name for an 37337 nick; most recently
there’s someone posting comments on LWN with login “havoc” (I am
“hp”). Unfortunately these are often not comments I agree with, and
people attribute them to me. Please don’t attribute them to me. For
all I know the same thing happens on Slashdot or other sites where I
don’t usually read the comments.

If you have a name like “John” or “Bob” then presumably people assume
there are lots of people named that, but for “Havoc” many assume I’m
the only one. Other than the X‑Men character and the GI Joe vehicle of
course.

Notably, both of those are really lame. Havok is one of the lamest
X‑Men, and as that url says about the vehicle “Most don’t like the
Havoc because it’s silly. I agree, it is.”

(This post was originally found at http://log.ometer.com/2005–04.html#2)

Nice article from a Word
for Mac developer echoing my stock reply to most Metacity feature
requests. Question: will I get more flames or less flames if I start
quoting Microsoft in bugzilla? 😉 (Of course the guy is wrong that
open source won’t do the root cause thinking, since he describes
pretty much exactly the GNOME party line over the last few years. For
that matter the first question out of almost any decent developer’s
mouth on hearing a request is “what are you really trying to do?”)

(This post was originally found at http://log.ometer.com/2005–03.html#13)

This
article says:

To sum it all up: OSS developers should learn to live with the fact
that their software is no longer exclusively used by geeks and
enthusiasts– and if they cannot cope with this, then clearly state
that your software is a hobby project– this will save a lot of hassle
and takes away the ground under the feet of ordinary users complaining
and whining over features.

The funny thing is that smart developers use interaction design
methods (such as ethnographic observation) and customer feedback in
order to set the roadmap based on user needs.

A “hobby project” would use all the random stuff geeks and enthusiasts
say on mailing lists to set the roadmap, because hobby projects
confuse mailing list posters with users, and Internet noise with data.

I offer the following definition of geek/enthusiast: you read
Slashdot, OSNews, eXpert Zone, Planet GNOME, this blog, or anything
else of that nature. Yes, this means you. No, you aren’t “an
ordinary user.” Sorry. Neither am I.

Does anyone else see the hypocrisy in posting to a mailing list with
the single canonical list of features that are absolutely necessary
to succeed on the desktop (if I had a nickel for every one of
those…); and then getting pissed when an “arrogant” developer has
their own opinion about the top priorities. How can developers who
have been working on the desktop for years, understand the resource
and technical tradeoffs, have access to direct customer feedback, and
so forth; possibly know more than J. Random Slashdot Reader;
J. Random, after all, is an ordinary user.

One reason Red Hat split Fedora from RHEL is that geeks/enthusiasts do
indeed have different priorities than the general population of paying
customers. This is even more true for desktop than it was for
server. In fact it’s kind of a defining feature of geeks/enthusiasts
that they aren’t paying customers, because they want to do it
themselves. People who pay are people who want a company to make it
work for them. So we split the OS to do two different things for two
different audiences.

I don’t think developers are always right or even usually right. But
they are more likely to be right than a mailing list thread or web
poll.

Your average Linux distribution has at least 50% of its effort and even
code size aimed at geeks and enthusiasts. If I were making a strict
desktop distribution, you’d see sendmail gone, the terminal gone,
Emacs gone, compiler gone, I bet I could even get rid of /bin/bash
eventually. And the product would be better for desktop users because
of the resulting focus and technical flexibility.

(This post was originally found at http://log.ometer.com/2005–03.html#11)

The thing I like about Pages is the handling of “semantic styles” and
their relationship to “text formatting” — while OO.org has a “styles”
feature it’s pretty clunky and the details are all wrong. Pages has a
good model for how styles relate to the text formatting — to see it
you have to enable the styles drawer and look at the features in there
and how the styles work. The way they have done it also allows
templates to be applied/changed after document creation (i.e. is a
template just sort of a blank starting document, or is it a style
sheet). Also I think making styles the default/encouraged way to
write documents was the right decision.

They have the advantages of something like DocBook, without losing
WYSIWYG or introducing weird markup. Making new styles is simple
enough for users to do, and if you want to skip styles and just change
formatting by hand that works also.

Basically what happened here is that the interaction design was done
with styles and templates as the focus, rather than bolting them on as
a bullet-point feature afterward. Keynote has the same basic idea at
its core (you can make an outline, then change its template).

(This post was originally found at http://log.ometer.com/2005–03.html#9)

Marc,
desktop-backgrounds-extra sounds plausible to me, if I remember that’s
the one with outer space photos. The only problem I can think of there
is if it’s the same SRPM as fedora-backgrounds (the add/remove
suggestions have to be according to SRPM, not RPM). If so then there’s
some work to split it apart or maybe just delete it. Probably nobody
thought about removing this package because they were looking at an
SRPM list and so just saw “desktop-backgrounds.”

Of course, if we make that space then why re-add AbiWord rather than
Gnumeric? You see the reality. Somebody is always going to get pissed
off.

I’m assuming that the exact disk space isn’t known until the last
minute (for example, right now we’re rebuilding with gcc 4, which
probably changes the size of everything). I don’t know if this means
that the “dropped from FC3” list will grow or shrink prior to the
final release.

Regarding updates of AbiWord, ultimately Core vs. Extras should not
matter there. We want to enable external maintainers of Core packages,
and we want most of the same rules to apply to Core and Extras. I
think the version of AbiWord in FC3 is probably just what was
available around the FC3 package freeze date.

Personally I think you’re focusing on the wrong thing. To me there is
a path to making AbiWord + Gnumeric + a presentation app into the new
default office suite, and if it were me, becoming the single default
would be my goal. I don’t think Core vs. Extras makes any difference
whatsoever in achieving this goal, honestly.

(This post was originally found at http://log.ometer.com/2005–03.html#3)