|
Mike
Sugarbaker Reports
October
28, 2002 | Up at the Lawrence Hall of Science, a schoolkid-oriented
museum at the crest of the Berkeley hills, the only trace of San
Francisco that's visible above the fog is Sutro Tower, or, as my
San Francisco hipster acquaintances like to call it, "Big Friend."
The Open Source Content Management Conference, being held up here
in the same auditorium where Mom used to leave me for Saturday afternoon
to watch obscure educational cartoons, is the kind of environment
where it's easy to believe that communications technology is all
that matters. It does matter, though. If you use a blog, are part
of an online community or read a major web publication, content
management software shapes the information you get, and the discourse
you make. Whether the open-source part matters to you depends on
your political lean and your pocketbook, but blogs wouldn't be anywhere
if so much blogging software weren't free as in beer - and new forms
for online communication come about when people are free to hack
around. That's why I went to cover the proceedings for Mindjack:
to see if anything new is possible, if any new doors have opened
towards making the dream shapes I have in mind.
The first attendee I talked to said he was building a business
and therefore building a CMS - or, he quickly added, trying not
to build one. I joked that it seemed almost as hard not to build
your own as to build it. Serious open-source CMSes often suffer
from unwieldiness and feature bloat, by necessity. As Joel Spolsky,
developer of the closed-source desktop CMS CityDesk, has said on
his popular blog, it may be true that 80% of users use 20% of the
features of a program, but they all use a different 20%. CMS has
this problem in spades, as every content flow has to fit a different
set of business processes - usually rife with politics and executives'
sometimes-oddball pet theories about what the UI should be. Weblog
managers like Blogger hit substantially less than the 20% the average
business needs, but what they do hit, they hit dead on, precisely
because they're so limited. It's like the example Alan Cooper quotes
in his usability opus The Inmates Are Running The Asylum: those
great little rolling suitcases with the pullout handles were designed
exclusively for stewardesses, but now everybody wants one because
they're so relentlessly good at the specific things that frequent
travelers need. Do certain things incredibly smoothly and be free,
and the world is yours.
The blog talk was very light at OSCOM, which surprised me. The
closest we got to a formal presentation on anything bloglike was
the presentation on PostNuke and Xoops, two spins on the community-weblog
model pioneered by Slashdot. Most of the software packages being
represented in formal talks were content management frameworks -
that is, not CMSes, but tools to build exactly the CMS your business
needs. If the scheduled talks were any indication, the attendees
wanted to hear about tools, not shapes - how to do it, not what
they might do. Community or new modes of communication were barely
in the picture.
Given all that, Charles Nesson's keynote seemed discordant, and
pleasantly so. The bulk of his talk centered around Jamaica, and
the various educational and correctional organizations that his
Internet and Society Conference at Harvard is working with. Specifics
of software weren't the point, although I was tantalized by a mention
of an education package called H20, in which professors pose questions,
students provide possible answers or comments, and then rate each
other's comments. Alone amongst the presenters, Nesson expressed
lots of interest in the possibilities of content management for
forms of content besides text. He brought up a Harvard-run digital
music project in a low-income computer cluster in Rock Springs,
Massachusetts, where underprivileged kids learned about computers
by making music. At first, I thought some jackass in the audience
hadn't turned his laptop speakers down, but it turns out Nesson
was playing the result of one of these workshops from his presentation
machine the whole time. I couldn't tell it from professionally produced
downtempo.
The "digital divide" debate obviously looks different to Nesson
than it does to many of the rest of us, who might conclude, like
Bill Gates famously did, that there's not much point in bringing
information technology to countries and communities that lack running
water and sanitation. Nesson's work in Jamaica through the Internet
and Society Conference shows that it isn't that simple. By targeting
educational and correctional facilities - the carrot and the stick
in high-poverty societies - IT professionals can contribute to some
real change. Maybe not in the poorest of poor countries, but Jamaica
at least has the infrastructure to plug computers in, in some buildings.
Giving teachers good tools in schools, and in educational environments
in prisons, is a small enough task to be doable and a meaningful
enough task to put energy into.
After Nessom's keynote, the technical talks began. I won't do the
complete rundown of all the "competing" open source content management
frameworks. I'll cut to the chase: the winner is Plone. This "productized"
take on the six-year-old web application framework Zope was the
package with the most tools, the most professionalism, the most
traction, and above all, the most buzz. (The folks from Red Hat,
representing the former Ars Digita, were certainly professional
but didn't have the same verve when it came to commenting on other
presentations and making themselves a part of the whole discussion
- professionalism in the sense of the sciences rather than business.
And it never hurts to have free T-shirts for everyone.) Plone is
not quite 1.0 yet, and in my own experimentation I found it to be
not quite as stable as unvarnished Zope, but the finished Plone
will have a lot more, well, finish. Where Zope is a wide-open toolset
that doesn't totally make it clear what you can do with it when
you open it up, Plone will help would-be content managers dive right
in.
Plone installs easily on Windows and Unix, and has its own built-in
web server - like every framework we actually saw a demo of, its
interface pops up in a web browser. Like in Zope, you create folders
and pages in what looks an awful lot like a file system. Unlike
Zope, Plone has a complete collection of pre-coded widgets to manage
things like workflow (CMS-speak for making sure you don't get anything
past your editor or your boss), revision control (sophisticated
Undo that reaches back forever), and live editing (fix your site
from inside your site). What presenter and engineer George Runyan
made the biggest deal of was the graphic design: "editors and content
creators don't want to use something ugly." I'm not saying that
nothing else at the show had any impressive aspects - Wyona does
neat tricks for inline WYSIWYG page editing that I've got to learn
more about, and other packages like Midgard and DBPrism offer aggressively
abstracted XML flows suitable for plugging into everything from
databases to printing presses (as does Plone, probably). But the
chief differences between most packages, and the bulk of the content
of their talks, were implementation details: WebDAV, Cocoon, PHP,
model-view-controller, various technical alphabet soups and component
stews. The Plone team grasps that this stuff isn't going to change
the way we communicate until it looks and feels more approachable
for non-coders.
If we'd seen more hands-on demonstration of Xoops and PostNuke,
they might have fit neatly into that category as well. PostNuke
is what gets called, sometimes dismissively, a slashclone - a software
package designed to run sites that end up looking by default very
much like the wildly successful tech-news site Slashdot. PostNuke
itself splintered from, and has mostly overtaken, a package called
PHPNuke, and Xoops (there doesn't appear to be a standard way to
pronounce that, sorry) derives from PostNuke in turn. This kind
of profusion is both a benefit and a drawback of open source, but
the confusion about what's what, and what's good, seems to be waning.
The point is that the Slashdot model - a webloggish front page with
long spurs of commentary off each post, and many categories by which
to organize the news - is adaptable to things that look much more
like, well, that content that you want to manage.
Michelle Alexandria was a fairly unusual person at the show - the
only female presenter, the only non-white presenter, and significantly,
the only presenter with a solid background in the non-techie communications
industry. When she isn't the documentation head and sort-of-product
manager for the open-source Xoops, she's the webmaster for the entertainment
magazine Eclipse. Eclipsemagazine.com uses Xoops, whose lead developers
all live overseas, to maintain a bustling site with multiple news
flows and a large forum. Alexandria talked about how content management
tools that not only include community but leverage it, to bring
in more content and do quality control, utterly changed what she
did, to the point where she got involved in the Xoops project.
Gregor Rothfuss of the PostNuke project used his time to talk generally
about the role of community in a content site. Rothfuss was also
one of the attending members of the OSCOM board. I couldn't help
but notice that out of all the presenters, the only ones who attempted
to do anything other than a strict dog-and-pony show about their
product were OSCOM organizers. Michael Wechner's talk on Wyona was
rather brief because he wanted to make time to talk about peer-to-peer
technology - the crowd seemed confused by the conflict between the
free-for-all models of common P2P apps, and the top-down nature
of most CMS architectures. (One attendee pointed to Groove Networks'
eponymous non-free platform, which uses the tools of peer-to-peer
without making it part of the user model, and could indeed have
CMS tools built into it.) I already covered Nessom's talk. Why was
there so little talk about what the forms we build for people to
pur themselves into are going to be, and so much talk abolut how
to build them?
It may just be that before open-source content management can
get creative, it has to get a foothold. That's what Friday's Interop
sessions were all about - the necessary nuts and bolts that all
these different frameworks have to have in common before they can
do things like have a standard Mozilla interface, for example. Paul
Everitt of Zope Corporation spoke eloquently about the challenges
and rewards of interoperability in his earlier Zope spiel. "The
client side is where interop gets rolling." In keeping with Plone's
focus on the end user, Everitt concentrated on motivating CMS coders
with real user goals. "Mozilla's not close enough to where people
live and do their work. Let's get into Word."
When I spoke to him, Rothfuss agreed that CMS had to get on firm
technical ground before innovating, but he put an interesting spin
on it: "We have to get social interop first. We want to get people
talking to one another. Then, we want to have something to show
for all the talk." Nesson elaborated on the same theme: "This morning
Eric [Wiseman, board member] asked me what I was thinking, and I
said, 'I'm trying to figure out what they all think content management
is.' " That's a kind of interoperability dialogue too - seeing if
the mental models fit. "They're all coders. They are only just realizing
that they're in the communications business. Whereas I'm coming
from education and communication and have only recently come around
to code." Communication precedes creativity. Hopefully, the next
few years will see CMS developers come to terms about the details,
so they can start talking about new wants and needs - then we users
can get some new communication tools to be creative with.
bio:
Mike Sugarbaker is a
writer, coder and layabout based in Berkeley. He previously reviewed
Lawrence Weschler's Lemon for
Mindjack.
|
advertise
here
email for info
|
