I heard Mitch Kapor today at Stanford talk about Chandler. I sat next
to Adam Cheyer (longtime friend of Doug and Bootstrap, and brother of
Jon, who's on this list), who turned to me after the talk, and said,
"There doesn't seem to be much there." I didn't feel quite as
negative as Adam did, but I was a bit underwhelmed. It might just be
the side effects of the hype machine. (01)
Kapor stated up-front that their primary goal is to make sharing PIM
information easy, not to make a kinder, gentler Outlook. The basic
design borrows heavily from Lotus Agenda, Kapor's first PIM baby.
Chandler will be written in Python with wxWindows as the UI toolkit
and an RDF-based repository. They have not resolved the license yet,
but it will most likely be dual-licensed, similar to the MySQL or
Sleepycat license. In other words, those who plan on incorporating
Chandler into a proprietary product will have to buy a license from
OSAF; everyone else will be able to use and modify it for free. (02)
Two elements bore great resemblance to the OHS: the RDF-based
repository and a View Construction Kit. It uses a graph-based data
model to represent all of the information in the system. Everything
is what Kapor calls "items" -- e-mails, calendar entries, contact
information, etc. Items have properties, and they can be linked,
annotated, and dynamically viewed. Custom views can be constructed
and reused, and the system will come with a bunch of default views.
The system is designed with extensibility in mind, and an alpha
release is expected in April. (03)
I think Adam was unimpressed by the promised feature set -- there
didn't seem to be much there that wasn't already in Outlook/Exchange.
I mostly agree with that assessment, but I will also be happy to see
all of this functionality available on Linux. I also think there's
much to be said about Kapor's cost-of-ownership argument. For an
individual or a small business, the cost of setting up
Outlook/Exchange is unnecessarily high. (04)
I didn't notice anything particularly innovative under the hood -- no
granular addressability, for example, which I think is unfortunate --
but I think that it will benefit from being built from the ground up
using only open standards. I also think that overall, this is a very
positive thing for the open source community. We're going to have a
greater selection of open source applications to choose from. We're
also going to see major improvements in various open source toolkits
-- wxWindows, ZODB -- because Chandler needs these improvements, and
Kapor is willing to fund them. (They currently have 14 full-time
people, nine of whom are paid, with more to come.) (05)
What impressed me most about Chandler was the approach. If Chandler
offers everything that it promises, then its underlying repository may
end up being its most important contribution. I view Chandler as a
repository implementation disguised as a PIM, similar in spirit to
Helium's relationship to KnownSpace. I think this approach would have
been a far more effective route towards convincing people to fund OHS
development. If you promise people a whizbang PIM that's going to
change their life, that's something that people can wrap their heads
around, and it gives you the excuse to build an OHS data repository,
which is the real innovation. Unfortunately, that's not the current
strategy. While Chandler's architecture obviously borrows a bit from
Doug's ideas, it only skims the surface. There is much, much more it
could do (again, granular addressability comes to mind). (06)
-Eugene (07)
--
This message is archived at: (08)
http://collab.blueoxen.net/forums/cgi-bin/mesg.cgi?a=tools-yak&i=20030130083030.GB7651@douge.blueoxen.net (09)
|