[kDev] Developers, Project Managers

["Dave Cridland [Home]" <dave@xxxxxxxxxxxx>]

Morning.

I'd be interested in either playing with the open source Kendra work, or
else getting paid to actually do some work as and when required.

For the others - I'm assuming that Daniel vaguely remembers me, as
hopefully does Kirit - I've worked on mission critical stuff for pretty
much all of my career, and worked on various collaborative tools before,
both as part of my work and as a habooy well before I started earning an
honest crust.

I'm confortable with describing myself as an expert on various aspects
of the Internet and Internet development, and I'm reasonably conversant
with most of the open standards that exist.

CV on request if you really want it, or else use Google to search for
"Dave Cridland" (in the quotes).

How I would aim to achieve the goals:

>1. Tools
>       * Profile
>         Users contact information such as addresses [each entity
>         stands in its own right so that many people can be employed by
>         a company].

On the one hand, LDAP would be ideal for a formal directory. On the
other, Consumers don't fit into a content directory, so perhaps an
alternative technology would be better, perhaps ACAP's addressbook
format. Perhaps both would be for the best.

Export format could be vCard/iCal.

>       * Calendar
>         Events and meetings.

There is an open standard for this in development - CAP - but it's been
through more radical changes than Madonna. The remaining alternatives
are using ACAP again, or developing an in-house system.

Obviously this needs to be capable of export, iCal/vCal should be fine
for this.

>       * Forum
>         Wiki, Majordomo, web bulletin board. Email lists with web
>         archiving like Yahoo but better. Need to know which emails are
>         bouncing too.

Some form of MLM, archiving either to web (pipermail, hypermail, etc),
or (my personal choice) to publically available IMAP mailboxes, with a
sane web front end.

>       * Funding
>         Connecting people/orgs to funding organisations for their
>         Kendra Participant Projects.

Not entirely sure I follow what this means in terms of technical
requirements.

>       * Accounts
>         All financial accounts will be displayed on the website.

So long as nobody expects a GL system to be built for Kendra, this
should just be a matter of sticking some Excel (or whatever) files up.

>       * News

Stored "somehow", exported using RSS.

>       * Events And Meetings
>         People organise meetings. Link in calendar. Conference
>         organisers enter info about their conferences. This then has
>         to be rated to make sure it passes criteria. Link in calendar.

Rated by whom? What criteria?

Linking things into the Calendar should be easy.

>       * Development
>         Need a development area.

In what sense? We could run our own Sourceforge, of course, and most
likely combine this with some form of PM software - there's
http://core-lan-org.sf.net/ which I've been involved with in the past,
and seems to be well regarded as a simple PM system that covers the
basics well.

The rating of [meta]data seems to be quite a key feature for Kendra. If
the intention is to provide data and metadata "directly" from the public
to the public, then obviously "trash" data will slowly amass, leading,
potentially, to a very unworkable system, since the bad data will
dissolkve the good.

So instead, I'm thinking that, while anyone could enter metadata that
they considered right (or indeed deliberately wrong), those reading the
metadata could equally well rate it. Rating ratings is even possible.

And here's an interesting idea. Consider three users, Alice, Ben, and
Charlie:

I could, for instance, decide that Alice's tastes are completely
different to mine, whereas I've found that Ben tends to enjoy the same
content.

If I were able to tell the system, then it could put a larger weight on
Ben's opinions as compared to Alice's, so I would see metadata and
ratings that were more likely to be aligned with my taste.

If Ben then decided that Charlie's opinions were in line with his, then
I might see Charlie's opinions with a higher weight, too.

In programming terms, given the following rating of a particular piece
of content:

		Their Rating		My "trust", out of 10.
Alice:		80%			1
Ben:		30%			9
Charlie:	20%			8 (From Ben trusting Charlie with 10)
David:		50%			5 (The default)

I might see a rating of:

1) The "general" rating, a straight mean: 45%
2) The "adjusted" rating, a weighted average: 33%

It strikes me that this could be much more useful - it's essentially
similar to Amazon's "Customers who bought this also bought...", and some
automation of finding people who's opinions aligned best would obviously
help, probably based on my own ratings.

I suppose I should really put something together as a demo of this
concept at some point, but I'm hoping people have got the general idea.

Dave.