forrest-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Stefano Mazzocchi <stef...@apache.org>
Subject Re: [RT] my Forrest dream-list
Date Thu, 21 Feb 2002 10:21:55 GMT
Steven Noels wrote:

> > My idea is to use the 'command line' features of Cocoon to generate
> > static snapshots of sites and produce mirrors directly using a sort of
> > rsync or other mean. Anyway, the goal is to make sure mirroring is as
> > easy as possible. No, easier than than that.
> 
> Doing CLI where appropriate sure would alleviate the issues I listed with the
> graph logs in my previous message: +1

Yes, I've looked into the way PHP.net does mirroring and they basically
*clone* the entire site using rsync.

The easy thing for them is installing apache+PHP is a breeze compared to
installing java+apache+tomcat+cocoon (even if we could prepare prebuild
distributions of everything... such as a 'apache-mirroring-kit' or
something.

Worth thinking about.

> > Ok, but what should Forrest give us:
> >
> > 1) coherence: all the site should look coherent, same graphics, same
> > look&feel, same information in the same locations, coherent and nice URI
> > space (build to last! so that broken links are reduced!)
> >
> > 2) structure: the site should look professional, the structure should be
> > flexible enough to fit every need but solid enough to guide users
> > browing and developers adding resources
> >
> >
> > Ok, but the best thing is 'functionality': everything you wanted to have
> > in your project web site.... here are the things I've been missing:
> >
> >  1) logs and community rating: I want to have numbers to judge the value
> > of a community/effort
> 
> Downloads, number of commiters, numbers of people subscribed on the users/dev
> list, numbers of mail messages / week on these lists, CVS activity...

Exactly. Sam and I exchanged visions on this privately and we came up
with exact the same numbers.
 
> The kind of parameters I use to 'judge' an Apache project I haven't been looking
> into before. Unfortunately, not all of those are easy to get hold off.

Downloads are easy, just process the apache logs.

Number of committers is just as easy, process the /etc/groups file

Number of people subscribed i think it's not hard, Sam knows how.

Number of mail messages shouldn't be hard when we implement our own mail
archive, before than, we can 'screen-scrape' MARC (which graphics
doesn't change that often :)

CVS activity could be measured with the above method: couting the email
messages on the cvs lists. (since only CVS can post on them)
 
> >  2) built-in search engine: users much have a simple way to find things
> 
> Lucene sure is an option, but maybe we can bug Google to index the
> xml.apache.org pages (even) more often, and reuse that search functionality
> instead.

No, I'd love to keep in internal because tomorrow this could give us
potential playground for more semantic web technologies. I'd say lucene
is the way to go for now. (besides all the machinery is already in
place)
 
> >  3) user-writable pages: people should have a way to add things to the
> > pages.
> 
> As in WikiWiki? Hmmm.
> As in 'Comment on this page' deferred to some discussion board?
> 
> I'm quite opinionated (as Stefano already knows ;-) on the how and when of
> in-place-editing XML content across a web interface, so please explain what you
> mean here.

Ok, look at the PHP web documentation (not the print-friendly version,
the other one): if you think there is something wrong with the page, or
have something more to say about it, you just go there and add your
personal comments to it.

Don't think of some slashdot-like forum thing, no, just a way for users
to "annotate" on the side of the pages, much like you do with your
pencil on your books when you study.

Jacob Nielsen says that the web lacks the functionality of paper that
you can always write on it, even if already printed.

My dream would be to be able to let users *take notes* ON their pages,
using some form of inline editing. Write comments, add sticky notes,
correct mistakes... and then a very simple workflow system could allow
us, developers, to know *what* the people changed and consider merging
the changes, the annotations, or simply their comments directly into the
page.

It would *ease* by a couple of order of magnitute the rate of
documentation production for OSS projects, because my personal opinion
is that it's not true that people don't like to write docs, just that
they are scared by the amount of 'structure' they have to give to their
content.

on the other hand, when you take notes, or write email, there is no such
structure.

So, it's way faster to exchange information that way.

The idea is to make it easier for people to write stuff, removing their
impression of structure-need when they are writing.

But this really *is* a dream because only IE 5.5+ on windows currently
has the functionality that we require on the client in order to do this.
 
> >  4) short bios for the project committers, with pictures: gives a better
> > sense of community. Having a weblog there would be great.
> >
> >  5) news, announcements, events and project calendaring: the news and
> > the annoucements will generate a RSS feed, events will be aggregated
> 
> You can count me in for the RSS stuff.

Cool.
 
> > into a project-wide calendar, or an effort-specific calendar, the events
> > will also be overwritten on the logs, to indicate how logs were
> > influenced by the events (say, a release or a new committer added)
> >
> >  6) book-like PDF versions of documentation: easy to print out docs.
> >
> >  7) integration with GUMP: dependencies, runs, committer that broke the
> > run
> >
> >  8) javadocs seemlessly integrated with the documentation and with the
> > search engine
> 
> Java Doclet? http://www.sun.com/software/xml/developers/doclet/

Yes, exactly. I'll look seriously into it.
 
> >  9) mail archive seemlessly integrated with the search engine (I should
> > look for something and get results from either docs, javadocs or emails
> > messages)
> >
> >  10) coherent-looking CVS view
> >
> >  11) document translation using Google translation services
> >
> >
> > Ok, I think it's enough for now.
> >
> > throw your comments in the lake.
> 
> Swimming :-)

Cool, nobody else?

-- 
Stefano Mazzocchi      One must still have chaos in oneself to be
                          able to give birth to a dancing star.
<stefano@apache.org>                             Friedrich Nietzsche
--------------------------------------------------------------------



Mime
View raw message