Experimental p.o.a site is up

[Ross Gardler <ro...@gardler.org>]

[CC'd to Forrest dev for info]

Thanks to some help from David Crossley we now have an instance of the 
*experimental* forrest geenrated p.a.o site up and running on the 
Forrest Zone.

Take a look at http://forrest.zones.apache.org/ft/build/apache-projects/

I still need to integrate the work on the XSLs and RDF extensions that 
others have done over here, so still only the Forrest and Projects DOAP 
are being processed.

How is the site updated?
------------------------

Currently it is rebuilt every 3 hours via a cron job. The build can be 
triggered in from any script.

What happens if there is a problem with a site build?
-----------------------------------------------------

Broken links and other such problems with a site build will be emailed 
to nominated email addresses. In this way any problem with the site is 
autmoatically reported. At present these come to David and I, but it 
would be more sensible to send it to an apropriate list.

How do we add a new project?
----------------------------

Simply add an entry to a locationmap file. This describes where all the 
projects DOAPs are, the idea is that they can be anywhere that Forrest 
can access (SVN, CVS, HTTP, file system and pretty much any  other 
protocol.

Do we have to manually maintain this file?
------------------------------------------

At present, yes. However, if we define a way of automatically 
discovering where these DOAPs are then it can easily be automated.

How do we add projects to the indexes?
--------------------------------------

If the index already exists, for example, the site currently has a 
"Java" and "All" index then you need not do anything. Any DOAP that has 
the relevant information to make it part of the index (e.g. have "Java" 
in the programming languages list) will automatically be included in 
that index on the next site build.

How do we add an index?
-----------------------

Create an XSL that will process the full list of projects, add a 
processing step to the projectInfo plugin in Forrest that uses this XSL 
and add a link to the output in the site.xml file.

This process can be simplified by more work on the projectInfo plugin, 
but for now I am happy to do this work manually (especially if someone 
provides an XSL).

Where do we go from here?
-------------------------

I'd like to put the config for this into the site-dev repository so that 
others here can see just how easy this really is. The Forrestbot that 
publishes the site will automatically retrieve updates

The skin is CSS based, we can customise that for colours etc.

Add more indexes.

Add more structure to the site.

????

Any questions? Observations? Requests?

Ross

Re: Experimental p.o.a site is up

[Ross Gardler <rg...@apache.org>]

Ross Gardler wrote:
> Ross Gardler wrote:
> 
>> Take a look at http://forrest.zones.apache.org/ft/build/apache-projects/
> 
> 
> Since there seems to be a problem with the forrestbot built site I've 
> built the site manually for you.

(sorry, wrong list)

Ross

Re: Experimental p.o.a site is up

[Ross Gardler <rg...@apache.org>]

Ross Gardler wrote:
> Take a look at http://forrest.zones.apache.org/ft/build/apache-projects/

Since there seems to be a problem with the forrestbot built site I've 
built the site manually for you.

See http://people.apache.org/~rgardler/apache-projects/

The skin (including all adornments such as breadcrumb trail, last 
modified timestamp, search box etc.) are all under the control of a 
simple skin configuration file and can be turned on/off etc. Everything 
else is CSS. So it's really easy to change the look and feel of the site 
to suit whatever design is required.

There's ony one index in there at the moment (by language). Others are 
very easy to add, however,

I believe it is time to make a few technical decisions as to how we want 
to proceed with site-dev. The perl route or the Forrest route (or even 
Maven if there are still folk here proposing Maven for this).

We should have that discussion in another thread when everyone has had 
time to look at the different implementations.

Ross