community-dev mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Shane Curcuru <>
Subject Re: Streamlining new committers info: projects.a.o
Date Sun, 11 Nov 2012 13:17:16 GMT
Yes, the grand plan for projects.a.o is to have a single place that 
provides an overview of all Apache software products (as well as Apache 
projects too).  This is an idea that needs some people to help drive it 
to both ensure it's updated, as well as to provide a better UI for users.

Imagine having the front pages and categories/search of ohloh, 
sourceforge, and Google Code rolled together, but just displaying Apache 
projects.  It'd be a huge win to let the public know what we build.

David R. and others (a while back) created the projects.a.o code (that 
uses a mess of different technologies, including perl and XSLT on a cron 
job, which means it's hard to decipher how it works 8-() that actually 
scans the DOAPs and generates the website, with various indices.  This 
is my main goal - to improve the usability of this site, both with a 
nicer UI (with introductions explaining what it is to average users) as 
well as with better data (for example, with an improved taxonomy as 
suggested below).

Now, actually improving the taxonomy is a two step process, that I'd 
love to see energetic volunteers working on:

- Figuring out an XML-compliant way to improve the taxonomy(ies) 
available in the DOAPs.  We need a 'good enough' solution, not a perfect 

- Working with projects (or simply doing it ourselves and submitting 
patches to projects) to actually update all the DOAPs with these 
improved taxonomies.  We can't rely on all projects being diligent about 
updating their DOAPs - experience has shown, some projects just never 
get around to it.

Using DOAP files also means we can provide a simple API (or set of 
URLs/REST service/whatever) for external entities to come crawl 
themselves, if they want to provide their own mapping of Apache 
software.  That's one of the reasons that providing a DOAP file is a 
branding requirement for all Apache projects.

* A nice-to-have would be a simple set of public instructions for people 
to figure out how to scan our set of DOAPs themselves.

Sounds like it was a great ApacheCon, and hoping to get to know some of 
you working on projects.a.o!  8-)

- Shane, sadly based in the US without funding to travel these days

On 11/10/2012 7:44 AM, Tammo van Lessen wrote:
> Hi,
> On Fri, Nov 9, 2012 at 4:14 PM, Gavin McDonald <>wrote:
>>> -----Original Message-----
>>> From: Igor Galić []
>>> Sent: Thursday, 8 November 2012 3:32 PM
>>> To:
>>> Cc: ASF Site-Dev
>>> Subject: Re: Streamlining new committers info
>>> Another idea that just came up in a BoF was to create more useful info on
>>> the *main* page about what the projects do and how to get the one you
>>> need: i.e.: I'm looking for a Database - it should be a Key-Value store,
>> in Java.
>>> And you're down to 2 projects: DB, Cassandra
>> Sounds like what's on [1] already.
>> Gav...
>> [1] -
> I like the idea to improve this aspect on the main page as well as I also
> have difficulties to understand which projects does what and how it
> compares to others (especially in the hadoop incubator ecosystem). I think
> DOAP (and semwebtech) is well suited to do this job, but what is IMHO
> missing right now is a taxonomy of categories. Currently, categories are
> flat and used as tags. By using something like SKOS, we could define a
> tree-like vocabulary of our projects' categories, e.g. a "Key-Value store"
> is a narrower term for "Database". See [1] to get an idea of such a
> taxonomy. If the doap files would then reference the most narrow category,
> Igors example could be answered easily datawise. UI-wise, the data could be
> materialized and then visualized, either graphically or with something like
> Exibit [2], see [3] for a demo. Just an idea, I'm happy to help if time
> permits.
> Best,
>    Tammo
> [1]
> [2]
> [3]

View raw message