From: Josef 'Jupp' SCHUGT <jupp@...>
Date: 2004-01-15T02:56:09+09:00
Subject: [ANN] Release of rubric 0.3

Hi!

I am proud to announce a new release of rubric, the Ruby-only news
feed aggregator.

Version 0.3 is a major rewrite of rubric. Quoting the changelog which
is implemented as a RSS news feed that is available at

    http://rubric.rubyforge.org/rubric.rubyforge.org.rss

> The most important change since 0.2.1 is that rubric now uses a
> database. The database is implemented using an index file and a
> database file.
> 
> The index file contains the date a news item was added to the
> database and its MD5 hash.
> 
> The database file contains the MD5 hash of a news item and the
> compressed and base64-encoded (except that line feeds are
> replaced by spaces) item.
> 
> You perhaps wonder what the MD5 hash good for. A cryptographic
> hash like MD5 is a 'genetic fingerprint' of the news item. The
> likelihood that two news items have identical MD5 hashes but are
> different is smaller than the likelihood that a genetic
> fingerprint matches by any chance.
> 
> This property of MD5 is used by rubric. The database file only
> contains the news items that are needed to generate the portal
> page (default hold time for this data is 3 days). The index file
> on the other hand contains the hashes of the news items to avoid
> that data is outdated news items are added again - which is a
> known problem of earlier versions of rubric. The default hold
> time for the data in the index file is 4 weeks (28 days).
> 
>   * New config options Hold_time_items (time in seconds to hold
>     news items) and Hold_time_hashes (time in seconds to hold
>     their hashes) have been added.
> 
>   * Config file options Input_limit and Output_limit are no
>     longer supported.
> 
>   * Input_limit and Output_limit are no longer supported.
> 
>   * Portal page now advertises HTML version used as 4.0
>     transitional.
> 
>   * HTML now uses description lists in place of tables.
> 
>   * Support for CDATA has been added.
> 
>   * Support for broken CDATA has been added (because some feeds
>     use it without knowing what it really means).
> 
>   * This feed now uses HTML and CDATA.
> 
>   * rubric now accepts items with missing titles.
> 
>   * Parsing of RSS feeds has been improved.
> 
>   * File names for local copies of feeds did change to base64
>     encoded representations of the feed URLs (exception: '=' is
>     transformed into '_' and '/' into '-', line break is removed)
>     with an added '.xml' suffix so that every feed has a unique
>     filename.
> 
>   * rss_fetcher now comes with a version number.
> 
>   * Instead of 'host', 'path', and 'port', you now can use 'url'.

BTW: rubric is what I had in mind when I recently posted my data
compression questions.

Now then[1],

Josef 'Jupp' SCHUGT

[1] Does one really use that? In German it would be 'nun denn' and in
    Japanese 'sore de wa'.
-- 
http://oss.erdfunkstelle.de/ruby/    -     German comp.lang.ruby-FAQ
http://rubyforge.org/users/jupp/     -     Ruby projects at Rubyforge
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-
Germany 2004: To boldly spy where no GESTAPO / STASI has spied before