[geeklog-devel] Google Summer of Code 2009: Geeklog

[saurabh gupta]

On Fri, Mar 13, 2009 at 12:41 AM, Dirk Haun <dirk at haun-online.de> wrote:
> saurabh gupta wrote:
>
>>The implementation part mainly consists of parsing of xml
>>files. I have already worked on the parsing of xml files using
>>libraries like msxml and nsxml (however not open source) and some of
>>my own wrappers.
>
> Geeklog also comes with classes to read and write feeds in various
> formats, so unless they are missing something for this project, you
> wouldn't have to step down to that level.
>
>
>>For example, if a blacklist item has the title as
>>"casino", then a new entry under the item for "casino" will tell that
>>how many posts or spams have been recognized by this keyword "casino".
>>This will have the benefit of determining the validity of any
>>blacklist item.
>
> Not sure what exactly this would add? The concept is called "web of
> trust" for a reason - if you find someone adding useless rules, don't
> use their feed.
>

all right. But I just wanted to say that while subscribing to a site
for spam blacklist, some modification can be put to choose the some
selective items and not the whole blacklist. Currently, only the
distance (i.e hops) is only the measure to choose the items from the
blacklist. More options can be given to the site admin in that case.

>
>>The overhead in this idea is
>>that for each post which is recognized as spam, the RSS feed file is
>>to be updated.
>
> The impact, e.g. during a massive spam wave, could be significant. For,
> IMO, very little added value.

yes, you are right.



-- 
Saurabh Gupta
Senior,
NSIT,New Delhi, India