[thelist] newspaper community site/structure

Hugh Miller hmiller at cfpress.co.uk
Wed Nov 26 11:15:56 CST 2008


Bob,

I work for a UK regional newspaper firm. We do a similar thing with a 
portal site that leads into the various local newspapers and from there 
we go into further localisation for the newspapers.

ie.
http://www.newsportal.com/newspaper/[news]/town/localarea/

This system is also set up with (for example);
http://www.newsportal.com/newspaper/[sport]/type-of-sport/teams/
http://www.newsportal.com/newspaper/[opinion/leaders]/blogs/bloggers/

It's all done in PHP / mySQL with a custom CMS rather than an off the 
shelf product so I can't be too much help there, but I'd certainly be 
happy to explain the basic way it tracks back and forward through the URL's.

Articles are published to various subsections which are controlled by 
allocating a parent ID to each one, so using bloggers as an example we get:
OPINION (section code 3, parent code 0)
BLOGS (section code 305, parent code 3)
BLOGGERS (section code 3050, parent code 305)

The numbers are not relevant I just use a style to make it easy to think 
through in my head, they can be any value.

Content is syndicated as required using a series of titlecodes, with a 
basic structure of town / county / national / international 
(international exists because we publish in the Republic of Ireland as 
well as UK, but it could equally be used for publishing relevant 
content). So depending on where you are in the system your SQL query 
requests content from related titlecodes which are stored for each title 
as titlecode, regioncode, nationalcode, internationalcode (I find it 
easier to have individual titles under number 500, regions 1000+, 
nationalcodes 1100+, but they could be anything). So the portal can take 
content from any area, but individual titles can only get syndicated 
content to display.

The load is spread across several databases - current archives hold 
about 40,000 articles, with a central one being used for site config and 
sections tables, one for all commercial tables and 5 for editorial copy 
(England, Ireland + 3 Scotland), Syndicated copy is handled by a 
cronscript rather than the CMS. Localised copy is automated via a parser 
in the CMS which looks for street names within the copy.

I also use a date format in the querystrings for all copy, so you get:

http://www.newsportal.com/newspaper/news/town/local/YYYY/MM/DD/articleID-headline-in/print

Which while good for SEO is also really easy to handle archive browsing 
and searching.

Regards,
Hugh




Bob Meetin wrote:
> Just a little brainstorming here.  I have a client-in-waiting who has a 
> newspaper, actually part of a community of newspapers.  Example:
>
> Visit www.newspaper.com
>
> At the toplevel you get general news/such pertinent to all (say for the 
> county). Then via a drop down menu you select a more localized smaller 
> newspaper, say CityA or CityB, and you get specifics for that city, 
> news, advertisements, etc.  So the jump would take you to possibly:
>
> www.newspaper.com/CityA
> www.newspaper.com/CityB
> www.newspaper.com/CityC
> etc...
>
> This structure is not in place yet, but this is the goal.  We are 
> probably going to set this up using Joomla or similar.  First of all I 
> suggested it would probably be better to set up the structure now vs. 
> retrofit.  If any of you have done something similar I'm entertaining 
> any gotchas in advance.  Thx.
>
>   

-- 
This message has been scanned for viruses and
dangerous content by MailScanner, and is
believed to be clean.




More information about the thelist mailing list