Re: RFC: Syndicating PEAR Weekly news?

Jesus M. Castagnetto wrote:

(CC'ing to the list too) Will start work on an input system this weekend. Definatly +1 on RSS feeds.

but please dont do an complex input system - it takes roughtly 2 hours to review and write the weekly news. (which got very tight this week as I was busy all friday on other stuff) - html forms are great for quick comment systems and simple field based entry, but are lousy at publishing systems. - (my current research is on webdav which will finally kill the nasty kludge of typing in textboxes for publishing..) uploading by coping/pasting the raw HTML into a text box is possible, but not really an improvement.. realistically - when I write it , I jump around many sections, then do a final proof, upload, check the list of released packages, then do the last tidy up. you should be able to tokenize on the '<', look for H1/H2/H3, BR etc. and break the document that way. - use the timestamp on the original HTML file(eg. CVS exported) to determine if the database needs updating.. That was the original idea. Overlaying the parsing with a pregmatch of the package names (and replacing with links would be another nice trick..) having done the parse to generate RSS/database, you could then reuse this to do the final HTML/email etc. It may be usefull for the translations / eg. translation of each paragraph. to use forms. I guess the RSS will probably look a bit like <item> <title>Discussions on APD profiler</title> <link>http://pear.php.net/en/20020908/weeklynews.php#01</link> <description>the clear text view of the text using striptags</description> <content:encoded><![CDATA[ the raw html that starts after <h1> and ends when <BR or <H1 is reached. ]]></content:encoded> </item> regards alan

--- Pierre-Alain Joye <paj@pearfr.org> wrote:

On Thu, 29 Aug 2002 14:05:05 -0700 (PDT) "Jesus M. Castagnetto" <jcastagnetto@yahoo.com> wrote:

I can imagine a simple web based form where Alan (and/or the translators)

put

the text for each new item, and valid RSS is generated automagically. Which

can

always be edited by hand by people with XML-warped minds and fingers ;-).

Then

once the news feed is aproved and ready to go, HTML is generated, an plain

text

email goes to pear-general, and it is copied to a canonical URL, e.g. http://pear.php.net/pwn.rss (for PEAR Weekly News)

Which format for the internal datas (links to package, users) ? keeping them as html link can do it.

We can do it two ways: 1) Use CDATA to include them all e.g. <description>An Foo depends on <![CDATA[<a href="http://pear.php.net/package/Bar">Bar</a>]]> for more complex printing</description> or 2) Use a namespace like the one for HTML or similar (DocBook, XML Link, etc.), or even the DC (Dublin Core), the Syndication, and/or Content. The last three are supported as extension modules in RSS 1.0 e.g. <item> <title>The Foo package</title> <link>http://pear.php.net/packages/Foo</link> <description>Useful when you need all those "foo" in your code</description> <dc:relation name='requires'>Depends on the Bar package</dc:relation> <content:encoded><![CDATA[See also: <a href="http://pear.php.net/packages/Bar">Bar</a>]]></content:encoded> </item> http://groups.yahoo.com/group/rss-dev/files/namespace.html I wrote the above manually and sort-of from memory (was talking about RSS recently that's how I learned about those modules), so I might have gotten one or more things wrong.

Any ideas, comments, better names for the WOGLON index? ;-)

+1 pa

===== --- Jesus M. Castagnetto (jcastagnetto@yahoo.com) Research: http://metallo.scripps.edu/ Personal: http://www.castagnetto.org/ __________________________________________________ Do You Yahoo!? Yahoo! Finance - Get real-time stock quotes http://finance.yahoo.com