delicious linkrolls

Time to revise this page. It was originally created to document the runes on how to use the delicious provided javascript linkroll generator, it then covered the slow and painful death of the service and my attempts to postpone the inevitable.  In late 2019, I found my backup copy and posted it to the blog,

Is this my delicious linkroll?

There are some problems with this, most obviously the tags are hidden. I have written some code in python, to parse the html and create a python dictionary, and also some code to build a tag index. I have some code to see if the linked sites are still active but there’s over 3000 bookmarks at it takes some time, and doesn’t distinguish between those that have gone away and those that are being resold. (I need to go through them for continued relevance also.)

The links are held in a pickle database, which could easily be converted to XML or JSON. I have code to delete a record on the basis of a numeric surrogate key. I can build two indices, on has the key and the tags, held as a string, the other has a list of key-tag pairs.

  • Document which sites no linger work and/or are no longer relevant.
  • Define the output formats, html, rss or an active query and if the latter how to host it.
  • Document the delicious date format field
  • Write some restartable HTML status checking code.
  • Amend the index generators ro  write to a file, so there is some persistence

The slow death of delicious’ rss feeds

I originally saved the javascript linkroll code as text in a page, and latterly, from Oct 2013 used Artiss Embeded Code and I located the scripts in code objects. delicious withdrew the javascript. I implemented JP’s Get RSS Feed to fix the withdrawal of the delicious javascript;

I rendered the feed using wordpress. At some time delicious changed their URL from back to At the same time, they decided to insert an advert into the first item of the feed.

This has led to all my bookmark feeds in this wiki breaking, and to asking the question as to whether one wants to fix it as I no longer control the content. I have posted to the wordpress support site to implement a “start from” parameter. I wonder if CSS display:none; properties could be used to supress the first element in the feed.

This wiki has a number of pages which used the feed to expand the articles and to point at my bookmarks. I haven’t decided whether to fix them or not.

I have asked that the plugin is adjusted to exclude the first element and I propose to examine planet to see if I can put a filter into the consumption code lines. See for my planet documentation.

See for my jp-rss-feed documentation and for Jimmy’s documentation.


Tagging Engines

While is a fantastic social bookmarking site, its not easy to implement inside a firewall, which needs to be done if the URL list to be tagged is an intranet resource.

Open Source alternatives

Here’s some alternatives,

which needs from

  • Glassfish

Installing Scuttle

  1. Download WAMP – this works pretty much out of the box, only problem is if IIS is already running in which case you need to either let Apache use another port, or turn it off. Use phpmyadmin to create a new database and a user with rights to use the new database. Its best to sue the same name as the O/S user, document the password.
  2. Download Scuttle – this is a zip file and I had problems with Windows Vista as it has problems dealing with the file hierarchy names. You can download your favourite evaluation version of a tool, or I downloaded cygwin which since I used it to run the .sql file, I found useful and there is no problem with license and uncompress | tar -x works fine.
  3. Download Cygwin – Not exactly mandatory, but I prefer it, every windows systems should have one.
B$ export PATH=${WAMPHOME}/bin/mysql/mysql${VERSION}/bin:${PATH}
B$ cd ${WAMPHOME}/www/${scuttlehome}
B$ mysql -P$DBPASSWD < tables.sql

Configure PHP as a database client. Copy all the scuttle files to an appropriate location with the web server’s public file system. ${ROOT}/scuttle is good, but I don’t use it. Copy to, and amend the user, password and database values to the ones you have used and set. Add the line


to this file, otherwise you get a bunch of forward compatibility warnings.

That’s enough I’ll write up my usability tests another day, I expect to repeat some of these notes on the project’s sourceforge pages. but didn’t!