I used a similar technique ten years ago, setting up msaccess as reporting tool for SAP R/3 with RFC dll’s and ActiveX. That remained stable for eight years without maintenance. Hey, I might get lucky with WordPress xml-rpc and xhr/ajax.

I called the example RpSequel.

adding sql rpc functions

To sync the data sets, I will duplicate a list with sequel operations from my desktop database as rfc call to my blog’s xmlrpc-endpoint. To handle that list, I plug a sql crud method into the xml-rpc method array in WordPress :

The basic INSERT method itself can be simple:

That in itself is not very exciting, it stuffs records in the host’s database. Which is fine, however, I want the blog to respond to individual operations.

I can write rpc-functions for every single function I want the blog to perform, that means truckloads of rpc-functions, on both client and server end. I am incredibly lazy, so I ain’t gonna go there.

adding hooks to sql rpc functions

Lucky for me, WordPress has hooks, hooks are cool.

Adding action hooks (before_insert and after_insert) to the crud method makes it more powerful. Two hooks are enough to separate the business logic of the desktop database from the blogs reporting logic.

Now it is more exciting.

• I can send a list of records and a rfc insert-method to the blogs xmlrpc endpoint
• the rfc crud-plugin can process the records one by one
• Before and after each insert operation, the method triggers an action.
• Before inserting, it exposes the record data.
• After inserting, it exposes the record id.

And the last two, was exactly what I wanted.

adding functionality with plugins

Now I can add tiny plugins, that ‘listen’ on the action hooks in the rfc-methods. If there is an INSERT into the database, my plugins read which table it affects. They can perform actions, either before the insert, with the new record data, or after the insert, with the new record id.

That’s basically all it takes. As technique, it has it’s limitations, but it can come in handy sometimes.

Using the Bing API to list search results is easier.

Bing TOS : not for seo rank checks

In the last paragraph of the api guide, Bing give a quick recap of their TOS, you can do max 7 queries per second, and using the results for SEO rank checks is explicitly prohibited.

These following snippets (text source) are hence explicitly not to be used for bing search engine result page (‘serp’) rank checks.

bing api with simplexml

So here is one for web results using php simplexml. The web api (which uses namespaces) allows for retrieving max 1000 results per term at max 50 results per query, you can specify the number of results and the offset, where to start grabbing results.

..and one for the pictures using php simplexml :

I actually like that api, I am going to use that.

bing api with json

Bing seem to prefer you use json, less bandwidth usage. After their example in the api basics guide :

Of course there is the old RSS-option, which doesnt require an appid but also falls under the api 2.0 tos, and a soap option.

other sources :
There is a bing api php class made over at routecafe, and a jquery bing plugin using json over at Einar Otto Stangvik’s blog.

I don’t so much have a problem with that, there is not even a release date set for PHP6.

I was pondering more in general on backward compatibility and legacy function support, ereg serves as a nice example.

Does it make sense to make a plugin with general php patches ? A central plugin would make it easier for some coders (me of course) to check what functions are dropped, and update code without code breaks caused by updates of php or wordpress core.

So I don’t have to reinvent the wheel. Something simple like :

Anyways, I’ll have to read about it some more I think.

This page covers most of it :
dynamically add functions to php classes @ www.gen-x-design.com. The class construct at the end of the comment thread, Martin Pietschmann’s contribution, is rather useful. This pattern revolves around importing functionality from pluggable classes and exposing it through one object instance (the ‘decorator’ pattern mentioned is mostly used for writing extended classes, different objects with the same base data and functionality, ‘views’ sort of).

I can include the file with base, import and user class into function.php, and on making a user object have it read the directory and import functions modules (or load functionality conditional based on user role/authorization).

For this example I used the wordpress options table. I write a new functions class

…store the added class name in the options table…

…and load the plugin classes when instantiating the user object :

In the wordpress template i can use the added functionality through the user instance :

Fun with classes :)

add. 3-8 (qed) :
a function to load a directory with plugin files, i add a header /* plugin pluginfilename */ and check the files if there is a header.

the feed has three namespaces (atom, opensearch and dxp/analytics), a simple way is accessing the ENTRY tags (from the Atom namespace), in that tag is one DXP: line and that has the answer to the question.

<dxp:metric confidenceInterval=’0.0′ name=’ga:pageviews’ type=’integer’ value=’755′/>

Important is using the (string) typecast, normally simplexml returns a simplexml object, when you force a string type, it gives the actual metric ga:pageview value attribute as number.