Shoutbox

Just had a good idea for a plugin. It would be great if you could add commands with urls of webservices. I'll explain:

1) User downloads webservices plugin
2) User adds a webservice
3) User types /servicename [parameter] e.g /cinema Cardiff
4) Plugin Recognises commands and loads the appropriate url (not visible to user)
5) Plugin puts text in message and sends to contact.

This would mean that people providing information on the internet could easily provide the information to messenger users without the need to develop an individual plugin.

Example:
Prior to the conversation, Richard has added www.script.com/cinema.php?name= with the command /cinema into the plugin settings

Dave says:
What's on at the cinema tonight?
Richard says:
I'll check

--Richard types /cinema Leicester Square
--Plugin visits script www.script.com/cinema.php?name=Leicester%20Square

Richard says:
Film Choices:
Team America: World Police (15)
Saturday 18.00, 19.00, 20.15.

Assault On Precinct 13 (15)
Saturday 17.45, 20.30, 21.00



Cheesy example but it would be quite cool, trouble is the only programming skills I have are in php, so wouldn't know where to begin coding a plugin!

In any case it would have to be /xwebservice cinema Cardiff

Anyway, who would add webservices?
I would understand if XML pages were more usual to rertrieve data easily, but how do you plan the plugin to parse the actual site without having to regex (to code an inteligent reader) like a madman?

The need for more XML sources seems really prominent to me

Ok, I maybe used the term web service a bit loosely, I'm not talking about xml web services here, just plain text web pages which are dynamically updated.

Anyway, XML is better than plain text. But there are no plaintext sources in most commercial websites.

If there's the need to parse the html of a site and extract the infromation (regexing)... It's just too much. And normal users won't be able to do that.

On the other hand, if it was usual to offer XML output of the billboards or any other boards and tables of services, the world would be much better.

The idea is all right, but it's really difficult to achieve a system that would work well easily.

For example:
If the public transport companies would have XML oputput I would code A system that would tell me exactly when to take a train if I want to be at a certain time in a certain place and I could add that to my digital scheduler without having to look at the timetables myself and calculate the time I need to get to some place.
Unluckily XML (or say plain text) is very rare in these cases. Neither companies nor webmasters care about standardised outputs that may interact way better with many other services.

I understand our point, but I was thinking the parsing would be done by a php/cgi script, which just outputs the information in plaintext

But the problem's still the same: you gotta get the pages heavily parsed. It's possible, but you may need hours for only one billboard to work properly.

However the variety of websites makes it a tough work to parse that amount of information.
This somehow reminds me of vWatch, a project of which I'm co-manager, that does automatically get the latest versions of any apps from reading and interpreting the author's webpage on-the-fly. It was really difficult to make a system that got only one piece of data, like the version number. Imagine what it would be to parse much more information with the risk of existing a layout change in the website which screws up all the work.

If there was a well founded and reputated project where people would help to get the data parsed and outputted in standard language, namely XML, it would be definitely easier.

ok, I made the example. It gets film times for Uk Odeon cinemas

http://sprinkles.me.uk/odeon.php?cinema=

Try These if you don't know any odeons
http://sprinkles.me.uk/odeon.php?cinema=Sheffield
http://sprinkles.me.uk/odeon.php?cinema=Bridgend
http://sprinkles.me.uk/odeon.php?cinema=Wrexham

It looks a mess on the page, but if you view the source, it's formatted with \n, one film per line

Takes ages to load, but looks nice.
Now get it in XML, otherways it's a pain in the ass to read for a desktop app or a CGI.

XML allows to read and parse the full thing into multidimensional arrays being directly available to use by any language (must have XML libraries, but that's usual).

Loading nearly instantly here, and i'm on dialup

I'll have a go at making it xml now

edit: Done:
http://sprinkles.me.uk/odeonxml.php?cinema=Bridgend
http://sprinkles.me.uk/odeonxml.php?cinema=West%20End
http://sprinkles.me.uk/odeonxml.php?cinema=Leicester%Square
http://sprinkles.me.uk/odeonxml.php?cinema=Sheffield
http://sprinkles.me.uk/odeonxml.php?cinema=Wrexham

etc...

The times tag is messy. Is it going to be shown like that to the user?

You could create a day tag.

<day name="THU">
<time>22:00</time>
</day>

Edit: btw, you might want to edit your post (in your post: edit button, next to quote, q. quote, etc.), instead of double-posting, as it is against the forum rules.