Aggregate user group information and events via microdata

[coderbyheart]

By reading microdata embedded in the usergroups' home page we would ergonomists the need for them to maintain a copy of their information in our XML files.

This could be implemented by parsing a groups URL and collecting

• Organization
• Person and
• Event microdata

[heiglandreas]

Sounds like a good idea to alow the "changeable" data to be hosted by the usergroup while the "static" data is available via the github repo.

[coderbyheart]

We could get rid of the XML file totally as the only information we need is an URL pointing us to the website of a usergroup.

[heiglandreas]

I wouldn't do that as it removes a lot of information! And It is great as it actually is. The only "drawback" (if one can say so) is that some information has to be provided twice. That is for one thing the calendaring niformation - which can also be provided as an Ical-URI - so it doesn't actually count. And that might be the information about locations - which can also be provided via the ical-file and doens't therefore count as well! And all the other information is not really easily retrieval from a simple URL except the usergroup provides the information from a defined file from their homepage. And then it's irrellevant whether that file is located at their homepage or in github.

So aggregating usergroup-informations via microdata is a goodie that is not really relevant for a working website.

[coderbyheart]

Have you looked at http://schema.org/Organization ? Everything we have in the XML can be describe with microformats … So the idea would be (like with the iCal feed) that this is an opt-in for usergroups. If the provide microdata on their website, they can state that via the XML-file, only then we would use the data from their website.

[heiglandreas]

Yeah, I've seen that.

So the XML-File at github would "simply" mash up a set of microformats. That does sound cool indeed.

But nevertheless the question is,

• what happens when microformat-file does not validate? Don't display the information (well obviously) and notify the person in charge. So a usergroup-entry would have to have an attribute personInCharge with an email-address (or a corresponding tag).
• And when are those microformats read? It has to be some sort of dynamic generation as they could change anytime. I'd opt for a kind of cache that retrieves the information once in every 24 hours. But as far as I see the information is always retrieved on runtime which might be a bit exhaustive.

Or am I wrong there?

[coderbyheart]

So the XML-File at github would "simply" mash up a set of microformats. That does sound cool indeed.

No.

If a group decides to use microformats, they set a switch in the XML file. Then we scrape their site (using the url provided in our xml file) to gather the data.

The data is marked up by them on their site. Check out this page which contains markdown describing a Recipe. Use the Google's tool for testing rich markup to see what is in there. Here is another example of how microformats are used.

[heiglandreas]

Ah, Sorry, didn't catch the concept at first. Got it now! So scraping on demand or once every n hours?

[coderbyheart]

So scraping on demand or once every n hours?

We would scrape it once a day.

On demand would require some kind of notification channel (webhooks). That is too complicated for most users. But I can image a simple form that triggers a scrape …