In keeping with the idea of only posting every few months... yeah okay I'm lazy.
Regardless, I recently found myself in an endeavour where I needed to retrieve the most recent content from a feed. On the surface this seems like a straight forward proposition, get the feed URL, fetch the feed and use the content from the first entry. One should be so lucky. It turns out that the initial feed was sponsored so new entries weren't always the top entry. Annoying but these things happen and after some investigation a "better" feed was revealed, one sorted by date instead of some variety of weight. Perfect!
Alas, perfect it was not. In fact the continued observation was that at times the middle of the feed changed but the top did not. Clearly entries were being added, they were in fact sorted by date, so my question became what the hell?
Okay, you're right, I didn't say hell.
A little prodding revealed that, by design, the date on the feed entries was a creation date for the item represented in the feed and not the update or insertion date. The insertion date? Gone, never recorded, unimportant for over seven years.
What can we learn from this? Don't loose <meta> DATEa. Don't loose meta-data either for that matter. Consider carefully when designing any sort of data store what is needed now and what may be needed in the future. I find that in many cases if there is any sort of valuable business content associating at least the insertion date is paramount. Update date and creation date, if there is any possibility that this may differ from insertion date, are just good ideas. So from experience, preserve "natural" meta-data, else someone will eventually need it and you'll have to re factor a less than ideal solution. Guaranteed!
cheers and happy <meta> DATEa days!
2008-01-16
<meta> DATEa
Subscribe to:
Post Comments (Atom)
Save all data! :)
ReplyDelete