RSS as an Update Announcer
To clarify what RSS is and why it's got everyone so thrilled, allow me to just begin on common floor with a little something we presently know, a conventional webpage. Historically, a website contained whatever content may have been place on it and that content could be static or might change consistently. The situation here has generally been that a person had no method of being aware of when or if that content material had improved apart from checking again periodically or becoming notified by someone.
RSS solved that challenge by "announcing" articles updates. A site proprietor produces a special file termed an RSS file in conjunction with a website link to it, and this generates a "web feed." A internet feed can be a data format employed for sending end users content updates. Users possess the alternative of "subscribing" to this feed either by means of a stand-alone desktop application called a "feed reader," by an internet based subject material aggregator like Newsburst, or, progressively, right by their common web browser. After subscribed, users are notified when there's new content posted to the feed. Which is all properly and very good, nevertheless it will not finish there.
RSS Website Feeds
An RSS net feed is in fact just an XML-based file that sits on the internet site like several other file and incorporates whatever material the web page owner desires to place into it for distribution. It can be designed and taken care of manually or dynamically (ideally the latter.) Part 3 of this series will cover the development of these feeds. You will find not actually a whole lot to it.
I do not need to shed you in the terminology or the acronyms here. An XML document--especially on the variety we're discussing here--is an extremely easy text document. The markup has specific similarities to HTML, but exactly where HTML defines ways to display the information, XML categorizes the content material (for example, figuring out title, description, author, etc.) and does it in a machine-readable format which implies that distinctive computer software on distinct functioning methods on various platforms can easily entry and display that articles. This is the part that allows for syndication. Quite a few apps exist that hunt for this sort of information, examine the information, and parse it back again out for use elsewhere.
RSS for Articles Syndication
Content syndication is (by definition) the primary utilization of RSS. Originally, RSS files outlined just the title of a piece, the creator, the date of publication, a hyperlink back for the unique material, plus a quick summary to act as a sort of teaser to obtain you to go back again to your original internet site to read the article--hence the mistaken look at that it absolutely was only great for blog site posts and information headlines. Now it consists of syndication of complete content--including HTML--along with "enclosures" to incorporate multimedia material like images, audio files, and video files and that content material is getting utilized in at any time much more imaginative methods. This needn't scare you far from applying it.
Ok, we’re quite very pleased with this - we think it’s a cool use of blogs, webfeeds, podcasting and an iPod Shuffle in a real live business scenario. We’d be interested to know what people think - comments welcome.
Our client is the classic Type-A personality, time-poor, stressed executive with too much to do and too little time to do it - he spends most of his life on planes in transit between meetings. He needs to keep up with the key developments in competitor intelligence, but gets very little opportunity to sit in front of a screen to browse through reports. Neither does he want to drag a large pile of paper around with him.
We put together a monitoring package, scanning a variety of aggregated news sources - PubSub, Google Alerts and so on. Based on criteria that the client supplies - “this week I need to know about X” - we scan the feeds and grab anything that seems likely.
Then we quickly scan what comes up - and using our skill and judgment - select the items that we judge to be the most interesting. We narrate a summary of each item (in our finest BBC English accent) as an MP3 file. They’re short podcasts, no more than 3 or 4 minutes each. And there’s maybe a half-dozen at a time.
The MP3 files are uploaded to our server, and we publish an RSS feed with the MP3s as enclosures. Our client runs a news aggregator, so as he’s collecting his email from whichever side of the world he’s on, he’s also downloading the RSS feeds.
His news aggregator grabs the MP3 enclosures and downloads them into iTunes, tagged into a specific ‘News Podcast’ playlist. He plugs in his iPod Shuffle, which is set to grab everything in the News Podcast playlist. At the same time, if there are any interesting background articles that we’ve found, those are sent as enclosures as well, and dumped into the ‘pendrive’ area of the Shuffle with a script.
When he’s sitting on the plane, he is able to listen the podcasts of the news items while he’s eating or just lying back with his feet up (he flies Business Class, so there’s room for his feet.) If there’s an interesting file to look at, we’ll mention that in the podcast - so he can then stick the Shuffle in the side of his laptop and access the files as if we’d emailed them.
The secret sauce:
A blog engine to provide the webfeed. Garageband and a couple of
tweaks to record the podcasts. Some minor configuration of iTunes to
grab the webfeed and download it to the Shuffle. A script to do the
same with any files.
A couple of days to knock together the RSS feed and enclosures. A few hours a week to record the podcasts. And $99 for an iPod Shuffle.
It’s a seamless solution for him - all he needs to remember to do is stick the Shuffle in the side of his laptop while he’s downloading his email. The Shuffle is small enough to fit in an inside pocket, so there’s no ‘luggage overhead’ to deal with. He can listen to the podcasts in small windows of time that would otherwise be wasted - in the back of taxis, queueing at the departure gate and so on. And he can get hold of documents that we’ve found without them being drowned out in the mass of email.
We’ve had a ton of interest about this since I put the post up, and one of the common themes was “why didn’t we just use text-to-speech to create the podcasts?”
That was something we tried in the early stages, and while it’s ideal for the producers (i.e. us) it’s not so ideal for the client (i.e. our stressed, Type A exec) The problem is that although the rendition of text to speech is flawless in the sense of taking the words and speaking them out loud, it doesn’t have the inflections and cadences that make conversation bearable. Imagine Stephen Hawking reading the telephone directory, and you’ve got a fair idea of what it sounds like - it’s fine for one-off quickies like a system that reads back airline departure details, for example, but you wouldn’t want to listen to it for long once you’re on the flight and your seat’s reclined.
The other factor is that recording the podcasts isn’t actually the most labour-intensive part of the process - the real work is editorial, when we have to decide what it is we’re going to record. It’s something that improves with practice, but it’s damn difficult to automate
Having said that, there’s some really neat software out there that will take an RSS feed and convert it directly (see the comments for the link), so that’s something that we’re going to take another look at.