Nick Bradbury

Personal ramblings from the creator of HomeSite, TopStyle and FeedDemon.

Last build:
Language:: en
Feed URL:: http://nick.typepad.com/blog/index.rss

RSS FEED IDEMS: Nick Bradbury

YouTunes: An Example Yahoo! Pipe
My friend Rex Hammock wants to know what Yahoo! Pipes is all about, so I whipped up a simple example "pipe" to show how it works.

A pipe takes input from different sources, mixes them together based on criteria you specify, then outputs the results as a single RSS feed. In this example, I created a pipe named "YouTunes" which links the top 10 songs from iTunes with videos from YouTube. Designing it was simple:
1. First I dropped a fetch source module onto the canvas, and told it to use the iTunes Top 10 Songs feed as the source.
2. Next I added a for each: replace operator module, which takes an input feed and replaces it with output from another feed. In this example, I used the iTunes feed as the input and a Yahoo! search feed as the output.
3. I instructed the Yahoo! feed to search for the song name from the iTunes feed (itms:song), then restricted the search to http://www.youtube.com/
4. The final step was attaching the "for each" module to the pipe output.
The end result is an RSS feed which links to YouTube videos of the iTunes top 10. It's not perfect - it needs additional filtering to weed out irrelevant videos - but I wanted to keep it simple for this example. If you'd like to check out the actual "YouTunes" pipe, you can find it here (Yahoo! account required).

PS: Here's how it looks in FeedDemon.

Fri, 09 Feb 2007 13:08:06 CST
Yahoo! Pipes is a Platform Play
Given that Yahoo! Pipes has been clogged much of the day, chances are you haven't been able to see it yet. So I'll take a few minutes to expand upon my brief previous post.

What does Pipes do? Pipes enables the creation of custom RSS feeds tailored to specific interests. Here's a very simple example: you could create a "pipe" which asks for a zip code, then generates a single feed containing information taken from a collection of weather feeds, news feeds, events feeds and sports feeds pertaining to that zip code. Or you could create a pipe which mixes news about a big event with Flickr photos of the same event. Or how about one that combines comments from the same person across multiple blogs?

However, while this feed filtering capability is exciting, the fact that Yahoo! is attempting a true web-based IDE is what makes Pipes stand out to me. Pipes offers a web-based development platform along with a web-based, drag-and-drop development tool which enables easily browsing, viewing and cloning code written by other developers working on the same platform.

In other words, this is Yahoo! making a platform play, not an aggregation play. Yahoo! is upping the ante in their competition with other web platform players (notably Google, Microsoft and Amazon).
I'm extremely impressed by how much Pipes tries to accomplish, but it does have sort of a "proof of concept" feel to it. Regardless of whether Pipes succeeds, though, it will be interesting to see whether Yahoo! creates other IDEs which build upon the lessons learned from this one.
BTW, if you'd like more information about Pipes, these links should help:
Thu, 08 Feb 2007 14:59:53 CST
Yahoo Pipes is Stunning
I'm not easily impressed, but this impressed the hell out of me.

It's geekier than it should be, but it's still a big step forward.

Wed, 07 Feb 2007 23:50:36 CST
Stupid Optimizations
One of my priorities for the next FeedDemon has been to fine-tune the "Popular Topics" feature so it's much faster than it currently is. If you're subscribed to a lot of feeds that talk about the same topics, it can take several seconds for FeedDemon to build the popular topics page - not painfully slow, but slow enough that you might not use the feature all the time.

And that to me is a problem. See, I've been pretty busy for the past couple of weeks and haven't had time to catch up with my feeds, so I've been marking them all read and relying on the popular topics page (along with a few watches) to bring me the latest buzz. It's been a very useful feature to me lately, so I hate that it's not as fast as it should be.

I spent several hours ripping through the code and tweaking performance, but it still wasn't fast enough. I was about to chuck it all and take a different approach when it dawned on me that the problem could be outside the code that builds the popular topics - and sure enough, after profiling the entire codebase I discovered that a crappy routine I wrote long ago which compares two URLs was bogging things down (side note: my routine was pretty bad, but not as bad as the one Dare Obasanjo blogged about).

I spent a few minutes rewriting that routine, and it made an enormous difference. After the initial building of the link database (which occurs the first time you use this feature), displaying the popular topics page now takes less than a second. And changing the filter to show popular topics from today, yesterday, etc., is virtually instantaneous.

When things like this happen, I'm torn between being excited and pummeling myself. I'm pleased that I was able to speed it up, but I'm embarrassed that the performance problem was caused by one stupid piece of code.

Fri, 02 Feb 2007 16:46:38 CST
Troubleshooting
While reading Nick Harris's post about troubleshooting, it occurred to me that FeedDemon could be a lot friendlier when an error occurs. Right now when something goes wrong, FeedDemon catches the error and displays detailed information about the problem - but most of that information is useful only to our support team, and there's no built-in way to send it to us. Even worse, when this information does reach our support team, we usually have to ask for additional information (such as the version of Internet Explorer being used).

So I decided to fix this.

Starting with the next version, instead of displaying a dialog stuffed with cryptic details, FeedDemon will display one like this:

If you choose to send the error report to us, it will include a single compressed attachment containing details about the error and your Windows environment, a screenshot of your desktop (optional), and a copy of your subscriptions (also optional).

The next step is to figure out how FeedDemon can help customers resolve certain problems. For example, we frequently see errors caused by buggy or outdated Internet Explorer add-ons, so it would be helpful to both us and our customers if FeedDemon could figure out when an IE add-on is mucking things up, and then explain how to disable it.

PS: The dialog shown above is my own creation, but I rely on the excellent EurekaLog to gather information about the exception.

Fri, 02 Feb 2007 15:57:10 CST
Nick Harris: Creating a Pleasant User Experience
NewsGator codemonster Nick Harris has been on a tear lately, authoring a ~~four~~ five-part series on creating a pleasant user experience:
Definitely worthwhile reading for software developers.

Wed, 31 Jan 2007 16:47:08 CST
Attention Nashville Bloggers!
I need your help spreading the word about Max It Out, a Nashville charity event inspired by five-year-old Max Royka. Max contracted bacterial meningitis when he was six weeks old, and every year his parents hold this event to raise money for meningitis research.

Max's parents are good friends of mine, and last week they had to rush Max to the hospital. I just returned from visiting them, and they're concerned not only about their son, but also about their charity event. They need to sell more tickets, but they can't focus on that while their son is in the hospital.

So I'd like your help letting other Nashvillians know about this great event, which will be held on February 3 (less than a week away!). My wife and I are among the sponsors of Max it Out, and I can tell you from personal experience that it's a good time.

If you're in (or around) Nashville, please spread the word to boost ticket sales for this worthy charity.

Update: Here's a list of bloggers who've talked about this so far - thanks, folks!
Sun, 28 Jan 2007 14:46:03 CST
The Hype Machine RSS Feeds
Jason Calacanis just pointed to The Hype Machine, a free service which tracks MP3 blogs and builds a list of MP3 links that are downloadable by a podcatcher such as FeedStation.

Even better, they offer RSS feeds of their search results. If you're using FeedDemon, simply perform a Hype Machine search from within FeedDemon's browser, then click on FeedDemon's autodiscovery icon after it lights up to subscribe to the search results.

Every item in the feed includes links to both Amazon and iTunes, where you can purchase the music (I've already purchased a CD I didn't know existed thanks to the Amazon links). One big missing feature, though, is Creative Commons license information - it would be nice to see that added in the future.

Thu, 25 Jan 2007 14:59:53 CST
White Hat Wikipedia
Last week I blogged about editing your own Wikipedia entry, and this week Wikipedia is being discussed all over the geekosphere following Microsoft's attempt to pay someone to edit an entry about Office 2007's new XML format.

I had planned to write a much longer post about this, but it turns out that my Nashville neighbor Rex Hammock has already said what I was going to say:

"If Wikipedia has become the platform of record for web-based knowledge, then having a voice there is going to be a requirement for corporate America. Wikipedia either needs to find an accepted “white hat” way for this to be done directly and transparently (and not some “in the discussions, off the website way), or dark-hat, Rube Goldberg solutions will naturally follow."

So I guess if you want to know what I think about the whole thing, just go read Rex's post :)

Thu, 25 Jan 2007 05:14:47 CST
Google Groups Feeds
Google Groups is finally out of beta, so I thought I'd update my blog post from 2004 which described how to subscribe to Google Groups feeds in FeedDemon (or any feed reader, for that matter).

To view the available feeds for a specific Usenet group, navigate to a URL which uses this format:

   http://groups.google.com/group/NAME-OF-GROUP/feeds

Or to skip that and subscribe to an Atom 1.0 feed containing the most recent 15 messages, use this format:

http://groups.google.com/group/NAME-OF-GROUP/feed/atom_v1_0_msgs.xml

For example, to see feeds for the alt.html newsgroup, browse to this page:

   http://groups.google.com/group/alt.html/feeds

And to subscribe to the Atom feed for alt.html, use this URL:

   http://groups.google.com/group/alt.html/feed/atom_v1_0_msgs.xml

Wed, 24 Jan 2007 17:04:45 CST

Nick Bradbury

RSS FEED IDEMS: Nick Bradbury

Submit your RSS Feed

Subscribe to this RSS Feed