Posts

Showing posts from April, 2004

Death to Aggregators

Frank Steele thinks that aggregators should be subsumed into browsers I’ve got a fair amount of sympathy for what he says. Syndication is about browsing, so why not do it in the browser? The case is even more persuasive because browsing in current aggregators generally sucks – NNW launches new browser instances, nothing in windows blocks popups and so on and so forth. But, having said that, Mozilla/Firefox has shown that the world is moving away from swiss army knife browsers. Feed browsing (for me and presumably a big chunk of other people) is a different kind of consumption experience than normal web browsing. The difference is subtle but it’s there. If I wanted to read feeds in my browser I could use bloglines or amphetadesk, but I don’t. I’d hope that rather than vanishing, desktop aggregators will raise their game when it comes to the browsing experience.

CVS Sucks

I should have set up my project as two separate projects in a solution, one for the UI and a lib for the core code. Having made this great realization, I’m screwed now in terms of CVS ‘cause I basically have to delete everything and start again, losing all my version histories. We use Perforce at work which, for all its shortcomings, handles this case rather well. I suppose there’s always that hope that I’ve missed something because of my weak CVS skills but google suggests otherwise.

Some sage advice

Hmm, so a post which was really nothing more than a bit of a whinge, actually got me some really good advice It hadn’t really occurred to me that people might want to hack the code to a project that hadn’t even released anything yet. In fact truth be told I hadn’t considered that there would ever be a development community any bigger than my code, Paul’s graphics and Phil’s constant nagging ! But, in for a penny, in for a pound. The code is now marginally better commented and there are instructions on how to build it As for test cases, currently they consist of subscribing to every feed that even slightly interests me. I should put something more structured in place, but first I have to work out why Russell Beatie’s large but otherwise uncomplicated-appearing feed won’t parse. While you wait however, go wild with the code, whoever you may be.

Release early, release often?

So FeedThing is now good enough to be the only aggregator I use. It’s responsive and it uses roughly 10% if the memory of .NET news readers such as Sauce Reader. The overwhelming tempation is to release a version. Isn’t that what you’re supposed to do in the wild and wooly world of open source? What worries me is that I think that probably only works if you have no competitors. Or maybe if you have no free competitors. Objectively speaking, if I downloaded FeedThing in its current state, I’d not bother with it again (I know this by the pile of dead aggregators in my recycle bin). The biggest weakness is adding feeds. No auto discovery, no integration of any sort with the browser. If I release early, even if I am lucky enough to come to people’s attention, I am likely to drive them straight away, never to return. PulpFiction is in the same position as me. Trying to release a new aggreagor against established competition. Now OK, their offering is commercial,

How to write an aggregator

This is a list of resources which are useful when building an RSS/Atom aggregator. I found them useful when building FeedThing, maybe other people will too. Expect this list to grow. Things That FeedThing Does Correctly Specs and Things HTTP Primer HTTP level behaviour Aggregator client HTTP tests Determining content-encoding Don’t hijack referer Postel’s Law (see bottom) Strict Stripping of unsafe HTML entities More unsafe HTML advice Do not parse HTML with regexs Unicode for syndication consumers (Also read the comments for a long argument over the merits of this). RSS Security Checks Feed Autodiscovery Things That Still Need Work in FeedThing Comment Feeds feed:// subscription How to Explain an Aggregator CNN Forbes

Timely reminder

Must remember this post when it comes to stripping and making safe the content in FeedThing. Interestingly, the body of that article exposes a bug which actually breaks the current unsafe display code, so there’s two lessons for the price of one.

Aggregation improvements

I think it’s cause I started writing an aggregator , but Phil has helpfully been off and started reviewing the competition. Something that struck me though is that none of the aggregators he looked at supported pop-up blocking. This was something I had planned from day one (admittedly with no real clue how to do it ;-). However it looks very much like I won’t have to bother now. For all its supposed faults Windows XP SP2 seems like it will, at a stroke, improve all the IE embedding desktop aggregators out there. Here’s the big step forward Phil was looking for, without anyone having to do anything.

Back in the blogging spirit

I abandoned my last blog and switched exclusively to livejournal about 3 months in. I think that was the right thing to do, the mix of techy bullshit and photos of my daughter was a strange one and something had to give. However every now and again I feel like I should be ranting about something or other and livejournal isn’t the place to do it. So here’s the plan: livejournal for my life, xurble.org for my geek stuff. Is there room for both? I guess we’ll find out.

About

Gareth Simpson entered a career in software engineering via the usual route of studies in modern languages, politics and philosophy. He lives in Burbank, California with his partner Emma and his daughter Madeline. All opinions are those of the author, and are in all likelihood wrong. Other Places Links Photos Sad memes Rants and Raves

Welcome

You have reached the personal website of Gareth Simpson. I am a UK based Software Engineer with the usual background in Politics and Philosophy. This site contains things of mine that need a home on the internet such as: My small coding projects: Grabbr - A screenshotting tool that uploads to flickr iChat for Miranda - A iChat/Rendezvous chat plugin for the Miranda Instant Messenger iScrobbler for Windows - An Audioscrobbler plugin for iTunes FeedThing - A simple but fast windows RSS aggregator Motherload - A small app for loading MP3 players from iTunes in a shuffle-esque style. jTechnorati - Some java classes for using the Techorati API jTextile - A java 1.3 implementation of the Textile HTML formatting library Greasemonkey Scripts - Little hacks to make the web nicer My weblog wherein you will find: Development notes on my projects Rants about programming Rants about politics Other random things