Syndic8 2004 Retrospective

2004 has been a really good year for Syndic8, and I wanted to take the time to enumerate all that’s happened. I have spent a lot of time coding and answering email, and not enough time talking about what I’ve built and what I’ve learned. I will try to do better in 2005. Anyway, here’s a quick summary of what’s transpired.

Feeds – We had 42,699 feeds at the beginning of the year, and added 220,372 during the year, for a grand total of 263,043. Making room for all of these feeds, and processing them efficiently, was a challenge earlier in the year, but the site is running smoothly now. I know that we can do better with Reviewing, and I’ll get to that in a bit.

Users – We had 5,196 users at the beginning of the year, and added 4,948 during the year, for a grand total of 10,144.

Hardware – We continued on our four-processor BocaCom box through the year, adding a 73 GB disk in the spring and another 250GB in the fall, along with an additional gigabyte of RAM, bringing the total disk space to 400 GB and the total RAM to 3 GB.

System Reliability – We had a lot of problems with system reliability earlier in the year. Many mornings I would find that the system had locked up overnight, and that an Apache restart was required. I added some advanced monitoring and automated the restart process, which made things considerably better. There was an issue with a CPU heat sink, which caused a few lockups as well. At mid-year we had a hardware failure on the root drive, and we ended up with a newer (and less reliable) Linux kernel. It took a really long time to diagnose the problem; until I fixed it the system would crash every 12-24 hours. The problem has been fixed and the uptime now stands at 17 days.

Ranking and Ratings – Our Alexa rating moved from around 15,181 up to 7,394. Google records about 4,110 backlinks to the site. Additionally, the number of hits from a Google “inurl search” grew from 691 to 16,500.

New Features – We added a lot of new features and content in 2004:

  • The Feed List allows filtering by language.
  • Multibyte strings and multiple character sets are now supported throughout the site.
  • The Atom syndication format is now supported. Jeremy Zawodny’s blog was the first Atom feed to be submitted.
  • Published a document on the Syndic8 Personal List system.
  • The Featured Feed system allows feed owners to publicize their feeds on Syndic8 at a low cost. This has worked out well, with an average of 10 feeds featured on any given week.
  • DNS hosting moved from a 10 year old 486/66 box in my home, to professional hosting at zoneedit.com (highly recommended).
  • Many more news aggregators are now supported through “Add” links on the Feed Info pages.
  • The RSS 1.0 <taxo:topics> tag is now parsed and used to set the category of the feed, if present.
  • It is now possible to generate an OPML or OCS list restricted by category, as seen on the category browser pages.
  • Compressed data is requested when polling for feeds, and the results are tracked as statistics. It would be excellent if more feeds supported compression, given that each Syndic8 poll now pulls down over 2 GB of data, and that the polls run twice per day.
  • Syndic8 now receives pings from a variety of sources, and it applies filtering and other checking to distinguish good pings from bad. The pings are displayed on the Recent Pinged Feeds page. Pinging and other real-time processing will be a big focus for Syndic8 in 2005.
  • Newly discovered headlines are now published to the RSS Scroller at mod-pubsub.org.
  • Syndic8 now tracks feed enclosures, and displays them on a new tab in the Feed Info pages. There are also some enclosure statistics, of course.
  • The XML-RPC interface now has a GetFeedsInCategory function.
  • The Feed Info pages now include a tab listing recent changes to the feed.
  • It is possible to use the XML interface to generate a list of feeds that have enclosures of a given type. This isn’t documented yet; if you need it drop me some email.

You should know that there’s no faceless mega-corporation behind Syndic8. You’ve got me, working on this site in the wee hours of the morning (usually 5:20 to 7:20), and Bill Kearney, who has a lot of other projects going. We have one hosted server, DSL connections in our respective homes, and that’s it. No venture capital, no silent partners, no big pot of VC gold. Having been a part of several venture-backed startups, I’m still impressed with what we’ve managed to do with the resources available to us.

We have found that text link advertising is a great way to support the site; hosting and bandwidth costs are no longer a drain on my credit cards. I appreciate your clicks and your tolerance for these ads.

I should add a big “Thank You” to Bill Kearney, who helps out with the server, with tons of great ideas, and with lots of messages to the mailing list. Thanks, Bill!

Ok, so what didn’t go so well?

A couple of things. System reliability sucked for most of the year, and it took far too long to get that under control. The reviewing system is simply not scaling with load – a lot of people would like to be Reviewers, but the system is hard to use, there’s more manual work required than should be, and the “rules” for reviewing are not really clear. Fixing all of this is going to be a big priority for the first part of 2005.

The site’s user interface is definitely outdated, unattractive, and not all that easy to use. Again, this is going to be a focus for early 2005 — I am going to hire a part-time web developer within the next week or so. The first step will be to make the existing site look better; after that we will look into structural changes and additions.

It would be great if the Personal List feature had some more use. Bill Kearney designed this feature several years ago, and its really, really cool. However, it just doesn’t get enough use.

Other focus areas for 2005 include more real-time processing, better support for and use of enclosures, and a focus on international issues. I am also going to try to be a little bit less of the “behind the scenes” guy, and to participate in the mailing list on a more frequent and regular basis.

I would like to thank all of you for your participation in the site — for your contributions of time, energy, suggestions, and more. I think that we are doing something of real value here — helping the world to be more informed, giving them the ability to have multiple viewpoints on a topic or an event, and perhaps even expanding their mind a bit in the process. That’s a pretty cool thing to be able to do with some code and some XML!

On a personal note, I would like to say that this has definitely been the most exciting project that I’ve ever had the privilege to participate in, and that the fun is just beginning as far as I am concerned. I recently convinced my 19 year old son Stephen to become the first “second generation” Syndic8 member. He’s contributed a bunch of feeds, and he’ll do some reviewing next month.

Thanks to all of you for your great emails and your support, and here’s to a safe, successful, and syndicated 2005!

Feeds with Bit Torrent Enclosures

On the podcasters mailing list, someone asked for feeds where the enclosure was a torrent file (media type application/x-bittorrent). I have been tracking these on Syndic8, so it was easy for me to make a list. Here’s what I have so far:

Mega-Hobbies

Whenever your wife complains that you are spending too much time or too much money on your hobby, it is good to be able to point to what someone else is doing and to say “Well, compared to this guy, my hobby is not that big of a deal.”

So, here’s my list, and you can thank me later:

Random Links

Time to clean out the old inbox:

  • MogileFS – an open source, distributed file system with a configurable level of redundancy.
  • Memcached – a high-performance, distributed memory object caching system.
  • RSS Popper” – a local RSS to Outlook gateway, designed to serve up RSS content to Outlook. New and still in beta.
  • Why There’s No Escaping the Blog – Blogs from the corporate perspective — how to use them (and how not to use them) to build buzz and to spread information. Covers some notable successes and also some notable failures. Well worth reading and distributing to your non-blogging friends.

Bruce Springsteen on the Importance of the News

Way back in October, Bruce Springsteen had something fundamentally important to say about what should be in the news:

Real news is the news we need to protect our freedoms. You get tabloid news, you get blood-and-guts news, you get news shot through with a self-glorifying facade of patriotism, but people have to sift too much for the news that we need to protect our freedoms. It should be gloriously presented to the people on a nightly basis. The loss of some of the soberness and seriousness of those institutions has had a devastating effect upon people’s ability to respond to the events of the day.

Read more in the complete interview.

Design In Flight

Even though I have no design skill whatsoever (look at syndic8 for living proof), I’m still trying to learn. I just ordered the first two issues of Design In Flight. The sample pages were not printable; I hope that the real issues are not locked in this way. If the first two issues are good, I will definitely subscribe.

Great Moments in Microprocessor History

IBM just published an informative article titled Great moments in microprocessor history. It was interesting to read through this and to see if I could come up with a personal link to each of the processors listed in the article. Here’s what I have so far:

  • Intel 4004 – In 1972 or 1973 my father was in the electronics distribution business. One day he gave me some Intel manuals and he started to tell me about this thing called a microprocessor. I remember reading those manuals on the deck of our house, and that the entire processing model made intuitive sense to me pretty much right away.
  • Intel 8080 – In 1975 I was working at the Retail Computer Store in Seattle. I saved my summer earnings and bought an Altair 8800 for myself.
  • Motorola 6800 – I never wrote any code for this machine, but I certainly sold some Altair 680′s and some Southwest Technical Products (SWTPC) goodies at the store.
  • MOS 6502 – My first real for-pay programming job was to write a macro assembler for the 6502, so I know this instruction set really well. Clean, elegant, and to the point. Even better, we were able to write macros (using my assembler) to fill in a few gaps in the instruction set, making it almost totally orthogonal.
  • LSI-11 – When we moved from Bellevue, Washington to Rockville, Maryland, I started working at the Computer Workshop in Rockville. I quickly latched on to the 16-bit Alpha Micro, and did a lot of cool stuff with that machine, none of which I can remember at any level of detail.
  • Motorola 68000 – While working for Intellimac in Rockville Maryland, I wrote a multi-user hypervisor which ran OS instances as tasks. This allowed us to run 4 copies of the Telesoft ROS operating system. The only redeeming feature of this OS was the fact that it ran the Telesoft Ada compiler. I wrote the boot ROM, the hypervisor, and all of the I/O drivers myself.
  • Zilog Z-80 – I wrote a bunch of Z-80 assembler code when I worked at Contel Information Systems. I wrote the code which ran some “toilet paper” printers on the floor of the American Stock Exchange, and I also re-implemented an X.25 protocol handler.
  • Intel 8086 – I didn’t do a whole lot of Intel coding. In fact, all I can remember writing is a very simple dynamic loader which used the dl() function to fault-in some pieces of code, resolve references to newly loaded functions, and then patch up the stack as if nothing had happened.

I think that’s about it! Ireally need to write up more of my career history at some point, before my brain turns to mush!

I did the first little bits here, and here as posts to the FoRK list

New Year’s Day Polar Plunge in Seattle

For years my wife had read about the various “polar bear plunge” events where a bunch of brave fools dive into ice cold water on New Year’s day. She’s an adventurous soul and always wanted to give this a try.

New Year’s day 2004, she finally got her chance. We had visitors, and we convinced them to go with us. We raced up to the designated location, a bit north of the University of Washington, and off they went. I didn’t get to see them go in, because we got there a bit late and I had to park the car. All I got to see was the pictures.

This year, we will get there early, and I will definitely give it a shot.

You can find more information here.

Amazon Web Services Blogs

In addition to the present blog, I am also the principal blogger behind the Amazon Web Services blog. My Developer Relations Group has been using this blog to communicate product status and to point out and feature interesting applications built by the Amazon Web Services developer community.

At first this blog was a bit of an experiment for Amazon — our PR (Public Relations) people, while behind the effort, reminded me that I already been through the Amazon media training process, and that I had to be mindful that I was a company spokesman as I wrote a blog post. As best as I can tell, things are working out well so far. I haven’t done anything controversial, no feathers have been ruffled, and I have never received a visit from the Men In Black (at least not as far as I can recall).

Our Web Services Evangelist in Japan has a similar blog for the AWS community there: Amazon Web ???? ???.