AmethystRSS Blog

I’m on vacation right now.  When I’m home in Austin, I hang out at Cospace.  The place was remodeled about a year ago over a weekend.  Now, they’re repainting the bathroom with a Donkey Kong motif.  I’m getting old, increasingly T-shirts and other artworks reference games, books, and movies or TV series I’ve never seen/played.

I read Seth Godin’s blog and have a couple of his books.  I like the title of “We are All Weird”.  I’ve aways been at least a little weird, what techie hasn’t and unofficial motto of Austin, Texas where I live is “Keep Austin Weird”.  One of the bumper stickers campaigning against a Walmart Supercenter was, “Walmart isn’t weird enough for Austin”.  Which is one of the points of the book, Walmart is about the mass market, Weird is about niche markets.

But his emphasis on tribes (e.g. the Apple tribe, the Nike tribe, or even the Ford tribe) bothers me.  Civilization’s progress is about expanding the boundaries of who is human, who is covered by law.  In tribalism, only people in the tribe are fully human, i.e., killing them is murder.  Killing someone outside the tribe is not murder.  Expelling someone from the tribe, puts them outside the laws protecting members of the tribe, i.e., they are an outlaw and killing them isn’t murder.

Giving the current culture of meanness in the US, I’m very leery of of increasing tribalism.  It looks like a regression, not progress.

El Camino Real was the road that connected early California’s missions.  Its modern namesake is the backbone of the Silicon Valley.  The original Byte Shop where the Steves sold their first Apples (and I bought my first computer, an IMSAI 8080) was on El Camino Real in Mountain View.  It forms the eastern boundary of Stanford University, where Donald Knuth taught, Google’s founders started up, and I bought my copies of The Art of Computer Programming.  The Stanford Linear Accelerator Center (SLAC) auditorium hosted the meetings of the Homebrew Computer Club.  It’s on Sand Hill Road, not far from El Camino Real.  I drove up and down El Camino Real to assemble the parts of my second computer, a DEC LSI-11 – the chassis and CPU bought at a trade show up El Camino Real.  The memory bought from Intel, down El Camino Real.  The Fry’s grocery store where the younger Frys started selling computer parts was just off El Camino Real in Mountain View.  The only food in the Fry’s in Austin is snack (junk) food.  I wonder if there are still any Fry’s that sell groceries?  Kepler’s Books on El Camino Real in Menlo Park hosted events, sold magazines and technical books, and was part of that early computer scene, as well as many others.  The fork I took into Zen Buddhism merged back into Christianity through A Taste of Water, a book I bought at Kepler’s.  When my wife was doing her chaplaincy internship at Stanford’s Children Hospital, I ate lunch in a Vietnamese-Chinese restaurant in a former drive-in restaurant on El Camino Real where I could hear startups of the Dot Com era being hatched at nearby tables.

The Byte Shop is long gone.  The Homebrew Computer Club lives only in documentaries (e.g., Revenge of the Nerds) and history books.  Donald Knuth is retired.  The Dot Com bubble popped half a decade ago. Kepler’s folded several years ago inspite of enthusiasts fund raisers.  And now Steve Jobs is gone.

Amethyst has been running on the same EC2 instance (Amazon cloud virtual computer) for over two years.  I installed a variant of Debian’s Lenny release (5.0) built by Eric Hammond and the team at Alestic. It has never crashed, though it has been rebooted several times due to kernel updates.  I have been happy with that decision, but Lenny is now the ‘old stable’ release.  It is time to update the OS.  Eric has chosen not to build a Debian 6.0 build for the Amazon cloud (reportedly, XFS and several other packages are broken on the cloud).  Some packages (e.g. Redis) do not build or fail to pass their own tests on Lenny, so I do need to update.  So I started looking around at what is available.

Amazon has announced it’s own Linux build, based on Red Hat.  I don’t like Red Hat for a number of reasons that basically boil down to they are a little too “bleeding edge”.  I value reliability highly, especially on production servers.  My wife’s laptop (Dell M1330N) has Ubuntu pre-installed and I’ve been maintaining it.  Not my favorite, but workable.  Ubuntu server is well supported on Amazon.  So I’ve spent the last two weeks migrating Amethyst from Debian Lenny to Ubuntu Maverick Meerkat.  Yes I know there are newer versions, but a key piece of software is only available on Ubuntu up through Maverick.

One of the attractions of Ubuntu is Cloud-init, a protocol to configure a generic Amazon Machine Image (AMI) for specific needs.  Cloud-init builds on Eric Hammond’s work with passing a text file, possibly a script, to an instance at startup.  Cloud-init supports several file types: shell scripts, Python run-parts, includes from a URL, and compression.

My goal was to go from a database backup, my laptop, and a good Internet connection to a running system in under an hour.  After considerable work, mistakes, tweaks, and more than a bit of Googling and reading, the goal is reached.  Sort of.  The system is working from the user’s perspective, but not all the monitoring is running.  That takes another half hour to hour depending on how much of the monitoring I want.  Today (July 13, 2011), I decided it was good enough for production work and switched instances.  Amethyst on Lenny is no more.

Since I’m a programmer/entrepreneur, not a sysadmin, improvements in the setup will have to wait.  I plan on spending some time improving the setup every month or so.  By bringing up new instances monthly, I expect needed changes to be smaller.  Plus, I’m essentially running the disaster recovery scenario and know it works.

I’ve swapped the primary and secondary domain names for this blog.  If you bookmarked is at abluz.dyndns.org, please change it to blog.AmethystRSS.net.  Both names will continue to work for the near future, but not forever.

I’ve been looking for a todo list manager for my Android smartphone to replace Life Balance on an aging Handspring Visor (PalmOS).  There are lots of candidates, many claiming to implement David Allen’s Getting Things Done (GTD) ranging in cost from free on up.  Working with them has pointed out that I’ve become sloppy and let my projects have no end point.

So I’m adding goal dates (probably too aggressive) and a clear cut success criteria (e.g., at least one additional user and referrals from search engines).

Projects need to be much smaller than I’m used to. (Life Balance handles sub-projects to arbitrary depth, leading to top level projects that span years and no end in sight.)

The longest revised project is due in 6 months and has an unambiguous success criteria, at least one paying customer.  Most are in the 1-3 month range.  This keeps my working time much  better focused.

Jon Postel is generally credited with:

Be conservative in what you do; be liberal in what you accept from others.

The consequence of trying to follow this is that sloppy programming lives much longer than it deserves and systems downstream have to go to great lengths to handle all the badly formatted data sent their way.

I have spent over a week digging into some of the feeds Amethyst gags on. A common problem is HTML pages indicating the Website and/or the RSS feed has moved. HTTP and several of the feed standards have a nice standard way to do this. And Amethyst can handle them. So how far should I change the code to extract the new Website or feed URL out of what is often minimalist, hand coded HTML (i.e., not strictly valid HTML)?

Amethyst uses libxml2, an widely used XML parsing library written in C and quite fast. It is rejecting several feeds as invalid. One has ASCII Vertical Tab characters in it. Should they be allowed? Is the libxml2 team correctly rejecting invalid XML? Or did they miss a case? Same for an <iframe> in a <creator> tag. And a raw URL inside a <guid> with ampersand (&) characters that libxml2 is trying to interpret as HTML entities.

I think the first two can be worked around. Trying to handle the last one is an endless series of hacks to hacks; experience says “Don’t go there.”

Do I code around them? Patch libxml2 to accept them? Or take my lumps as the big meany by rejecting them?

AmethystRSS, my RSS reader Web app, runs on Amazon’s cloud.  It just passed the 200 day uptime mark.  I have read some blogs of people who spend a fair amount of effort keeping their servers running: 24×7 monitoring with automated restart, etc.  Network connections seem to occasionally bog down for a bit, but the server itself has never crashed.

Another amazing, dense, rich monologue from Bruce Sterling. A great way to end SxSWi. Hopefully a full video will eventually make its way onto the Internet. Most of the great lines are on Twitter, #brucesterling.  Time to rest and start working through the backlog of URLs I e-mailed to myself, Twitter hashtags for sessions, and my notes.