Sometimes you have to do something dumb because software is being a butt. This has been one of those times.
Well, that was nice. Because of the CMS switch, I can no longer generate a feed of comments (which doesn’t really matter much; I don’t think anyone was using it.) I’m happy to see that telling FeedBurner to redirect it to the source feed, then telling httpd to return “410: Gone” does exactly what I was hoping it would do, at least in NetNewsWire.
This is all I had to do in my
Redirect gone /ffg/index.commentfeed
I’ve changed the software that serves this weblog. If I’ve done my job right, most readers won’t notice much of a difference.
Even before I let this blog be down for an entire year, I'd pretty much stopped writing in it completely. I was trying to figure out why that happened -- early on, I was writing, or at least linking, stuff pretty much every day. So I looked back at the old entries, and a lot of them were one-liners, posts of pictures (my own or ones found on the net), quick links.
Looking at other people's blogs, I notice the same thing. It's no great mystery, then. When I have a quick thought, it ends up on Twitter. If I see a silly pic, I post it on mlkshk. If I take "better" pictures, they get posted on Flickr. Collections of links go to Pinboard. Stuff that doesn't really fit elsewhere goes on Google+.
It's not perfect, though. Anything you post at a place you don't own isn't really yours (I'm thinking of all the stuff I posted on Delicious, Posterous, LiveJournal -- stuff that, if it even exists anymore, is in a form where it doesn't do me any good.)
I'm not promising I'm going to suddenly start blogging again like it's 2002. I'm just saying, well, having your own place is nice.
After almost exactly a year of downtime, I finally got a real (small) webhost and stood my blog up again.
For its entire previous existence, this blog lived on whatever desktop Mac I happened to own at the time. In retrospect, I think I just did it to prove, yes, Macs are actual Unix machines (see, real servers), but, of course, using your desktop machine as a live webserver is really an unnecessary pain in the ass.
Moving it over to an actual webhost was really, really simple: less than 3 hours of work from creating the VM to a fully functioning blog, and that includes the time I spent trying to find the right backup and rsyncing a bunch of crap around. For abandonware, Blosxom is still a pretty robust bit of software.
I’m sure there’s some broken stuff (most of work was weaving a scary web of symlinks so as to avoid having to edit old entries), so if you see anything obviously borked please leave a comment.
I may even start writing again. The past 18 months of my life have been ridiculously eventful.
It was buggy, and I didn’t have any evidence anyone was using it, so it’s the dark theme or bust. If you’re really attached to the old white theme, I’ll bring it back, but I’ve shaved enough yaks for one morning so I’m turning it off for now.
Hardware problems, the usual, whatever.
I haven’t actually tested it externally, but dynamic page generation should be much, much faster. It still doesn’t get around the problem of me not ever really writing anything, but “Small moves, Ellie. Small moves.”
I have a new miniblog for linking to linking to visual stuff, silliness, and random crap. I have a couple of weeks worth of content(?) there already. I finally got around to doing this because Pulse makes it pretty much effortless. If you have an iPad, it’s one of the first things I’d recommend you install.
Actual writing, of course (whenever that happens) will still live here.
Chris added location stuff to his blog and blamed me. :) His solution was a little more technically involved than mine— he used an actual framework, where I just parse a feed into its constituent atoms (pun intended.)
(Disclaimer: as Chris argues, and I fully agree, I’m not worried about the privacy implications of the service— my stored location is only updated when I choose it to be, and I completely control the update frequency and geo-resolution of the updates)
I signed up for Yahoo’s Fire Eagle long ago, back before I even had an idea of what it could be good for. I liked the architectural purity of putting location data in one place and permitting apps to use it, as long as I could opt-in / opt-out at will. Rather than updating several “silos” with location data, Fire Eagle provides a location store and a defined API for apps to fetch it. Fire Eagle looked like it would be fun to play around with, but I didn’t do anything at all with it for a few months:
A few months ago, that changed. I signed up with Brightkite during their closed beta. Brightkite is a social network that lets users leave 140-character updates, photos, etc. (sound familiar?) but with one killer extra feature— every update has a geographical component. Optionally, you can just use it to “check in” at locations. When you do so, your Brightkite location gets updated, and if you “connect the dots” so does your Fire Eagle store. They’ve got a very nice mobile site and an even better iPhone app.
A guy named Richard Metzler wrote a mashup-enabler called Eaglefeed that talks to the Fire Eagle service and provides location data in several easy to digest forms: namely, Atom, JSON, and GIF. I julienne the Atom feed with the feed parser and plop the peelings into the left sidebar of the blog along with a PNG version of the map graphic. Like almost everything else in the sidebars, I display the location info statically, refreshing it once an hour— since I generate my blog pages dynamically, making the sidebars static keeps page load times reasonable.
For the 0.0 of you who care, I did some long-overdue cruft cleansing from the ffg front page. The template shrank in size by nearly a third, and (more importantly) I dropped two old script files that were almost never used. I haven't profiled it, but it certainly seems to load faster now.
It’s probably not useful to anyone other than me and Dad. :)
I’d heard about Tumblr before, but I’d never quite figured out what it would give me that I couldn’t already achieve via this blog, or Twitter, or del.icio.us, or moodmat, or any of the other places I sporadically attempt to bore people. I’ve finally figured it out — it works best for things that are a little too substantial to be tweets, are more than just random linkdrops, but don’t quite justify a full blog entry (general purpose nor music specialty). So I have fractional horsepower freeform goodness.
The stalker feed covers it all, anyway, and then some.
The newly free NetNewsWire has this way of letting you know that your weblog isn’t updated enough — it turns its display of the feed title a lovely Zune-y brown:
I’m in Philadelphia for the week, if anything fun happens I’ll let ya know.
OK, not really. I'm about to move to Leopard on the machine that hosts this blog, so, if it doesn't come back and you can't read this, you know why. :).
Quiet over here but we're posting like crazy at Moodmat. Take a look.
There has also been some work to allow the interesting use of macro names.
For example, if you wanted all of your “creat()” calls to include read
permissions for everyone, you could say
#define creat(file, mode) creat(file, mode | 0444)
I would recommend against this kind of thing in general, since it
hides the changed semantics of “creat()” in a macro, potentially far away
from its uses.
To allow this use of macros, the preprocessor uses a process that
is worth describing, if for no other reason than that we get to use one of
the more amusing terms introduced into the C lexicon. While a macro is
being expanded, it is temporarily undefined, and any recurrence of the macro
name is “painted blue” — I kid you not, this is the official terminology
— so that in future scans of the text the macro will not be expanded
recursively. (I do not know why the color blue was chosen; I’m sure it
was the result of a long debate, spread over several meetings.)
— From Ken Arnold’s “C Advisor” column in Unix Review