0xDECAFBAD

The latest mini-sensations to arrive through my firehoses are identi.ca, a Twitter-clone / microblogging site, and the Open Source software Laconica, which powers the aforementioned site.

Having started and neglected two Twitter cloning attempts of my own, Cuckoo and OpenInterocitor, seeing someone else carry the torch with any modicum of momentum is attractive to me. So, I spent a little bit last night getting the code running on my own servers, and managed to do it twice:

• decafbad.com/laconica
• lmorchard.com/laconica

See, the interesting thing promised by Laconica—and something I wanted in my own clones—is the ability to federate instances of the software. That is, users on one Laconica-based site should be able to subscribe to the updates from users on another site, by way of the OpenMicroblogging specification. Although federation isn't a silver bullet to a web-scale Twitter clone, I do think it's one of the most important bootstrap steps—but that's another blog post entirely.

Thus, since I'd like to see you run a Laconica site (or something like it) for mine to talk to, I figured I'd document how I got the thing running. My server is running Ubuntu Gutsy, so your mileage may vary. This is a long one, so check out the how-to after the jump...

Read More »

Update: Oh, and rumor has it that this widget will switch to reporting on downloads, rather than pledges, once the main event has begun.

Update 2: It didn't quite go like clockwork, but this widget is now showing estimated downloads by countries, rather than pledges.

Thanks to the completion of bug 435967, I can offer this totally unofficial hack of a mega widget (better late than never):

If you'd like, you can include this on your own site with a Copy-and-Paste of the following:

<script type="text/javascript" src="http://decafbad.com/2008/download-day-top-ten.js"></script>

This is a total 2-hour cut, paste, reformat, and slight rejigger of Kent Brewster's work on Content Syndication with Case-Hardened JavaScript. Hopefully it works, because I'm releasing early before I've had a chance to check it out on anything besides the browsers in my lap.

I use Markdown to write for this blog, and I don't think I could write much online without it. (Of course, that's not saying much lately—but nevermind that for now.)

Although I do appreciate all the effort put toward enhancing <textareas> on the web, I'm not a fan of WYSIWYG writing interfaces. They hide too much in an effort to make things "easy"—yet, for all that ease, they make me miss Reveal Codes in WordPerfect 5.1 for DOS. Hell, I wish I could write books using Markdown instead of Microsoft Word. (And, yeah, I know that some O'Reilly authors have written books in perlpod—but that's a bit too far even for me.)

Anyway, I think the issue is that it really bothers me to compose HTML in the same mental context as writing prose—HTML is too much like code. Coding and prosing are two different modes for me, and I don't like the constant context-switching thrash. (And, yes, I've just learned today that prosing is a word.)

So, apropos of this, there's been a thread in the blogophere lately naysaying non-HTML, wiki-text, and smart-ASCII formats. They seem to be mostly attacking the idea from a technical perspective, where it's more like a usability problem to me.

I don't see where Jeff Atwood's "Is HTML a Humane Markup Language?" makes the case for HTML as humane. As far as I can tell, the inhumanity is that markup schemes are harder to keep track of than HTML. Yet, there's a kind of coder macho thing going on there that's decidedly not humane—i.e. real programmers should be fine with getting their eyes poked by angle brackets on a regular basis.

Anyway, I do agree that some of these formats are horrendous. Wikipedia markup and BBCode are nasty in particular. Though, Wikipedia markup at least has some semantic purpose beyond HTML. But, I can't follow from that to seeing how the whole idea is wrong.

I used Textile for awhile, and then switched to Markdown when it came around. At this point, I've been using Markdown since it came out in 2004. I now plug Markdown into everything I can where I'll be doing any writing. There's a version of Markdown in just about every language known to the web these days—you can even find Markdown in Javascript. I write notes to myself in Markdown, even when there's no URL to view it. I've even used Markdown semi-consciously when writing by hand in a paper journal.

Perhaps I've said too much.

Anyway, I'll expend as many brain cells as it takes to not need to compose raw HTML by hand in the course of writing. As it happens, I can usually just use Markdown, so those brain cells were spent years ago. And thus, I feel I've saved more brain cells in avoiding coding/prosing thrash. So, it could just be personal preference, but I agree with the reasons why the original wiki doesn't use HTML.

When I first created this blog back in February 2002, I was talking to myself and I knew it. But, that didn't really matter: Since becoming newly digital in 1983, I'd already spent years plugging away at personal projects big and small, frequently abandoned, talking to myself all the time anyway.

So, blogging, I thought, would give me a real or imagined motivation to finish some projects. The difference between writing a blog and working alone in the basement, is that there's a chance someone outside my physical neighborhood might catch a few mutterings and knock on my door to have a chat—rather than just walking a little faster and avoiding my house in the future, like the local neighbors usually do when I find myself really going off on a tear.

Thus if someone, somewhere showed interest in things I was doing and describing—then maybe I might just stick with one or two of them long enough to declare a version worthy of a freshmeat.net release. (It's been awhile, but that used to be my gold standard for Real Software.)

However, I've realized that occasional visitors won't magically help me finish software. I've come to accept that I'm far too much a serial enthusiast to release much at v1.0 without a paycheck or some other substantial incentive. But that's okay, because tinkering with unfinished software still seems to attract occasional visitors. And, where there are visitors, there are contacts and conversations. And, where there are contacts and conversations, there's learning and fulfillment—two of my favorite things.

Apropos of conversations, though, I've gone habitually quiet in the past few years. I've gone from claiming that hacking is my World of Warcraft to just getting sucked into WoW proper. It's gotten to the point that I've wondered what I ever talked about here. Yeah, I've written this kind of entry many times before—but I don't feel it's gotten better.

I'm really out of the flow here, but I don't think giving it up is the right thing to do. I've been doing some retrospection (in volumes one and two), and have been reviewing what I've done here. I'm getting some ideas, and hoping I can get myself back into acting on them.

It might be a little while before I produce anything on a regular basis, because I'm currently working on a third book, entitled Professional JavaScript Frameworks. And, on top of that, I've got a bit of a transition going on in my life right now.

But, waking things up here is something I'd like to make part of that transition.

As you may have noticed, I've been shaking things up around here a little. Sorry for turbulence in the feeds! I've moved to a new server, and I've been trying out Movable Type again, but with mixed luck.

Why? Because I've been using WordPress for a few years, and I felt like a change. Maybe in another few months I'll convert over to Blosxom again, or try ExpressionEngine for the first time. Now, if only I could get all the timestamps right so some of my late-night posts don't have broken permalinks due to day changes. Maybe I'll end up just sticking with WordPress after all.

But anyway, I'm also having another bout of periodic existential doubt about what I'm doing here with a blog. So, in navel-gaze mode, I spent time a few nights ago skimming and reading through all 1000 or so entries I've written here. I do this periodically with the scores of paper journals I've accumulated to recenter and reconnect with myself, but I've never done it here.

So, for my own reference more than anything, here's 0xDECAFBAD Greatest Hits Vol 1, the Tinkery Bits.

It may be presumptuous to ask, but if anyone actually reads through this list and thinks of something worth surfacing that I missed—by all means, let me know. I'm trying to remind myself what I started this thing for.

• 2008 Jan 31: Metablogging
  • A look at all the different tools I've tried using for blogging lately.
• 2007 Dec 24: An Accumulator for all my stuff
  • About building my own lifestreaming thingy.
• 2007 Oct 17: OPML reading lists in FeedMagick2
  • More work on OPML reading lists and merging feeds.
• 2007 Jul 23: A Curmudgeon Playing with the CodeIgniter Framework and OpenID
  • Reporting on learning a new PHP framework.
• 2006 Apr 21: An S3 AJAX Wiki
  • Building a wiki based on Amazon S3 and client-side code.
• 2006 Mar 25: About XoxoOutliner
  • Building a browser-based outliner that deals in XHTML and XOXO.
• 2005 Dec 13: FeedMagick, the feed filter that doesn’t know much about feeds
  • Munging, converting, and merging feeds without destroying content.
• 2005 Oct 05: FeedSpool is progressing nicely
  • Fetching, caching, and spooling up feed updates without destroying content.
• 2005 Jul 12: In-browser XOXO Outliner Experiment
  • Early experiments in microformat-based outlining.
• 2005 Jun 08: Magic Microformat Forms Redux, Now with GreaseMonkey!
  • Expressing my enthusiasm for microformats in code.
• 2004 Dec 12: Building an Address Book as a Modern Web App
  • An unfinished attempt to demonstrate what everyone would soon be doing with AJAX.
• 2004 Nov 30: Next generation web apps using REST, XML, XSLT, and XmlHTTPRequest
  • Wherein I try to talk about AJAX before the the name was coined.
• 2004 Jun 16: Wish-of-the-Month Club, Part 1 of 3
  • The first in a series of articles building a wishlist project atop Amazon web services, an attempt at writing something publication-worthy.
• 2003 Sep 02: Using web services and XSLT to scrape RSS from HTML
  • Progressing from perl scraping hacks to cleaner XSLT-based feed scraping.
• 2003 Aug 31: Bookmark Blogger in Python
  • I would later abandon this entirely for del.icio.us, but I've been bookmark-blogging for awhile.
• 2003 Aug 22: Scraping HTML with curl, tidy, and XSL
  • Trying to bring some cleaner approaches to web scraping beyond regular expressions.
• 2003 Jun 19: Desktop server + Flash GUI?
  • Eventually, things like Adobe Air would fill this niche, I think.
• 2003 Mar 06: Building Pipelines with Web Services
  • I still haven't seen much work toward using pipelined HTTP-based web services.
• 2003 Feb 15: Say hello to Agent Frank
  • This project got abandoned, but I did release this overly-complex Java-based personal HTTP proxy.
• 2003 Feb 04: Bookmark blogging from Safari via a quick hack
  • Prescient of my eventual extreme enthusiasm for del.icio.us, I started blogging my bookmarks.
• 2002 Dec 28: Universal personal proxies & agent companions
  • Initial thoughts toward Agent Frank and filtering / analyzing my own web consumption.
• 2002 Nov 26: XML-RPC, a mini case study
  • Wherein I talk about how XML-RPC has worked for me.
• 2002 May 31: RSS autodiscovery via the HTML LINK element
  • It's the orange feed button everywhere now—here's how it started.
• 2002 Apr 18: Auto-backlinks from blog entries via referral logs
  • Building backlinks from referer logs to help weave the blogosphere. It was great until the spammers ruined it, though Sam Ruby has a neat implementation.
• 2002 Apr 18: Pipelining the web makes for messy URLs
  • It's awkward to express pipelines on the web using URLs.
• 2002 Apr 17: REST. What is it good for?
  • At one point, I was pretty dubious about this REST thing.
• 2002 Feb 26: Almost done with another implementation of the Wiki XML-RPC API
  • I was working on XML-RPC APIs for various wiki installs at one point.