(mu)blogs are the new trades

Blogging and microblogging tend to get disrespect from folks who “just don’t get” the purpose for them and consider them at best mindless entertainment and at worst an attention-sapping pest.

As a second-generation programmer, I found that I “got” them pretty quickly.

My dad programs for embedded and industrial-control systems; I grew up watching him bring home stacks of trade magazines — not to read every article in detail, but to skim through as an environmental scan, updating his awareness of the state of the art. If anything the ads and editorials were far more useful to him than the articles!

As a web developer in the 2000s, I started to use blogs and microblogs much the same way: little bits of information here and there which fill in my background map of what’s current among my peers (say, everything awesome in web browser work).

Scars of the past

Poking around Google Earth to optimize a hypothetical bicycle commute to the office, I noticed an unusual diagonal strip of alleys and divided lots running through San Francisco’s Mission District from about 22nd & Harrison to around Cesar Chavez & Guerrero:

Mystery line!

I figured it had to be an old rail right-of-way, but wasn’t sure what for. At first I thought it might be BART-related, since it runs right by the 24th & Mission station, but BART runs underground straight north-south to the 16th St & Mission station.

A little Googling on a portion of the strip cryptically labeled “Juri Cmns” turned up a page about the Juri Commons park which explained that the strip was originally part of the San Francisco and San Jose Rail Road. After the 1906 earthquake, the portions running through the Mission were torn up replaced with a line running nearer the bay — the same route today’s Caltrain operates on.

[Side note: I composed this post offline in Yojimbo’s rich text editor and tried to copy-paste it to WordPress’s WYSIWYG editor in Firefox. Total paste fail when including the picture; links stripped when pasting paragraphs only. Tried it again in Safari — all links came through intact, just missing the image.]

Notarista mockup

I’ve mentioned a few times that I’m a big fan of Yojimbo for note-taking, collecting web receipts, storing encrypted passwords, etc.

The biggest two failings of Yojimbo for me are:

  • Mac-only — no Linux or Windows client
  • No native web client; WebJimbo is more or less functional but very awkward
  • Sync is tied to MobileMe, and gets increasingly fragile as your library grows

I finally reached the point where I couldn’t sync between my iMac at work and my laptop for home & travel, so out went the iMac. :(

I’ve become quite enamored of my Linux netbook for travel, but limited access to my notes and password store makes it limited for me. I’ve just gotta replace Yojimbo…

There are some other apps in the note space which I haven’t been too happy with:

  • Tomboy — Mono/GNOME-ey note taker. Decent for quick notes, but sync is hard to set up, the Mac/Windows ports are immature, and no support for images or encrypted passwords.
  • BasKet Notes — a KDE app. I find its interface confusing, and it’s insanely slow on my netbook. Doesn’t satisfy my syncing, web, or transfer requirements.
  • Evernote — this comes close… online syncing, nice web and Windows clients as well as Mac. No Linux client, no encrypted notes/password store, and transfer from Yojimbo loses metadata like dates and tags.

It’s beginning to look like I’ve just got to write my own. :)

I spent a couple hours tonight whipping up a UI mockup of Notarista using jQuery and CKEditor:

Notarista mockup screenshot

Initially it looks pretty much just like Yojimbo without the sidebar. ;) I’ll play with some alternate layouts that work better on a small netbook-size screen though; I think the note list can probably be hidden when not actively searching, for instance.

Update: I’m loooooooving jQuery! :D My UI mockup is now pulling sample data out of wikitech.wikimedia.org — search works and it’ll pull up actual page text you can pull into the editor. Of course it doesn’t save back… ;)

Epic sort fail

Was transferring some screen shots from my iPhone with Mac OS X’s “Image Capture” app when I discovered that the sort-by-date seems to have some problems:


Yes, it’s sorting by ASCII string value of the formatted date. 3/4 comes after 3/22, and 3/22/08 comes after 3/20/09. How’d Steve Jobs let this one out the door? I can only assume nobody had a memory card or camera with old photos on it when they tested…

Updates a-coming…

Last week was busy — we had our Wikimedia Foundation all-staff meeting here in San Francisco, so some of our out-of-towners were in for planning and wackiness.

Plenty of new stuff to come over the next while…

This week so far:

Coming soon…

  • A big code review update and bump to the live sites will come in the next couple days. Yeah, yeah. :)
  • Drafts extension will go live.
  • Fixes for video on Firefox 3.1beta…
  • Hopefully, more support for large file uploads — we’ll try some experimentation with upload-by-URL.
  • Tomasz will be summarizing architecture work on the data dump design.

And I’ll have plenty more to blog about…

Applying for Summer of Code 2009…

I’ve just put in Wikimedia’s org application for Google Summer of Code 2009… Hopefully we’ll get in. :)

We’ve had mixed luck in previous years with GSoC, but I think we’ve got enough internal bandwidth this year that we can make sure there’s enough effort put into interacting with the student candidates ahead of time to pick the coolest and most go-get-em self-starter awesome projects and then support them through the project term.

I’ve tossed up a student application template if you want to get started early. :)

Update 2009-03-18: We’re in!

Testing user testing

Part of the Wikipedia Usability Initiative‘s work plan is to do some genuine real end-user testing — this’ll give us a more solid idea of a) what problems to prioritize and b) a solid measure of progress as improvements get made.

An initial test run is starting now: a banner is running for a small percentage of English Wikipedia anonymous page views inviting folks to participate in the study:


This leads to an online survey thingy which the testing firm we’ve contracted uses to pick out candidates for remote or in-person testing:


A subset of the respondants will get a callback from the testing firm, and in a couple weeks we’ll get folks in a lab and smack ourselves with how confusing our site is while we tape them. :)

Currently we’re mostly targeting San Francisco Bay Area locals; next time we think we can trim down the notice a little better to ensure it’s not showing to non-local visitors.

Inline video test

Just a quick test of Ogg-friendly video embedding with mv_embed in my WordPress blog…

Here’s Elephant Mud Bath.ogg from Wikimedia Commons:


  • Seems to work nice in Firefox 3.0 with Java
  • Seems to sort of work in Safari 3.2 with Xiph Quicktime components (have to play, then pause, then play again

Haven’t tested other browsers or FF 3.1 yet.


  • Still have to edit some settings in mv_embed.js manually to make it work (fix path, disable MediaWiki-dependent script loader)
  • Currently have to manually make/pick a thumbnail image
  • Seeking doesn’t currently seem to work (needs server support for oggz_chop?)
  • WordPress’s wysiwyg editor eats the entire <video> tag if I switch out of HTML mode! It leaves the <script> at least. :P
  • Integrating support into the upload/insert tool would be awesome of course, beating manual writing of the HTML. :)