Screen Shot 2012 03 22 at 3 46 41 PM

A few weeks ago, standing at baggage claim with my family, I checked Twitter to see Austin Kleon mentioning me with a screenshot I didn’t recognize of something I wrote here a bit ago.

Totally out of context, I had no idea what was going on. By the time we were in our car I discovered Austin had used the line from that blog post in his book “Steal Like an Artist”.

Now a few weeks on, after reading the book, I can say I am so unbelievably proud to have such a tiny part to do with that book. It’s a subject I have talked about with friends, but never with the clarity and precision Austin does in this book. You should buy this book. You should buy two copies of this book because someone is going to borrow your copy and never give it back to you.

I have always wanted to write a book. I have these little ideas and thoughts I end up either throwing here or launching into over beers with friends. I also have had these thoughts about taste and how just simply having it and knowing you have it is good enough to get started. But this book is better than the book I could have written on that subject.

So maybe I never get to write that book, but holy crap am I happy to know I have a couple inches of space in the book that so perfectly nails it.

My dream writing platform is not Dustin Curtis’ new application called Svbtle. It’s a great looking application. I am jealous he was able to get it out the door. But as far as I can tell it’s not the one I’ve been looking for, so I figure it’s time I write this.

Screen Shot 2012 03 22 at 5 36 29 PM

I have been thinking about writing platforms and talking to people (some of which worked on the platform you’re reading this on) and even keeping notes in nvAlt whenever a new idea pops into my head.

The more I think about it, though, the more I think I will never have it unless I make it. And that is:

  1. an impossibility because I have a company now with employees and clients and no time to waste on side-projects.
  2. a really dumb idea that has always been a dumb idea (FOR EVERY DEVELOPER WHO HAS HAD IT) right up until the point at which you publish something with it and people read it.

Previously my solution was to skip over points 1. and 2. and jump to point 3: write a front-end that publishes (using the Atom/MetaWeblog/Blogger/MovableType APIs) with an abstracted connector. That is, you push a publish button and my web app figures out how to publish on your service of choice. Your site continues to serve but the underlying application (Typepad/Wordpress/Blogger) is hidden behind a sort of driver that does not know what system it is publishing to, just that it has content to save and update.

This is fine in theory, but turns out to be the sort of pain in the ass you are wanting to avoid by bundling all the worst things about writing a blogging engine into one code base. You spend a lot of time reliving the stupid arguments of 2003 while learning how to publish to the MetaWeblogAPI. You then get to revisit the creation of an entire blog publishing standard from 2004 (that materialized out of an entirely separate argument people had with the self-proclaimed godfather of RSS for reasons related to his self-proclaimationizing ways) when learning how to use the Atom API.

These standards are the wrong path. They solve jumping straight to option 3 but basically uphold the sometimes head-scratching standards people conformed to when there were very good reasons to conform to them. They’re outdated. They are holding blogging back. I can’t use nearly 10 year-old standards.

That brings us to my pain points with every tool out there. My idea of an ideal tool is just that. It has all the little quirks and features only I want and I don’t think I will ever get everything I want without taking a year off and drinking a lot.

Now, I know Typepad and Say: have a hell of a lot more things going on at their company than spending time adding new features to Typepad. I am a minor speck in their publisher galaxy and that’s fine. I think the application is still the best place for me to write since I use MarsEdit and rarely have to use the web interface.

But since Typepad doesn’t let me save drafts to the server (like Tumblr) so I can share them with friends (and since we’re spitballing: have my friends make edits in place), blogging is a very solitary sort of activity. The best posts (I think) I’ve made have been because Anil or Mat looked them over.

I need a new tool. I have reasons I don’t use Tumblr regularly (it is clearly the best publishing platform for blogging) and perhaps I should give it another look, but I think if I sat down with my nvAlt and described what I want it would not look anything like Tumblr. It wouldn’t look like anything that exists.

I hate that.

One of the things about having a child is you find yourself returning to things you loved as a kid. We went to Disneyland recently and every single damn ride was AMAZING. I was watching my son’s face light up as we turned corners inside of It’s A Small World when all of a sudden I was grinning at the dancing kids and singing along.

Same with Legos. Painting. Rolling around on the floor. And rockets.

I was a HUGE model rocket collector and launcher. Here’s a family launching one if you’ve never seen one go before:

We built every sort of rocket and dramatized stories around each launch. We pretended to be astronauts and walked slowly around empty fields like we were performing astronaut duties like fixing transponders (whatever those were).

So the other day I saw this:

It’s a Windows/Mac game/simulation that lets you do an incredible amount of planning, building, and launching of rockets into space, and to the moon if you wish. Apparently they plan is to also allow your astronauts to explore other planets as well. So cool!

My son is 3 so not quite ready for this, but I cannot wait until we finally get to the point where this stuff is interesting to him.

The Curator’s Code (more) is a cumbersome solution based on a good intention. Attribution should be stored in the meta-data of a tweet or post, not in the post itself. Unstructured text only complicates the job of any software trying to make sense of these codes.

Screen Shot 2012 03 19 at 12 28 23 PM

Twitter once promised to allow arbitrary meta-data in the form of “annotations”. The tools themselves could be encoding this within the post or tweet object for you rather than inserting into free text that can’t be parsed and aggregated. It’s too bad Twitter never did it, it seems like more and more we need meta-data to clear up the growing list of symbols and bad tech (URL shorteners) that clutter Tweets and blog posts.

Attribution and appreciation are good things! And we should seek to include them whenever we create/curate. But I think the execution of the Curators Code is flawed.

If anything the team that made this should petition Twitter and other publishing tools like Tumblr and Typepad to agree on a standardized method of encoding this kind of stuff in meta-data.

(By the way, I hooked up Typepad to my @andre_torrez Twitter account in case you would rather get notified of new posts there.)

So not only does my little idea exist, but they’ve solved some of the problems I thought I’d have dealing with films a user hasn’t seen. Flickchart takes a Hot Or Not approach to rating movies, and gives you a way to skip movies you haven’t seen.

I’ve been playing with it for an hour or so and it seems to be building a pretty good profile of the movies I like and don’t like. Check it out:

Last night I wanted to watch a movie so I went to iTunes and started sorting by stars when I realized something: I don’t generally like the most popular films in theaters, so why am I using everyone else’s star ratings to find good movies?

What does a four out of five star ranking actually mean?

Then I remembered an old idea I kicked around a bit before MLKSHK that I hope someone will someday make.

I originally registered the name which describes how I think it would work on the front-end.

The one sentence way to describe it is this: Drag and drop movies* to order them from left to right, ranking them as good to bad. The data generated from these, what I call “taste fingerprints”, would be used to help find other movies that people have ranked above identical taste fingerprints of others.

That is, if I ranked some heist movies like this:


And you ranked some heist movies like this:


The service would possibly suggest Sneakers as a movie you might enjoy because we ranked The Getaway and Reservoir Dogs the same. (I say possibly because it should rely on a larger set of data, not just my single ranking.)

The assumption here is if people rank movies in a certain order relative to each other as you rank the same movies in a similar order, then there is a likelihood you might enjoy a movie they ranked higher than that match.

You could do this with genres, but also arbitrary categories like Woody Allen films or films shot in a certain location. As long as someone creates a set and enough people contribute rankings there would be enough data to make guesses for you.

I will build this someday if you don‘t.

* or books, or video games, or iOS apps.

I’m a quitter. I quit using Vico the text editor for Mac that replicates many of Vim’s features in a Mac-like UI. (Earlier post on the subject)


I had a big day of coding last week and found myself flipping to Sublime Text to deal with a tricky multi-line selection. I was surprised to find out in the few months since I’d left Sublime Text it had received a multitude of updates and fixes. The tabs work and look better. It became difficult to switch back to Vico that day, and a this Monday when I needed to write a lot of code I went straight for Sublime Text.

Vico was very good and I was certainly faster. I could write code faster when I wasn’t having to learn a new keyboard command. And that was basically the problem, I got sick of learning. Yeah, I have to reach over and use the touchpad every few minutes while editing, but since I do that already in this (MarsEdit) and Google documents and any other edit window that is not Vico it seemed strange to have to switch my mind to vim editing.

I switch Twitter desktop applications often. For reading on iOS it’s either Twitterrific or Tweetbot, but on the desktop I want something much more robust with all the noodle-y little features a heavy user like myself would like to waste/save time with.

(So I decided to write them here, but it’s times like this I wish I had a wiki-type site instead of a weblog so I could keep editing this file and have it rise to the top.)

  1. Show me what my friends have liked. Similar to but I want it in an app and without a follow limit. Also I would love to see the most popular or retweeted tweets for the past 24 hours in one view.
  2. Ability to turn off all retweets or only show retweets. I love retweets because they usually have great links in them, but I also dislike them when they get in the way of simply catching up with people.
  3. Save for later. I hate using favorites to track save things I want to read later.
  4. Twitter’s Tweetie-era app still allows for keyboard commands. This is essential for me. I am not so sure they will stick around when Twitter launches the next version given their iOS (BUY TAMPONS SPONSORED TWEET) release.
  5. Ability to quickly add someone to a must-read list. I keep some private lists for stuff I absolutely can’t miss, but adding and removing people from it is so hard to do.
  6. Pre-load every image and display it inline. I hate waiting for images to load even though they’re already thumb-nailed inline. I have bandwidth to burn, please use it.
  7. Separate Webkit view for URLs. Two reasons for this:
    1. I can trace where the URL was originally opened from.
    2. If it has tabs I can keep my work browsing separate from my Twitter browsing.

Itchy coding fingers wish I could make this. But…if I were going to dedicate some time to a desktop application it’d probably be an email app that was smart enough to know what to do with email as it arrives. Hotel reservations? Add to iCal. Super extensible rule plugins. Keyboard navigation. Better filing. Blah…blah…blah.

Mat’s post about Twitter making a tiny change to the way @ replies work reminds me of an idea I had. Someone please do it.

Basically every night a set of scripts run that test whether all the things we know about Twitter: how replies work, how RT’s work, how URLs are shortened, how following works, and it publishes the results into a nice one page screen people can check in with when a new version rolls out.

In addition the data returned in JSON responses would also tip that there are new features to check out.