Software over the rainbow

Some months ago I listened to this beautiful remix by 3spds. It’s based on a record of a religious group whose 900 followers committed a mass suicide.

He tells the story in a light but touching way.

At the time of this recording, they believed they had found the key to happiness and equality.

Next time I tried to login somewhere I still had them in mind.

Instruction set is a recently launched site that gives some instructions every month and asks for implementations in code. Think Sol Lewitt goes 2.0 (sorry I couldn’t resist). It is a beautiful simple idea and It’s stealing me lots of braincycles (see).

I’ve heard raindrops here and there about design by limitation lately. Number of chars in twitter, flickr video length, 1 photo/day in fotolog… There are a couple of limitations in instructionset which I missed at first but on a second thought I found interesting:

• There’s no clear way to post a result (images, applets, audio…).

You can always give a url, but it is not even suggested. I thought it might be an obvious improvement, but after trying to come up with solutions for the 2 first instructions, I’ve discovered that results could be a distraction. Instructions are about language and interpretation. I’ve found myself repeating the sentence, trying to define exactly every word, looking for a fold, or a flaw in which to introduce a deviation… Thinking how they relate to the programming language, how I relate to the programming language (specially with the first one “draw a straight line and follow it”)… I think that this relationship between everyday and programming language is a powerful one to explore.

• There is no form of comments/discussion other than the description of you implementation.

After seeing some of the contributions, I’ve found interesting approaches and I’d like to give feedback, but given the overwhelming noise ratio in today’s software playgrounds, I appreciate that there’s no way to do it. Paraphrasing Cage: just respond with your implementation.

Now, I don’t know if these are intended features or result of an early release, but if one could vote against new features… :)

Thinking about the traces we leave behind in all the software we use, I was wondering:

How will the biography of a data-intensive person – you know, blogs, twitter, delicious, tumblr, last.fm…- look like in the future? Will those “domestic privacies” be taken into account?

How will the fact that we have a clear, unambiguous point of reference for our own thoughts, places visited, people met, images, consumed products… affect the way we perceive ourselves? or the way we will be perceived in the future?

Just thinking out loud, to hear how it sounds like.

So I bit the bullet and have just opened a tumblr account. I’ve been somehow reluctant to try just another service (despite passionate, reliable recommendations). The way I use delicious (keeping an annotated track of interesting stuff) is pretty similar to what I could get using tumblr, so why should I add another tool?

The main point for storing things (es) is retrieving them. Common sense dictates that to find your things easily (and I’m rather obsessed with that), you should keep similar things in the same place, so the use of multiple services to keep track of similar information might seem counterproductive.

I have the tentative feeling that this is not so:

The richness of our homes results from our influence over their every characteristics and their accumulation of the traces of our activities. This richness is missing from our digital dwellings (by which I mean file systems, application windows, blogs, and mobile phones as well as the virtual locations in online worlds that more closely mimic physical homes).

Exploring personal digital archives for non-functional purposes. David A. Mellis.

One of the things I’ve learned to love about storing bookmarks in delicious is to forget about being organised at all. Taking advantage of idiosincracies related to personal experience, jokes, fuzzy relationships, ongoing thoughts etc. in titles descriptions and tags helps me find my information more easily. It adds the rich trace of my activities that software usually lacks.

Then, using different services adds, I believe, an additional layer of context. Episodic memory is characterized by having a unique context associated with a learning episode. The context of storing a bit of information is both physical (at work, at home…) and digital (where it comes from, where you store it…). Adding variety at this level might help a bit. You’ll save it in a slightly different way, and to a different place, for a peculiar reason. The little thoughts that you have when deciding to store it in either place will also be part of its context.

So my experimental feeling is that a reasonable amount (it’s not necessary to become francis bacon) of scattering could inject some healthy variety to our experience of digital information, giving us a richer context in which to manage our own data. To put it Mellis words again, we might be making our storage less efficient, but we’d be improving our memory of it.

And after all, the honorable 2.0 tradition of giving outer access to personal data makes scattering a simple problem of aggregation. You can always centralize your traces.

Let’s see how it goes.

Just a quick code tip for working with dates in mysql and php.

I’m building a script to regularly mirror my last.fm recent tracks to a mysql database of my own. I’m interested in experimenting with daily and hourly statistics.

One problem that has taken me some head scratching has been the fact that the last.fm recentracks web service gives me the date a given track was played in a timestamp of my current timezone (+02:00 , Europe/Paris, as I’ve just learnt), but my web server is on a different timezone (-05:00, America/Los_Angeles), so when I try to insert a date in the database, even though I’m giving it a timestamp, it gets interpreted as being 7 hours less than it actually is. So a track played at let’s say 31 Mar 2007, 23:55 would be stored as having been played at 31 Mar 2007, 16:55. Not good.

Not wanting to fiddle with configurations or anything beyond my knowledge, I discovered that I could set the timezone in both php and mysql for a given script or db query.

Setting the timezone in php

If you want to know which timezone your server is in, you can guess it with


date_default_timezone_get();


Which should give you a string like “America/Los_Angeles” or any other of the supported timezones list

If it is different than your desired timezone, look for the one you want in the list and before doing any time operation, call:


// set your timezone as gmt +02:00
date_default_timezone_set("Europe/Paris");


From then on, all your php code should understand timestamps and dates in that timezone.

Then for mysql use CONVERT_TZ

Even though it is corrected in php, you’ll have to do it also when inserting information on you mysql db, because it will interpret timestamps again in its time zone. You can guess which is it by having a look at the system time zone variable.

In my case, after trying some hacky alternatives I discovered CONVERT_TZ in one of the latest comments in this post

From the mysql manual


CONVERT_TZ(dt,from_tz,to_tz)


CONVERT_TZ() converts a datetime value dt from the time zone given by from_tz to the time zone given by to_tz and returns the resulting value.

You have to pass it the timestamp, current timezone (‘SYSTEM’ is the server time zone) and target timezone and it happily converts between them.

So now my query would look like this:

$query = "INSERT INTO recenttracks (artist, name, url, date, type) 
	VALUES ( '$lastFmTrack->artist', 
	'$lastFmTrack->name', 
	'$lastFmTrack->url', 
	CONVERT_TZ(FROM_UNIXTIME('$lastFmTrack->playDate'), 'SYSTEM', '+02:00'), 
	'$lastFmTrack->type')";

Basically that’s it. Hope it is useful for someone.

Mind you, I’m nothing of an expert in php or mysql, this is just a method which worked for me. Corrections are welcome.

Antecedent:

A couple of months ago, having a look at some old stats in my last.fm profile, I realised how much I could remember a given time by just seeing the music I used to listen to (that old (unfinished?) project by marcos weskamp and didier hilhorst came to mind inmediately).

The source:

Last.fm keeps weekly data about what we listen to. We also can (could) construct a radio link based on various artists. For example:

lastfm://artist/bibio/similarartists

The application

A web page that, given some artists that I used to listen to in a given time (e.g: december 2005), constructs the url of a last.fm radio with those bands, so I can somehow “transport” myself to that time by listening to similar music…

And I’ve done it, but…

Last friday, the multiple artists station feature stopped working. In fact, I had some suspicion, but I thought that… nothing. I didn’t think of it and kept working.

So it is basically useless in its actual form, but it was finished (as a proof of concept, at least), so here it is:

Last.fm time machine (if it had worked i’d have looked for a better name).

I’m not sure if the idea can take another direction to become useful. I’ve thought of having a look at xspf to see if I can generate playlists instead of radios, but by now I don’t really know. If you have any ideas…

At the very least, it’s been useful to clean the dust over my php, use the last.fm webservices, a little bit of ajax (thanks mr.sofa naranja) and above all, to finish something.

By the way, avidos let me stay in their hosting to do some tests while mine hadn’t php5, and ignasi tudela tried to help me with the design, but apart from using Georgia and taking his colors for the different seasons, I didn’t pay him much (deserved) attention, and you can see the results. Thanks to both.

Should we be learning from las vegas?