nick.recoil.org

I’ve just got hold of my first edition print of Computer Lib/Dream Machines by Ted Nelson. It’s an amazingly inspirational read, and covers a bewildering number of topics. It’s a shame that it’s a little too large to take on holiday with me to Italy.

For some reason, I was reading the Slate article on the mysterious Short Capuccino from Starbucks, and I thought I’d try it out in my local. The guy serving me, as described in the Slate article, didn’t bat an eyelid when I asked. 1.95 UKP for a regular Tall cappuccino, and 1.75 UKP for a Short cappuccino, and it’s definitely a better amount of milk in my opinion.

This has been brought to you by the trashy coffee department

Today was an example of the perfect English summer's day, with a clear sky, temperatures around 26^oC, and a cool breeze, all enjoyed with a glass of Pimms in hand. As part of the Open Garden Squares Weekend, ours (Nevern Square) was open to the general public, and while Jasper and I were lazing around the flat, we heard a lovely bit of light jazz drifting up from outside our window. We decided to investigate!

It turns out that the Garden Square Committee had organised a Saxophone quartet to come and play in the square between 2pm and 4pm. The Glyn Williams Quartet were four very nice gentlemen who kept everyone entertained whilst many of us were eating our picnics (including Jasper and I), or having a browse around the Herbaceous borders.

The source of the movie above was taken with my JVC Everio GZ-MG36EK, which I lovingly refer to as the iPod of the camcorder world. No more fussing with DV tapes, this is captured straight to an internal hard disk drive. The only slight issue I have is with the range of editing tools that support GOP-style editing on MPEG2 streams. They are few and far between, especially if you're using a Mac.

I've moved this blog installation from my locally hosted MythTV server over to the recoil.org box we've designated for hosting dynamic content. In doing so, I've re-hosted all of the content from MySQL to SQLite, as this is going to be easier to maintain in the future in a secure and simple way.

In doing this, I've discovered that there's apparently no neat way of doing this through the current Rails rake tasks. Eventually I've settled on preparing a blank SQLite database via:

rake db:schema:load

And then running the following script, which was modified from something found on the SQLite trac installation:

cp db/typo_blank.db /tmp/temp.db ; /usr/local/mysql/bin/mysqldump --no-create-info --compact --extended-insert=FALSE --quote-names=FALSE --complete-insert -u root -pYOURROOTPASSWORD typo | perl -pe '
        if (/^(INSERT.+?)\(/) {
                $a=$1;
                s/\\'\''/'\'\''/g;
                s/\\\"/"/g;
                s/\\n/\n/g;
                s/\\r/\r/g;
                s/\),\(/\);\n$a\(/g;
        }
' | sqlite /tmp/temp.db;  mv /tmp/temp.db db/typo_dev.db

And the last thing I found I needed to do, which may or may not have something to do with being on the SVN trunk:

UPDATE contents SET published='t';

After the migration, the contents.published column seemed to contain a 1 instead of a 't', which is related, I assume, to the way the differing systems choose to store boolean values

I've installed the breveCreatures screensaver at work, which has been entertaining my co-workers whenever I'm not at my desk.

It sets up a simulation that evolves a walking creature based on a simple genetic algorithm. The state of the simulation is saved each time the screensaver is interrupted, the state is saved, and will pick up again the next time it starts. Very entertaining eye-candy.

I've also got so used to certain aspects of Firefox extension functionality, I've installed SafariStand to enable my typical browser session state to be saved and loaded with ease. Not quite as nice as having it saved when you quit, but it'll do.

Things I have been enjoying:

• Samorost 2 - The followup to the original, which was released back in 2003 (via BoingBoing)
• Google Earth for Mac OS X - Reasonably snappy on my 1.67GHz AlBook (via Haddock)
• This new theme, called Origami, from a talented chap called Leevi Graham - Winner of the Typogarden theme competition
• The new V-Sag section of hybridized.org - More good music sets to work to

And I should also mention that the explosion at the Buncefield Fuel Depot managed to wake me up this morning. I was just dozing back to sleep after feeding the cats at 5:30am or so, and I suddenly heard a boom, and felt all the windows shake at the front of the house, and a car alarm started to go off somewhere in the distance. Very peculiar, I thought, and only put two and two together as I watched the news later on. There are some spectacular pictures on Flickr

I've uploaded a picture to Flickr demonstrating the artifacts you get from Flickr when submitting a photo that requires resizing to fit their maximum horizontal resolution of 1024 pixels.

I think they're definitely applying a second-stage sharpen process, rather than the sharpening being inherent in the choice of spacial filter used. This discussion came about because of a discussion on Haddock on the ways in which Flickr can mess with your originals.

It also seems that Flickr will strip some of the image metadata in the JFIF headers that relate to the way in which the image should be displayed (such as white point), thus causing differences in rendering between an image which has been uploaded and downloaded via Flickr, and the original, untouched picture.

Just a quick rendition of the four types of Sparklines as allowed via the library for ruby At the moment they are a bit throw-away, but if they can be integrated into some dynamic data where you care about trends rather than figures, such as load average, then it could be useful. The other feature I think it's missing is the lack of being able to tag a point with a value, such as this example up in the top left corner, as does the original post from Mr Tufte. This is crucial in presenting a single numeric value, and the reader being able to infer whether this is high, low, or average.

Anyway, on with the Typo show...

Update: Since I've now moved to Mephisto, I've ditched the content of this post

Dominic and I attended the last CopyFighters meeting of the year, where Dom gave an impromptu speech on the trend of technical standards bodies. Instead of merely creating specifications on how to package and distribute data, they are becoming increasingly intertwined with mandates on how these technologies become deployed in the market place, and what restrictions are placed against what can be done with the content.

You can see more pictures on Flickr

I'm giving the Flock Browser a trial run, and am posting this through its interface.

Blog entries are rather simplistic at the moment, but the del.icio.us integration is pretty nice, although I'm trying to get over my innate reservation over all of these shared resources. The benefits you have of sharing your recent activity outweighs my desire to be the owner of the services themselves.

Centralisation vs. decentralisation

My approach to anything related to the recoil.org domain Anil and I own was to build it ourselves. This was very much what was necessary when we bought it back in the late 90's, and it was also our chance to educate ourselves in what was required of basic unix systems administration.

Skip forward several years and here we are in 2005, year of networked social services. Small scale community aggregation sites like sn17 work well, but are limited to those individuals who can be bothered to explicitly arrange to take part.

The strength of flickr, del.icio.us, last.fm and all the rest of the other services that have sprung up has been in the amount of effort it takes to make social connections, and the effort involved in information derived from those connections to be displayed to you, the end user.

My problem is that I have no backup of the data held within their systems, unless the data was originally on a computer I own. This is true for a majority of the Flickr pictures, where I use an export program from iPhoto, but what of the ones sent via a mobile phone? I have no such protection against loss of data, or of more concern, changes in terms and conditions of the site. One particular example springs to mind, which was the ownership of CDDB entries, when the company went from offering what was in its databases for free, to a paid model. Many users were up in arms over this, but there was nothing that could be done.

I have the same fear in the back of my mind with all of these internet services, which stops me making extensive use of them, as much as I would if I felt my interests were entirely protected.

I didn't notice before now, but there's a beta release of Omnigraffle 4 out. Looks extremely nice, with Layers support, and SVG export capabilities to name but a few new features.

I've been playing with the Lomo filter for the Gimp from Goose24 after reading a tutorial on faking a lomo look (via The Tao of Mac). Take any boring looking picture with flat contrast, put it through the filter, and out pops a much nicer looking snap.

It seems it works best on pictures that haven't necessarily been through an 'auto-levels' process, and flat areas of colour obviously help, where you can see the effects of the slight vignette. Throw-away, but nice.

Update: It looks like I've lost those pics after a spot of unexpected data loss. Easy come, easy go!

I've added my name to the growing list of people who want to sponsor an organisation to campaign for digital rights in the UK. You can find more details a the Pledgebank site. It was started during the Open Tech 2005 event in Hammersmith last week.