13 Nov 2005

fun license plates


spotted in the streets near mill road in cambridge.

... Read More

13 Nov 2005

comments on john gruber's review of the new powerbook

john gruber (daring fireball) has a philosophical review of the powerbook. interesting points:
"... I suspect the aluminum PowerBooks look and feel more rugged than they actually are. The corners, for example, are prone to denting, which in turn can lead to problems with the power adapter socket."
so true! however, i'd much rather have a dented powerbook than a cracked case.
"..iBooks also get better AirPort reception than do PowerBooks."
of course, what antenna would perform well if it was enclosed in aluminium?
"One area where PowerBook hardware is clearly superior to that of iBooks is the keyboard. I consider PowerBook keyboards superior in every regard."
i totally agree with this one. i've played with all sorts of laptops, and definitely the powerbook keyboard feels the most comfortable.
"Apple claims the new 15-inch PowerBook display is 20-some percent brighter than before, and, indeed, it’s quite bright for a laptop. Next to my 20-inch Cinema Display, however but not unexpectedly, it seems a tad dingy."
i would love a brighter powerbook. that is one of the things that is attracting me to getting a 15". i've never seen a laptop that has a brighter screen than a regular LCD. compared to the dell LCD and CRT on my desk, my powerbook is pretty dim.

... Read More

12 Nov 2005

10 most powerful women in blogging only has 9 women

haha, i read this article on the 10 most powerful women in blogging and was confused by why joi ito was on the list. as far as i know, he wasn't female or hadn't recently had a sex change. or at least he hadn't blogged about it.

c'mon, anybody who blogs about techy stuff should know who joi ito is. nice way for the fine fools network and who ever's blog this is on to generate some hits. i'm getting sick of these blog networks who are trying to emulate weblogs inc, gawker and 9rules right now. funny also, look at that blog and the "blogs that he hates" -- i suppose if i had a sidebar like that, i'd like to put ... nevermind.

... Read More

12 Nov 2005

google open house in london

i went down to london last night to check out the google open house party in london. partly because i was curious about the new google engineering office in london (hey, i'm supposed to be finishing my phd, what then?), partly because i needed to escape from cambridge (sigh!) and also for the chance to meet up with nick who i've only known from anil's mouth and silc.

well, lets start off with google first. they've got some offices near victoria station in london. a nice new building called belgrave house (flash), they're nearly the first tenants in there, or so it seems. the party was so-so. lots of people, nearly an orgy of google colours. the desserts we very nice, lots of free booze (thats where your money is going google shareholders :), trademark scrolling google searches displayed in the lobby.

why are people searching for anil sex?

learnt a few things i didn't know, such as that they have an engineering office in norway in addition to the one in zurich. they seem to also have one in japan, but they didn't send anyone over. another interesting fact is that they're spending some time improving translation services, even won some translating competition using statistical information from their database.

although, the point of the party was a bit loss. they didn't seem to be recruiting yet (no engineering jobs here), even tho they only have very few engineers there. it seems like they are going to headhunt some kick ass people first before experimenting with hiring new blood.

we did manage to leave the party with a bit of google swag. alex who came down with us also managed to win a bluetooth headset for his phone which may or may not have bluetooth. he so nearly won a video ipod! we also got a t-shirt which is too big for me, and boutique style paper google bag and a flashing annoying google badge that i'm going to use for my rear bike lights.

turn off the bloody lights google
(turn those bloody lights off, google)

other non-google things, we had some ramen and sashimi near leicester square at zipangu. their ramen wasn't spectacular (it didn't taste like the ramen in japan), but at least i had sashimi. why is it that they have sashimi in london for a reasonable price (ok, reasonable-ish) and double the price in teriyaki in cambridge. surely an hours ride for the dead fish can't be that expensive. lastly, finally managed to meet nick in real life, didn't get enough chance to really chat because of the bloody hypnotic google colours, but there's always next time!

oh for the thrifty uk readers (yes, all 4 of you): apparently you can get 4 tickets for the price of 2 if you travel in groups of four when travelling on the train. cheaper than using your young persons rail card.

... Read More

12 Nov 2005

liquidx.net updated to a less eyesore colour scheme

i've just added 2 or 3 images to the site and changed the colour scheme a little to remove the clash of colours that was polluting the blog. sorry for those who might have been hospitalised by the ugly colour scheme.

i've just been so busy recently that i haven't had time to revamp the theme. promise that it'll look better soon. the most pressing issue is the rip out django's templating system to i can generate static pages on non cgi/python sites to the same theme. anyway, i'll get there eventually.

... Read More

08 Nov 2005

cambridge black spots

cambridge evening news has an article about questionable hygiene at some popular restaurants. i don't think i'll be eating out any time soon.

At Tandoori Palace, Histon Road, Arbury, spilt rodent poison was discovered in the storage shed next to the onions in March 2004. The poison was laid down after a rat was spotted entering the shed.

... Read More

04 Nov 2005

choosing a personal note taking wiki

i've just recently stumbled across pwyky, an extremely lightweight wiki implementation in python that is self contained in a single python script and stores things in html form in text files in a single directory. it's simplicity makes it incredibly brilliant. just copy this python script into a directory on your web site, rename it to cgi (or anything that is going to be executed as cgi) and off you go.

pwyky is ideal for personal note taking where you don't need all the cruft of revisioning and user accounts, etc. you're just presented with a simple page, the ability to edit it and a cruft free mark up syntax. the wiki itself stores all the files as html (which makes alot of sense) and it translates that back to wiki style syntax when you want to edit it. being in html form, it is easy for you to move all your notes to static form or possibly even something else without having to write a decoder for the millions of wiki syntax's out there.

apart from pwyky, i've been using voodoo pad lite and notational velocity. both great in their own ways. i really enjoyed notational velocity's, emm .. speed and minimalism. lightning fast searches paired with great keyboard shortcuts makes it a dream thing to use for taking short notes.

voodoo pad on the other hand is a rich text editor with built in links like a wiki. you edit it like a wiki and navigate it in that way as well. the interface initially very attractive and able to store all sorts of things like pictures and fonts soon gets in the way of quick note taking.

what the ideal note taking software should do is combine notational velocity' simplicity with a simple link markup, and then add a dash of totally open text exporting features. when i have some time, i'm going to have to hack on a replacement for notational velocity and maybe combine that with syncing notes to a web service, or even the dreaded .mac.

side note on progress with spotstamp

although, before i do hack on that, i need to finish my spotstamp project that has been polished a bit more by being able to handle multiple files in a nice way. search interface still needs some work and still have to solve the dreaded problem of metadata being lost on copy or a user's accidental flushing the contents of the metadata server. unfortunately, apple has decided that copying a file means you shouldn't copy the metadata too. so if you download a file, safari puts the origin url in the metadata, but if you copy that file, you lose the data on the new file. why they didn't make (at the very least) finder copy preserve the metadata is anyone's guess, but thatis a major minus in my book.

the major thing that is holding me back from releasing a preview beta is the fact that i need a way to backup the user created metadata content in case these scenarios happen, and so the users won't get frustrated with losing all their hard categorisation work by running some random command someone's given them to cure their spotlight problems :(

... Read More

01 Nov 2005

free .be domain names

(via threadwatch)

apparently you can register a .be domain for free until 31st january 2006. after that it costs 16.50 pounds to renew. doesn't apply to existing registration. dunno, could be fun to think of interesting .be domain names. most of them are already taken though.

nice way to lock your own citizens/company out of domain names, belgium.

... Read More

01 Nov 2005

new web design slowly coming together

due to my lack of time on photoshop, i've been slowly cobbling my web site from pure html and css. i like doing web design incrementally, seeing what works and what doesn't and slowly updating the site. once i know that i have a layout i like, i'll then spruce it up with some graphics.

so here i have some tabs on the top that i feel are quite cute -- although they don't come out very well on internet explorer. the pager will eventually become a scrollable javascript thing. i haven't quite sussed out what sort of interface for the pager should be, either by date or by item number.

lastly, the feel of the page is very similar to apple's front page, the main headline story (blog entry) slap bang in the middle, followed by three dynamically updated boxes of information. i'm going to try to restrict myself to only three, because its the right about that doesn't overflow people with useless information. although the information contained is rather useless :) where it starts to differ is that scroll down off the page and you'll get your regular front page. so if you enjoy reading some older entries, they're easily accessible.

there's still more to do with re-theming some of the static pages. i need to figure out where i can rip out django's templating system to do static pages like i did with phptemplates in drupal.

... Read More

30 Oct 2005

bare minimum xmlrpc support on my blog for ecto

the last couple of days i've been having some ecto withdrawal symptoms. this is purely caused by the lack of xmlrpc support after i switched to django. finally, between glibc and kernel recompiles i've rectified that problem.

it turned out to be much simpler than i thought in python. i basically pieced together the small xmlrpc.py django contrib from Amit Upadhyay and wrote some conversion functions from my blog entry model into the mt entry style suitable for sending over xmlrpc.

unfortunately, the view isn't that useful because everyone's blog entry is different, so it is rather hard to make a generic framework to open source for munging your blog entry into the movabletype api entries.

one bonus thing though, my problems with ecto doing weird things with my linebreaks has disappeared with my new blog. i'm not quite sure how or what happened, but i suspect it is because my new blog doesn't attempt to be smart and return any text filters, so ecto doesn't get confused by it.

finally, the one thing that would make ecto perfect would be to add a text box so that i could type in tags in a string format in delicious style rather than clicking on check boxes. meanwhile, i'll probably have to add a hack to give some tag hints inline in the blog post and strip them over on the server side.

... Read More