|♩♬♫♩♪ |

Aller au contenu | Aller au menu | Aller à la recherche

jeudi 4 mars 2010

samedi 27 février 2010

arsed

<drobilla> I just can't be bothered to explain The Great Culmination Of All This Stuff to people

a mail client ive been working on provides a decent demonstration ground for "all this stuff", my version anyways. daves off in mutopia somewhere

where are we going

IOW, i need to reconfigure the interface with regularity

mail is just another message, with its own peculiarities

so back to views on my stream

  • Filesystems*

my stuff has to go somewhere. im not about to upload it to some silo where i cant hack about

i'm not going to upload it into Riak, because by next week i'd proably like Redis better.

danbri made a great post about NoSQL. check it

to compose a message , i can use the mailto window that pops up in a browser, or mutt. I use emacs in mail-mode, so this is already covered

i use getmail because the name is intensely self-descriptive

the fact that SPARQL came out of W3 is remarkable, which completely eschews REST, in favor of a RPC "endpoint"

it becomes quickly apparent you cant use a SPARQL endpoint to efficiently serve a document

why? because a resource actually contains subresources - items with an unique "id" inside a document

http://volkswiki.org/Cars#Karmann-Ghia

the browser strips off #Karmann-Ghia and sends the rest.

once you have http://volkswiki.org/Cars, youre suppposed to find all these resources

in SPARQL, youd have to do something like this:

 SELECT DISTINCT(?S) WHERE ?S ?P ?O FILTER(REGEX(str(?S)+'#*'))

basically run a string search on every URI in the system - this could take hours. you could develop workarounds including wrapping insert scripts to duplicate all the triples into a graph at the document root's URI or so, and then join on all of them in your queries, not necessarily efficient in space or time however

so basically documents have these 'children' which are almost never expressed in triples themselves, but are needed right away

linked-data tutorials tell you need to create a

cars.rdf, and 303 to it - a file @ a particular point in a filesystem

also the path component inside URI schemes is well known and used by Apache, PHP and so on

vendredi 26 février 2010

rdf

<DanC> The RDF semantics doc is about 7000 times more complex than it need be.

a bit of an exaggeration but i think the complexity of the specs and RDF/XML format does not encourage one to cast off hesitation

what ive been working on with element is bringing an ease of developing RDF enabled applications for those who expect the productivity of eg Rails

serverside web development is about 2 to 3 times more complex than it need to be

mercredi 3 février 2010

Name binding blog

A blog all about techniques for handling name binding

mardi 26 janvier 2010

Caroliner

mercredi 20 janvier 2010

Caitlin

this guy had emailed my friend at school a very very odd message, pertaining to the name ‘Caitlin,’ which is her name, and ‘poop.’ It was literally one of the creepiest things I’ve ever seen: a two-page message about the name ‘Caitlin’ and its semantic relation to ‘poop.’ We found out that he had actually sent it to the first two hundred Caitlins he found on Facebook search.

- page 2 de 24 -