Created
July 2, 2009 19:40
-
-
Save sween/139662 to your computer and use it in GitHub Desktop.
This file contains bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Jabber IM with circuitshaman <[email protected]>7/2/09 2:38 PM | |
Ron Sweeney | |
ping | |
2:54 PM | |
circuitshaman | |
pong | |
3:02 PM | |
Ron Sweeney | |
are you familiar with document based db's ? | |
circuitshaman | |
familiar, yes | |
circuitshaman | |
used, no | |
circuitshaman | |
couchdb is the thing though | |
circuitshaman | |
fedora doesn't scale worth a darn | |
circuitshaman | |
it's poo | |
Ron Sweeney | |
wrong, but lets have another discussion about that... | |
circuitshaman | |
haha | |
Ron Sweeney | |
get off the couch | |
Ron Sweeney | |
anyway... | |
circuitshaman | |
no good? | |
Ron Sweeney | |
triple stores... | |
Ron Sweeney | |
mongodb > couch | |
circuitshaman | |
yes, I know those | |
circuitshaman | |
sweet | |
Ron Sweeney | |
if triple stores are just a storage mechanism for triples (pairs I assume)... | |
circuitshaman | |
usually fours in fact | |
circuitshaman | |
but at least triples | |
Ron Sweeney | |
ok, so these are like hash of hash of hashes ? | |
circuitshaman | |
well in the abstract sense, which is to say what a programmer cares about, they really are triples | |
circuitshaman | |
but the storage mechanism/method changes from triple store to triple store | |
circuitshaman | |
subject predicate object | |
circuitshaman | |
myOnt:Banana rdfs:subClassOf owl:Thing | |
circuitshaman | |
that's a triple | |
circuitshaman | |
but I guess I'm confused about your actual question | |
Ron Sweeney | |
not really, red light blinking on my forehead | |
Ron Sweeney | |
if I confused you, its because I dont know what im talking about... | |
Ron Sweeney | |
my underlying interest was to see if a document based db would be an alternative to a triple store... | |
circuitshaman | |
nope | |
circuitshaman | |
very very different needs | |
circuitshaman | |
I need to do some serious writing on the subject because there is jack squat out there to explain what this stuff really is and what it's good for | |
Ron Sweeney | |
right... im having a hard time mapping the niche it provides | |
Ron Sweeney | |
s/it/they/g | |
circuitshaman | |
it's kind of a chicken and egg situation. The problems that it's good for solving are usually ones that people don't think are solvable or at least not pressing | |
circuitshaman | |
he's an example at a high level | |
circuitshaman | |
When I say TB, what do you think? | |
Ron Sweeney | |
turberculosis, terabyte | |
circuitshaman | |
right, now how did you get there? | |
circuitshaman | |
and are those two thing related? | |
Ron Sweeney | |
not at all | |
circuitshaman | |
but in your head there is a mapping that happens on the acronym TB for both | |
Ron Sweeney | |
yup | |
circuitshaman | |
what's even more important is that while you know TB is an acronym for both turberculosis and terabytes are very different things and totally unrelated | |
circuitshaman | |
you an understanding of those words well beyond their spelling | |
Ron Sweeney | |
right, and one does not "overpower" the other, they are equal. | |
Ron Sweeney | |
meaning, one is not a "better" representation of TB | |
circuitshaman | |
sure, they're just related. But you know that relation isn't important in this case | |
circuitshaman | |
but if I asked, what do tuberculosis and terabyte have in common you'd have to conclude TB | |
Ron Sweeney | |
right, like that show on NPR | |
Ron Sweeney | |
dont let my banter hinder your lesson | |
circuitshaman | |
so with the idea that there is a far greater depth to a relationship than a hash map think of a relational database of the taxonomy of life | |
circuitshaman | |
and with that you also have diseases that are related to species | |
circuitshaman | |
and those species are susceptible to those diseases after a certain age. | |
circuitshaman | |
I'm not  | |
Ron Sweeney | |
ok, now stop. | |
Ron Sweeney | |
take species... | |
Ron Sweeney | |
its a "model" so to speak with properties too, like => 'age' | |
circuitshaman | |
well here's the point I'm making, try to come up with for a second the models that would make this work in a relational database | |
Ron Sweeney | |
ok, im tracking you... | |
circuitshaman | |
the whole taxonomy of species (kingdom phila,etc), disease that may have their own taxonomy, relationships between the two | |
circuitshaman | |
it gets insanely complicated very fast | |
Ron Sweeney | |
yeah, im there too. | |
circuitshaman | |
and the queries are ever more disgusting even if you manage to come up with a schema | |
Ron Sweeney | |
the queries even themsevles seem hard to "validate" or "test" | |
circuitshaman | |
right | |
Ron Sweeney | |
but, that is insanely powerful | |
Ron Sweeney | |
because, you know the answer to a question that hasn't been asked before really. | |
circuitshaman | |
exactly! | |
Ron Sweeney | |
i think this chat window is going to end up as a gist | |
circuitshaman | |
can you imagine a researcher trying to figure out which animal to test his new drug for a particular disease | |
circuitshaman | |
haha | |
Ron Sweeney | |
you'd first need to know which animal is suseptible | |
circuitshaman | |
right, and what age they are susceptible | |
circuitshaman | |
and maybe you want it to be a certain size | |
Ron Sweeney | |
weight | |
Ron Sweeney | |
right. | |
circuitshaman | |
right, the relational query is mammoth | |
circuitshaman | |
but it would look something like this in sparql: | |
Ron Sweeney | |
so populating this beast is just as easy ? | |
circuitshaman | |
SELECT ?animal WHERE { ?animal owl:Is_A Animal . ?animal susceptible_to MyDisease ...} and so on | |
circuitshaman | |
it's pattern matching | |
circuitshaman | |
adding to this beasty is just adding a series of declaration or triples | |
circuitshaman | |
Tiger owl:Is_A Animal | |
Ron Sweeney | |
ohhhhhhh | |
circuitshaman | |
and whether Animal is described further doesn't matter | |
Ron Sweeney | |
tiger owl platypus bsddevil:Is_animal | |
Ron Sweeney | |
thats a triple ? | |
circuitshaman | |
the store knows that Animal exists if only to be related to Tiger by owl:Is_A | |
circuitshaman | |
no, but I think you're going for: | |
circuitshaman | |
tiger owl:Is_A bsddevil | |
circuitshaman | |
owl is the web ontology language | |
Ron Sweeney | |
hahahaha | |
circuitshaman | |
it's a namespace of already declared things | |
Ron Sweeney | |
yeah, I actually knew that kind of | |
Ron Sweeney | |
just because this is going to be a gist... | |
Ron Sweeney | |
hahaa | |
circuitshaman | |
haha | |
circuitshaman | |
so obviously this is powerful stuff, no? | |
circuitshaman | |
it's using old computer science techniques (pattern matching and mathematical graphs) to attack the problem of complex relationships and semantic information | |
Ron Sweeney | |
absolutely... maybe this will help me get a grip a bit and use active_sesame to excercise what I've seen here | |
circuitshaman | |
well let me give you the new version | |
circuitshaman | |
it will make way more sense | |
circuitshaman | |
the original version is quite the hack and obfuscates way too much | |
Ron Sweeney | |
let me give you a use case..., you tell me if this could help | |
circuitshaman | |
sure | |
Ron Sweeney | |
ICD-10 | |
circuitshaman | |
sounds good so far  | |
Ron Sweeney | |
the relationship of the code to the procedure to the related procedures to make up a single code for cascading procedures | |
circuitshaman | |
yes | |
circuitshaman | |
that would be an excellent and extensible use case | |
Ron Sweeney | |
dood, you're a rock star. | |
circuitshaman | |
hahaha | |
circuitshaman | |
I'd love to work on it with you | |
circuitshaman | |
having someone actively involved in solving a real problem would help me finish the library | |
Ron Sweeney | |
I have 2 weeks off this month...im hacking every single day of them... | |
circuitshaman | |
plus I'd love a test writer, it's never been my strong suit | |
circuitshaman | |
hahaha | |
circuitshaman | |
I'll throw up the latest version as a branch on github | |
circuitshaman | |
in the mean time you'll want to grab allegrograph | |
Ron Sweeney | |
yup, seems we've gotten to this point before, I will make it happen now with my recent enlightenment... | |
circuitshaman | |
haha | |
circuitshaman | |
you also have a clear problem | |
circuitshaman | |
and I've recently started writing it again | |
circuitshaman | |
it needs to be called open sesame though | |
Ron Sweeney | |
hahahah | |
Ron Sweeney | |
awesome, gotta run, have a good holiday weekend. thanks too! | |
circuitshaman | |
no problem, enjoy the fight! | |
circuitshaman | |
that's way cool | |
Ron Sweeney | |
yup, we'll take lotsa pics | |
circuitshaman | |
nice |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment