|
|
crazygeekyblatheridea
|
|
anne-girl
|
dude i just had the craziest idea and it'd never work, and i wouldn't know where to start, but here goes... the idea is... we index all of blather into a database and use a distributed computing approach to do said thing like == make app to grab bits and pieces of blather and figure out from grabbed pages useful information like who's_blathed_on_this_page and stuff and it'd be a gigantic database and take tons of processor time but if someone made an app... and people downloaded it... and it was made in such a way that the computed data could be uploaded to an anonymous ftp server or something which could then be integrated into database for length-of-time-it-might-take considerations all i know is that making index_of_i_v2 and friends took at least an hour each (the bottleneck's downloading the pages, for that operation at least) (with perl code, said app would preferably haveta be written in c++ or something fast) that'd be cool imagine the crazy blathertools one could make with such a database? it'd be, like, magical! in a geeky way, anyway and it might be too much information... blather's nice the way it is whaddya think?
|
050706
|
|
... |
|
Doar
|
I'm not sure what all of that idea involved but it sure does sound geekerific...ummm....doodly....deeederrrr...... er....
|
050706
|
|
... |
|
u24
|
we could ask sage for (read only) access to the actual database... i doubt it though. I'm in the process of making a blather_lib for PHP...
|
050707
|
|
... |
|
u24
|
are you thinking of actually re-creating the blatherdb?
|
050707
|
|
... |
|
crazy spinning girl
|
ok, and now in laymens terms? you load all of blather into a database... you lost me after that. what does it do?
|
050707
|
|
... |
|
anne-girl
|
yeah... asking sage occured to me, but i didn't think so either not exactly recreating the blatherdb... i don't know exactly what's in the database already, but i think putting a tad more information than's in there would make sense note that i've never actually worked with databases... i just know that they hold information in laymens terms: (i'll try) load all of blather into a database and then us geeks can use the database to make cool blather tools because it has all of blather's information inside it, and makes it really easy to access so the database by itself doesn't really do anything like, for example, if we had a blatherdb, could make a sparkling new interface to blather, with jumping monkeys and such and have proper indices (no more index_of_i) and maybe even *gasp* a good blather search not that i know anything about either databases or search i'm also kind of unsure about whether spiffy new tools and features would actually be a good thing
|
050707
|
|
... |
|
anne-girl
|
problems -- would downloading massive amounts of blather cause undue strain on blather's servers? -- would it just take too long? -- how much space does blather take up? -- how much space would blatherdb take up?
|
050707
|
|
... |
|
u24
|
thisisblather might help, though it's mostly outdated, badly formatted conjecture. I'm good with SQL, but it'd take a -very- long time to process.
|
050708
|
|
... |
|
anne-girl
|
bah i dunno
|
050708
|
|
... |
|
u24
|
although as a distributed app, it's so crazy it just might work.. hang on, you've got a list of (nearly) all blathes, Ja? so we could have a central server that gives chunks of 50 or so blathes to each client that connects (similar to seti's work units) then each client could strip each of it's pages, and add that data to its own SQL database then, when its finished its work unit, send the SQL back to the central server. then, periodically, the central server merges the uploaded databases into the main db.. then, when the list is done... how to keep it up to date..... hmm.. suggestions?
|
050725
|
|
... |
|
anne-girl
|
keeping it up to date should be fairly straightforward - just use something analogous to the app I wrote to get all the names of blathes grap all the showday pages from the last update to now, sort && remove duplicates (the central server could maybe do it itself, just write an update script and have it run as a daily cron job or something)
|
050725
|
|
|
what's it to you?
who
go
|
blather
from
|